Unstabilized HeNe lasers - the more conventional type that most people are familiar with from Physics 101 - are covered in the chapter: Commercial HeNe Lasers.
For current production lasers, the manufacturers' Web sites often provide basic specifications. A Google search is usually the easiest way to find them, but most are also linked from the chapter: Laser and Parts Sources. For older lasers, it's often difficult to obtain detailed specs so estimates based on physical size, and then testing may be the only option.
An introduction to stabilized HeNe lasers and typical locking schemes can be found in the chapter: Helium-Neon Lasers.
Nearly every model of stabilized HeNe laser ever sold commercially is listed in the chart below. Many are described in separate sections of this chapter, arranged approximately in alphabetical order by manufacturer. For these, the level of detail here is probably several orders of magnitude greater than from any other source, except perhaps the operation and service manual for the laser (which with few exceptions, is generally not available in the public domain). Where there is no entry for a particular laser, a Google search using the manufacturer (with or without model number) will usually locate what little information is available. Sometimes, a research paper referencing the specific laser will even have the most information!
Unless otherwise noted (below), these data were obtained from manufacturers' Web sites, brochures, spec sheets, or user manuals for each laser. Contributions (including stabilized HeNe lasers I've missed) and corrections are welcome. Please contact me via the Sci.Electronics.Repair FAQ Email Links Page.
All stability values are in parts per billion (ppb).
<--- Frequency Stability Time Scale --->
Model Type/AP Sec Min Hour Day Year Life
-------------------------------------------------------------------------------
*Aerotech LZR2000 (20) SM M +/-2 +/-20 (1 mo)
*Aerotech LZR3000 (20) SM M +/-2 +/-20 (1 mo)
*Aerotech S100 (2) SM S +/-1 +/-2 +/-3 (8 hrs)
*Axsys 150 (8) DM M +/-2 +/-6 +/-20 (24 hrs)
CDHC-Optics DH-HN250P SM S
*Coherent 200 DM S +/-2 (5 min) +/-10 (long)
*Excel 1001A/B/F AZ M 20 (unspecified time)
Feanor LN 10 DM M +/-1 +/-80
Feanor LP 30 AZ M +/-2 +/-20
Feanor LSP 30 AZ M +/-2 +/-20
*Frazier 100 (21) I2 S +/-0.01
*HP-5500A/B/C, 5501A/B AZ M +/-2 +/-20
*HP/Agilent 5517 (all) AZ M +/-2 +/-20
*HP/Agilent 5518A, 5519A/B AZ M +/-2 +/-20
*HP/Agilent N1211A (22) AZ M +/-5 (5 min)
JENAer ZL 600 (10) DM M 2.0 20
JENAer ZL 700 (10) DM M 2.0 20
JENAer ZL 800 (10) DM M 2.0 20
*Lab for Science 200 DM S 0.03 0.05 0.2 0.5
*Lab for Science 210 DM S 0.03 0.05 0.2 0.5
*Lab for Science 220 TZ S 0.01 0.02 0.05 0.2
*Lab for Science 260 TM S 0.02 0.02 0.1 0.4
*Laseangle RB-1 (3) DM S 0.01 0.1
Lasertex Frequency Standard I2 S +/-0.001 (10 s) +/-0.025
Lasertex HPI-3D SM M +/-1 (short term) +/-1
Lasertex LL 10 DM M +/-2 (short term) +/-30
Lasertex LS 10 DM M +/-2 (short term) +/-30
Lasertex LSP 30 AZ M +/-2 (short term) +/-20
Laser Metric Systems SFL-1 DM M 2 (unspecified time)
Limtek LS 10.3 GP M 20 (unspecified time)
LINOS FS Series DM S +/-2 +/-10 +/-20
LINOS FAS Series DM S +/-1 +/-2 +/-10
*Mark-Tech 7900 DM M +/-2 (const. temp.)
Mark-Tech 7910 (6) DM M +/-2 (const. temp.)
*Melles Griot STP-901 (4) DM S +/-1 +/-4 +/-6 (8 hrs)
*Melles Griot STP-909/911 (2) SM S +/-1 +/-2 +/-3 (8 hrs)
*Melles Griot STP-910/912 (2) SM S +/-1 +/-2 +/-3 (8 hrs)
Micro-G ML-1 DM S 0.2 (10 ms) 1.6 (const. temp.)
NEOARK NEO-262 TZ M 1 (unspecified time)
NEOARK MODEL-430 (11) DM S 30 (unspecified time)
NEOARK 430-R4 (11) LD S 10 (unspecified time)
NEOARK NEO-9111 (11) AZ S 1 10 (1 wk)
NEOARK NEO-92SI-NF (11) I2 S 0.025
NEOARK NEO-OL101K (11) OL S 0.0001 (10 seconds)
NEOARK NEO-2MSS (11) PS S
NEOARK NEO-5MSS (11) PS S
*Newport NL-1 (3) DM S 0.01 0.1
*Nikon NKL-85 (14) LD S
Nikon NN-1 I2 S
NPL Hexagon (13) ?? S 0.01 +/-2
NPL I2 543 nm I2 S +/-0.25
NPL I2 633 nm I2 S +/-0.2
Optodyne L-103 ?? M
Optodyne L-104 ?? M
*Optodyne L-109 DM M
*Optra Optralite AZ M
*Perkin Elmer 5800 LD S
*PLASMA LGN-212 AZ M 10 (unspecified time)
*PLASMA LGN-302 DM S 10 (unspecified time)
*PLASMA LGN-303 DM S 10 (unspecified time)
*PLASMA LGN-304 DM S 10 (unspecified time)
*Renishaw HS-10 DM M +/-100 (unspecified time)
*Renishaw ML-10 DM M +/-20 +/-50
*Renishaw XL-80 DM M +/-50 (unspecified time)
REO SHL DM S +/-2 +/-2 +/-4
SIOS SL 02 DM S +/-2 +/-10 +/-20
SIOS SL 03 DM S +/-1 +/-2 +/-10
*Spectra-Physics 117 (5) DM S
*Spectra-Physics 117A (4) DM S +/-1 +/-4 +/-6 (8 hrs)
*Spectra-Physics 117B (5) DM S +/-1 +/-4 +/-6 (8 hrs)
*Spectra-Physics 117C (5) DM S
*Spectra-Physics 119 (7) LD S +/-2
*Spindler and Hoyer ZL-150 AZ M
*Teletrac 150-IV (8) DM M
*Teletrac 150 (Long) (8) DM M +/-2 +/-6 +/-20 (24 hrs)
*Teletrac 150 (Short) (8) DM M +/-2 +/-6 +/-20 (24 hrs)
*Teletrac 150 (Long) (8) AZ M +/-2 +/-6 +/-20 (24 hrs)
*Thorlabs HRS015 (19) DM S +/-2 +/-2 +/-4 (8 hours)
Tropel 100 (15) DM S
*Tropel 200 (15) DM S +/-2 (5 min) +/-10 (long)
Uniphase 1220 (16) DM ?
VM-TIM LHN-212-1 (18) AZ M +/-10 (8 hrs)
VM-TIM LHN-303 DM S +/-10 (8 hrs)
VM-TIM LHN-220SF (18) ?? S +/-10 (8 hrs)
*Wavetronics WT307 (all) (17) AZ M +/-2 +/-20
Winters 100 I2 S 0.025
Winters 200 (12) I2 S 0.025
*Zygo Axiom 2/20 (9) DM M +/-2 +/-10 +/-100
*Zygo 7701/7702 (9) DM M +/-2 +/-10 +/-100
*Zygo 7705 (9) AZ M +/-10 +/-20 +/-200
*Zygo 7712/7714 (9) DM M +/-0.5 +/-1
*Zygo 7722/7724 (9) DM M +/-0.5 +/-10
The * denotes lasers that are covered in detail elsewhere in this chapter.
Type Legend:
AP (Application) Legend:
A metrology laser can generally also be used for scientific/research applications since all have very tightly controlled optical frequency. And while the converse is often (but not always) true in principle, it's not usually practical or worthwhile except for experimental purposes since metrology systems may require laser characteristics (like two-frequency) that aren't present in laboratory stabilized lasers. In addition, the optics and cabling/electronics requirements would likely make their adaptation potentially complex, if possible at all.
Notes:
Among commercial instruments, the Coherent LaserCheck and Sper Scientific 840011 are convenient and relatively low cost. However, for observing the warmup and locking behavior of stabilized lasers, a power meter with graphing capability or a data acquisition system (possibly multi-channel) may be desirable.
Commercial SFPIs are available (though not as many as in years gone by). Suppliers include Ophir and Thorlabs. Expect to spend $2,500 or more. However, it's possible to build an SFPI capable of easily checking for SLM operation of stabilized HeNe lasers (as well as doing a lot more) for under $100 (excluding the required oscilloscope). See the sections starting with: Scanning Fabry-Perot Interferometers.
The simplest tests would be to monitor the output (or orthogonal polarizations if both are present) for a few hours with a recording laser power meter or photodiode and data acquisition system. If high frequency noise is a concern, use a fast photodiode and RF spectrum analyzer to search for residual ripple from the HeNe laser power supply and PZT/heater driver making its way into the output.
However, to determine the optical frequency to 10 or 11 significant figures requires an iodine-stabilized HeNe. If you have to ask what that is, you definitely can't afford one. ;-)
When making measurements on the output of most lasers, but particularly stabilized HeNe lasers, whether using a power meter, Scanning Fabry Perot Interferometer (SFPI), or another test instrument, care must be taken to avoid back-reflections into the laser that may, well, destabilize it. With some, even slight contamination on the surface of the output mirror, or a piece of clear tape over the output aperture will cause lock to be lost resulting in random fluctuations in output power and/or optical frequency. With most of these, no harm is generally done, but they would then more appropriately need to be called destabilized lasers. :)
Calibration was done using a carefully wound single layer 160 turn, 94 mm long electromagnet solenoid at 1 A, which works out to 21.4 g.
The gauss meter is perfect. :) But real World magnets are not. They actually quite terrible in terms of field uniformity. For the typical HP/Agilent magnet, only the central 1/3rd or so is even reasonably constant with the field strength tapering off toward the ends. Yet, the active discharge in these lasers is often exactly the length of the magnet! And, the field strength may not be symmetric even well inside as these Alnico magnets can have large local variations either from the way they are manufactured, or from deliberate or accidental demagnetization. For example, simply rolling a 20 penny (~3") steel nail around the outside will reduce the field inside by a few percent permanently. And much more extreme effects are possible by applying and removing Alnico or rare earth magnets. More on these effects in conjunction with HP/Agilent lasers, below.
With a single photodiode sampling the beam, only the amplitude of one of the two polarized modes can be stabilized. Even so, the frequency stability will still be quite good once the laser tube has reached thermal equilibrium and its power has leveled off. As the tube ages and its power declines, the output power from the laser will remain constant until it approaches what's available from the HeNe laser tube. At that time, it will lose lock and may even be damaged, more below. With dual mode frequency stabilization, locking will still be possible even when the power output from the tube is very low because it is the difference of the polarized mode amplitudes that produces the error signal, not a specific value. In addition, with intensity stabilization, the frequency will drift as the tube gets weaker and the lock point moves with respect to the neon gain curve. Why frequency stabilization was not implemented instead, or in addition to intensity stabilization as in the 05-STP-901, is a mystery as it would have been a very straightforward enhancement - primarily a second photodiode! Nearly everything else is already there. More on this below.
Lasers based on the Aerotech technology are now sold as the Melles Griot STP1 with specific model numbers of 25-STP-910 and 25-STP-912. [They may also be found as 05-STP-910 and 05-STP-912. Whether "05" or "25" is used simply depends on if it is considered a component (05) or system (25), and sometimes at random!] There was also an 05-STP-914, now discontinued. It had an output power similar to that of an 05-STP-910, but a larger diameter beam with lower divergence. The Melles Griot lasers are physically and functionally very similar to the Syncrolase 100 but both of these use a separate HeNe laser power supply instead of having one inside the laser head. (The Melles Griot 25-STP-909 and 25-STP-911 had the built-in power supply but have been discontinued.) While, it is not known how much the electronics differ compared to the S100 version, all indications are that very little has been changed since obtaining the technology as part of the acquisition of the HeNe laser division of Aerotech. Melles Griot calls them "frequency stabilized lasers" though the description indicates that the same amplitude stabilization technique as the Syncrolase is used (and examination of the locking adapters confirms that there is only a single photodiode). Searching for "Melles Griot 25-STP-910" or "Melles Griot STP1" should return a description and spec sheet. Here is a summary of the specifications for the Melles Griot versions:
And, if you'd like to order a few, the Melles Griot price in 2012 was $4,388 each for the low power version (0.5 to 0.95 mW) and $4,662 each for the high power version (0.6 to 1.4 mW). I wonder how they came up with those especially round numbers for the prices.
The output power is a user adjustment (a trim-pot) that sets the intensity stabilization point, and indirectly the operating frequency. In addition to versions based on output power, the Syncrolase came in two versions based on whether a pair of DC wall adapters were used to power an internal HeNe laser power supply and the locking controller, or whether the laser head had a standard Alden connector to attach to an external lab-style HeNe laser power supply, which is included in the price, along with the wall adapter for the locking controller! :) Now, why weren't the two combined, given the warning in the operation manual: "Application of power to the SFA adapter (locking controller) in excess of 5 minutes with the head de-energized may damage the SFA adapter". There's more on this below.
One of the unique features of this system is that rather than using a resistance heater over a substantial part of the HeNe laser tube as is done in most commercial stabilized HeNe lasers, these lasers use a compact "locking adapter" attached to the end of the laser head containing a coil surrounding only the OC mirror mount stem to directly heat the metal mount via RF induction. A very simple MOSFET driver can provide over 10 W directly to the mount resulting in a very rapid response. Based on tests I've done, I estimate that at maximum RF power, it will increase the temperature of the mirror mount stem itself by greater than 1 °C per second. This is more than an order of magnitude faster than traditional resistance heaters surrounding the glass portion of the tube. A temperature sensor in close proximity to the mirror mount stem senses its temperature and is used both to switch the feedback loop on when hot enough, as well as to shut the heater off if the temperature goes too high. Warmup to fully stable operation still takes 20 or 30 minutes because the rest of the laser head has to come into thermal equilibrium as well as the mirror mount stem. But, initial locking is very quick - typically 3 to 5 minutes. And once locked, it should use less power and be more immune to ambient temperature variations, and the faster response also improves frequency stability.
The same locking adapter may be used with any compatible laser head requiring at most minor electronic adjustments. In addition, the use of this technique allows for the possibility at least in principle of converting almost any HeNe laser tube with a suitable mode structure into a stabilized laser by simply attaching the locking adapter to its output end. However, in practice, minor details like the mirror mount stem dimensions and the long exhaust tip-off may make this rather difficult.
The DC wall adapter (either version, 1 or 2 required depending on the model of the laser) is rated 13 VDC, 1.3 A. Measurements show it to have an open circuit output of 16.5 V. The plug is 5.5 mm/2.5 mm center positive. Since there is a 7812 +12 V regulator in the controller (see the schematic below), the output of that DC adapter must be greater than about 14.5 V to assure proper regulation. So, at least once the feedback loop is closed, the input voltage should never dip below 14.5 V. I do not know the official specifications for the external HeNe laser power supply (where required), but based on the length of the tube and other typical Aerotech tubes (and the supply that comes with an OEM version of the 05-STP-910), it is probably around 1,500 V at 4 mA.
That Melles Griot OEM version (typically found in some high-end wavelength meters as a reference) is otherwise identical to the normal one except that its output is fiber-coupled. One example is shown in Melles Griot Fiber-Coupled 05-STP-910 Stabilized HeNe Laser. There is an additional assembly that screws onto the output end of the laser (and is then glued) with a 4 position shutter which can be set to block the beam, pass it to the fiber, or divert it at right angles out the side so the laser can be set up independent of the fiber. Some versions have a fully adjustable fiber port enabling a broken or damaged fiber to be replaced and realigned relatively easily, while others are totally glued with rock-hard Epoxy at the factory with no chance of alignment in the field. This also means that even removing and reinstalling the laser head - or replacing it with a new one - is likely to affect alignment in a way that is difficult to remedy. A more recent version uses a modified laser head assembly that bolts on rather than screwing on, which is more precise and immune to misalignment. See Melles Griot 05-STP-910-536 Reference Laser for the Agilent 86122B Multi-Wavelength Meter. It's functionally identical to the one shown above. Elongated holes in the laser head flange allow for the polarization orientation to be fine tuned. An adjustable fiber coupler attaches to the output-end of this laser. (A Web search will easily find information on the 86122B including specifications and operation manual.) Some versions also include a beam sampler and photodiode so that the output power can be monitored prior to the fiber. And as can be seen, these OEM systems also have the status LED and power connections brought out as twisted wire pairs:
The HeNe laser tube in the Syncrolase 100 and 05-STP-9xx lasers is between 6 and 8 inches long, depending on version. The ballast is typically 80K ohms made up of multiple thick film resistors on a ceramic substrate potted in a rubbery ring that slips onto the anode mirror mount stem. The default current is 4 mA for all STP-9xx heads (though it may have been 4.5 mA for at least some Aerotech versions). New STP-910 tubes will remain lit down to 3 mA or less; new STP-912 tubes down to 3.5 mA or less. This dropout current tends to increase as the tube is run. Eventually, it may be necessary to turn up the power supply current to 4.5 or even 5 mA to squeeze out a few months more life from of a high mileage tube.
A common 6 to 9 inch random polarized tube with cathode-end output (high voltage far away from the electronics!) would probably work except that the mirror mount stem needs to be a about an inch long with the exhaust tip-off cut off close to the end-cap so as not to interfere with the coupling coil assembly. Very few tubes have these characteristics, though some are close enough to be usable in a pinch. In addition, using too long a tube might result in a second longitudinal mode being present if the Output Adjustment is set so the main lasing line is too close to the neon gain center.
For details on theory and implementation see U.S. Patent #4,819,246: Single Frequency Adapter.
A schematic diagram of the electronics for the Syncrolase 100 can be found at Schematic of Aerotech Syncrolase 100 Controller. This may not yet be quite complete and numerous errors are possible since the PCB is tight, it is a 4 layer board, and the soldermask is almost totally opaque. It was not much fun to trace the circuit. Part numbers are not available for a half dozen components because (1) they might have been obscured and (2) there were several added parts that appear to be in the "oops" category. :-) (Those added parts relate to the overtemperature protection - more on this below.) But I bet this schematic provides infinitely more information than what's available anywhere else! :) Melles Griot has redesigned the PCB at least twice, the latest version using mostly surface mount parts. (Photo and comments on this and the other controllers below and in the next section.)
The gate of a power MOSFET is driven by a simple oscillator, running at between 500 kHz and 1 MHz (I measured about 700 kHz on one unit). The feedback signal is summed into the gate junction from the error amp and serves to modulate the output of the induction heater to maintain lock once the operating temperature has been reached. The coil is just short of 9 turns of #24 AWG wire close wound on a 1.35 cm form. Due to the way the leads enter through the back of the form, the final turn is short changed! :) This is probably not terribly critical though.
(The following applies directly to the original Aerotech design except as noted.)
The RF driver consists of a Hex Schmitt trigger (MC14584BCP similar to a CMOS 40106) with one section used as the oscillator and the remaining sections paralleled to buffer its output. An RC network converts the squarewave of the oscillator to a roughly triangle waveshape at the MOSFET gate. The output of the Error Integrator feeds into the gate as well with the effect of modifying its DC offset. Since the MOSFET gate threshold is fixed, this produces a modulation effect which is a combination of amplitude and pulse width, with the net result of controlling the amount of RF power transferred to the HeNe laser tube mirror mount stem. A significant part of the capacitance in the waveshaping network is the internal input capacitance of the MOSFET gate itself, and this may exceed 1 nF. Thus, it's possible that if the MOSFET needs to be replaced, the value of the capacitor between the gate and ground (C13) may need to be adjusted as well to maintain approximately the same net capacitance and waveshape. The MOSFET gate capacitance can vary by a factor of over 2:1 between MOSFETs with the same part number, or by even more if a MOSFET with otherwise acceptable specifications is substituted. On the unit I have, it was about 1.3 nF.
Newer versions include a ULN2003 Darlington array, possibly for driving the MOSFET in place of the HEX Schmitt Trigger. (But that is still present.) They may also use a thermistor sensor in place of the thermocouple - it looks like a 1N4148 with no markings but tests like a 10K ohm resistor at room temperature. That's much cheaper and easier to use!
The control functions are implemented in the four sections of a TLC27L4CN quad op-amp as follows:
The output of A1A also feeds the Over-Temperature protection circuit that is supposed to turn the heater off if the temperature goes too high. However, on early units, this almost looks like an afterthough with its adjustment pot hanging in mid-air!
When powered up, the temperature sensor is initially cool so the RF driver comes on at full power. When the mirror mount stem reaches the operating temperature (something like 80 °C in 30 seconds or so), the feedback loop becomes active and the SYNC LED comes on indicating that lock is being maintained at the current mode position. However, since the remainder of the laser tube is still increasing in temperature due to the normal heating of the discharge and hasn't reached thermal equilibrium, the RF drive gradually decreases cooling the mirror mount stem so that the total distance between the mirrors remains constant. Eventually, when the mirror mount temperature gets to be too low, the system will switch back to continuous heating for a time based on the hysteresis of the Sync Enable Comparator. After 20 to 30 minutes, the laser tube will reach thermal equilibrium and the system will then remain locked forever. (Unfortunately, many people take this literally and leave the laser on until it dies, which is considerably sooner than forever!)
On the Melles Griot locking adapter, a bi-color LED is used that serves a dual function and is labeled "Stable/Overtemp". It can be off, green, or red:
Note that the LED being green doesn't mean the output is stable, only that the feedback loop is active. For example, the laser could still lose lock due to back-reflections or be unable to lock as a result of the output level being set too high. (The latter should not be possible on a new laser if correctly set up at the factory, though it could occur with a high mileage tube that has lower output power.)
Here are some photos:
There are at least three major variations on the design and PCB layout of the locking adapter controller:
There is more on the locking adapter PCBs in the next section.
The coupling coil assembly on the first Syncrolase 100 of mine had disintegrated due to excessive temperature. (Actually the magnet wire and its insulation is in fine shape but the plastic form on which the coil was wound is no longer intact and it's not even possible to determine much about it.) I've tested the induction heating winding a test coil on a tube made from insulating plastic sheet. The effect is impressive considering the simplicity of the circuitry (see the schematic below) raising the temperature of a dummy mirror mount stem by more than 1 °C per second even with a coil that is probably far from optimal.
I do not know for sure if the cause of the destroyed coil form was due to a part failure rather than simply a result of the laser being been left on for 7 years continuously! :) The HeNe laser power supply was indeed dead, probably due to the tube being very hard to start and impossible to run for more than a few seconds regardless of power supply or ballast resistance. So it's possible that when the tube decided it was tired of doing its thing and the power supply shorted out, the controller ended up cycling on the over-temperature condition. The Melles Griot manual does warn against running without the laser on. And, electrical tests seem to indicate that the controller is working properly.
It's likely that the Over-Temperature (OT) adjustment was incorrect and too high all along. Since it's not something that affects normal behavior, it would be all too easy to neglect setting it properly! I've also been told by the former owner that this laser always ran very hot. If the tube fails - even if someone forgets to plug in its wall adapter! - the heater tends to be on and bad things can then happen if the OT setting is too high. Ask me how I found out. :( :) OK, I'll tell you. I acquired another Syncrolase with a good tube but that would not stabilize. I traced the problem to what I believe may have been a short in the temperature sensor and then adjusted it to operate at a reasonable temperature set-point. But I accidentally left the controller powered after turning off the laser and went away. When I returned (after lunch!), the entire assembly was too hot to touch and the platic coil form at is cover had melted!!! Apparently, either the OT setting was way too high (it's possible someone before me messed with it) or it isn't effective or was broken.
Interestingly, on one of those rare occasions where I was able to get the tube to remain on long enough with a lab power supply to watch a few mode sweep cycles, it is a classic FLIPPER! I suppose that the flipperitis could have happened in its old age (it is also weak - about 0.7 mW - and with brown crud in the bore), but normally the flipper or non-flipper status of a tube doesn't change over the course of its life. I do have another Aerotech laser head that would screw right on to the controller but it too is a flipper! :( :) In fact, its behavior shown in Plot of "Flipper" Aerotech OEM1R HeNe Laser Head During First Part of Warmup and the merged version in Plot of "Flipper" Aerotech OEM1R HeNe Laser Head During First Part of Warmup (Combined) looks virtually identical to that of the Syncrolase tube (over the few mode sweep cycles I could see before the tube went out). But, even more interstingly, the flipping of the tube in the plots ceases entirely and it becomes perfectly well behaved once nearly warmed up as shown in Plot of "Flipper" Aerotech OEM1R HeNe Laser Head at Transition to Normal Behavior (Combined). Perhaps that tube was intended for a Syncrolase as it in unusual in having the required long mirror mount stem and short cutoff exhaust pipe. Perhaps it was a reject due to the flipping. Or perhaps for unknown reasons, all these tubes flip when cold. Since the Syncrolase 100 would be operating well beyond this point, there's a chance that the flipping is irrelevant and it would work just fine. In fact, that one working genuine Syncrolase tube is also a flipper until it warmed up! More on this below.
I built a replacement coil using the wire from the first dead Syncrolase on a roll of plastic. It works, though the temperature response is faster probably because the thermocouple is not in the same location as the original. So, it locks more quickly, but also loses lock more frequently during warmup but is otherwise functional. Perhaps changing the temperature set-point would correct that. It's amazing how much variability can be tolerated with this design.
Adjusting the temperature set-point is an interesting exercise. Ideally, it should be slightly above the equilibrium temperature of the laser head with only the laser tube powered. Set too high and the laser will run excessively hot, but there will be a fewer number of lost lock events during warmup. Set too low and it may lose lock eventually when the tube equilibrium temperature exceeds the set-point temperature.
One way to do the adjustment might be to initially set the Temp. Gain pot (R1) fully CW (for a very low temperature) and power *only* the laser head (not the controller) for at least an hour so it reaches thermal equilibrium. Then, power up the controller and slowly turn R1 CCW to slightly beyond the point where the SYNC LED goes out. Monitoring the Temperature Amp output (A1 pin 1) will indicate how effective this is. The voltage on A1 pin 1 should remain between approximately 1.5 V and 2.75 V when the laser is locked. If it goes below about 1.5 V, the feedback loop is disabled and the heater turns on full (SYNC LED OFF). This state continues until the temperature increases to the point where A1 pin 1 exceeds about 2.75 V and the feedback loop is enabled (SYNC LED ON). Better to start out with the temperature set-point adjusted too low should the over-temperature protection fail. :( :)
I built another temporary coil for the first laser to check it out. This coil is wound on a plastic cylinder found in a junk pile that was glued to the remains of the original coil form. The Epoxy seems to stick rather well, which is a bit surprising. I didn't have any #24 AWG magnet wire, so I used #20, which just fits 9 turns in the available space. The laser works quite well now except that the speed of heating is not quite as fast, possibly due to the coil being slightly longer and larger in diameter. However, this is probably of little consequence in the grand scheme of the Universe. :) Lowering the RF frequency improved the response, though there was no resonance.
Finally, I built a new coil form for the third laser. It has approximately the same dimensions as the original so it behaves very well. But the plastic is too think and there is very little clearance between the form and mirror mount stem. So the genuine Syncrolase laser head won't fit because its tube is too off-center. (This must have been a result of the way it was manufactured since its beam is well centered.) But my "flipper" head fits just fine and works just fine. :)
Power for the locking adapter is 13.5 to 15.5 VDC at 2 A max. The center contact of the 5.5 mm/2.5 mm power jack is positive. On the OEM Melles Griot, version with yellow and blue wires, yellow is positive. Also on the OEM version, the bi-color status LED is brought out to red and green wires. If no LED is already present, connect a two-pin bi-color LED between the red and green wires such that positive on red results in green light. Or wire up a pair of red and green (normal) LEDs in parallel with opposite polarity.
A laser head with an Alden connector can be powered either from a lab supply for a 1 to 2 mW HeNe laser, set to 4 mA, or a suitable HeNe laser power supply brick. The brick provided with the OEM systems runs on 12 VDC. For the laser head with built-in HeNe laesr power supply, the spec is the same as for the locking adapter - 13.5 to 15.5 VDC, center positive. (Aerotech provided a pair of identical wall adapters rated 13 VDC but which actually put out much more.)
There is only a single user adjustment - the set-point for output power.
There is only a single indicator - the "Locked" LED (Aerotech, red) or "Stable/Overtemp" LED (Melles Griot, green/red).
If available, use a fast responding laser power meter to monitor the output. For the Melles Griot fiber-coupled lasers, the output of the fiber can be used, though it's best to measure power in the raw beam if it is accessible to determine the health of the laser tube. On some versions, there is a four position shutter wheel in the fiber-coupler assembly. Remove the hex cap screw in the shutter and rotate the shutter by plus or minus 90 degrees - One of these positions places a mirror in the beam path to divert the beam out the side. (Of course, if nothing comes out of the fiber when the laser is powered regardless of shutter position, the fiber alignment may be bad or the fiber may be broken - or the tube may be dead.) But not all versions have a shutter wheel. Others may have a beam-sampler plate with a silicon photodiode, so that can be used to monitor relative power.
Before applying power, make sure the laser head is securely attached to the locking adapter with the polarization references lined up. The head should be screwed in nearly all the way with the locking ring tightened against it, and the set-screw tightened. (If there is no polarization reference mark on the laser head, the orientation will need to be determined once it is powered, see below.)
CAUTION: For the fiber-coupled lasers, even just removing and replacing the laser head may cause alignment to be compromised. On those with an adjustable fiber port, that can be fine tuned to optimize alignment. But on those with no fiber alignment adjustments, this could be bad. So, if the head is originally tightly secured, don't loosen it. If it is already loose, then it may be possible to find an orientation very close to the optimal based on the polarization reference marks where alignment is acceptable with the lock-ring tight. Additional details are left for the advanced course. :-)
CAUTION: If the laser beam starts flickering at any time (not to be confused with the normal power variation due to accelerated mode sweep, or induced mode flipping from back-reflections), the HeNe laser power supply current may be set too low and/or the tube may be high mileage with an increased dropout current. Immediately adjust the power supply or power off and replace the power supply (or tube). This condition will not magically recover on its own. A current 0.5 mA higher than the default (4.0 mA for the Melles Griot lasers) should be acceptable with little impact on performance. But once the dropout current reaches this point, life expectancy is probably measured in months, not years. If increased current doesn't help, the tube may be end-of-life, have a damaged ballast resistor, or the power supply may not be operating properly. DO NOT allow a flickering condition to continue as the tube and/or HeNe laser power supply may be damaged.
CAUTION: DO NOT allow the locking adapter to be powered if the laser is off or doesn't start for any reason. While there is supposed to be protection against an over-temperature condition, don't count on it, especially for the Aerotech version.
Aerotech S100 Syncrolase controller
Closeup of Aerotech Syncrolase 100 Controller - Left Side View and Closeup of Aerotech Syncrolase 100 Controller - Right Side View show what is probably the original design. As can be seen, there are 4 trim-pots on the PCB (From left to right: Temp. Gain, RF Frequency, PD Gain, and the user-accessible Output Level Adjust), plus the one hanging in mid-air which is Temp. Limit, probably being an afterthought added after too many locking adapters self destructed).
Trim-Pot Function Comments
---------------------------------------------------------------------------
R1 Temperature Gain Adjusts temperature set-point
R? Temperature Limit Prevents meltdown if laser unpowered
R10 RF Frequency Controls power to induction heater
P12 P Mode Photodiode Gain Sets range of user set-point adjustment
R? Output Level Set-Point User adjustment via hole in cover
The Schematic of Aerotech Syncrolase 100 Controller applies directly to this version. Clearly, some engineering changes were needed as in addition to the floating trim-pot, several componenst are installed at peculiar angles and generally shoe-horned into place. ;-)
Melles Griot 05-STP-9xx Syncrolase through-hole controller
At some point after Melles Griot acquired the Syncrolase, they redesigned the PCB (still through-hole) and added a connector so the PCB or induction coil and temperature sensor assembly could be easily replaced without requiring any soldering. They also eliminated both of the temperature trim-points - these are presumably set up during initial testing and should not change when replacing laser heads.
Trim-Pot Function Comments
---------------------------------------------------------------------------
P1 P Mode Photodiode Gain Sets range of user set-point adjustment
P2 RF Frequency Controls power to induction heater
R22 Output Level Set-Point User adjustment via hole in cover
There is also a 4 pin header with test-points (pin 1 on the right):
Test-Point Function --------------------------------------------------------- Pin 1 P Mode (output) amplitude Pin 2 ?? Pin 3 ?? Pin 4 RF drive (digital levels, around 10 V p-p)
Melles Griot 05-STP-91x surface mount (SMT) controller
Melles Griot 05-STP-91x (Syncrolase) Controller Surface Mount PCB shows closeups of one of the latest versions of the controller I have seen. It is clearly based on the older design with some parts simply being surface mount versions of the originals. But there have been changes as well. There is also at least 1 very minor variation on this layout, the only obvious difference being the substitution of a through-hole aluminum electrolytic capacitor for an SMT cap, which was perhaps not large enough. The PCB layout with the SMT cap typically has a through-hole cap soldered to the SMT pads. Reverse engineering the SMT PCB would be even more challenging than for the older one, and no, I'm not volunteering. :-) Note that even though Melles Griot no longer calls these "Syncrolase", the PCB arwork still use that designation!
There are 4 trim-pots on this PCB:
Trim-Pot Function Comments
---------------------------------------------------------------------------
P1 RF Frequency Controls power to induction heater
P3 S Mode Photodiode Gain This is for dual mode option, no effect
P4 P Mode Photodiode Gain Sets range of user set-point adjustment
P6 Output Level Set-Point User adjustment through hole in cover
As with the previous version, there are no temperature adjustments, but a resistor that appears to be selected for each unit is soldered into standoffs (visible at the top of the board in the photos, 6.8K ohms).
And here is what is known about the test-points:
Test-Point Function
--------------------------------------------------------
TP1 P mode amplitude
TP2 ??
TP3 Temperature sensor voltage
TP4 Output amplitude
TP5 S mode amplitude
TP6 RF drive (digital levels, around 10 V p-p)
I would suggest adding a separate temperature sensor used only for protection. The circuit could be as simple as a 10K NTC thermistor and fixed resistor or rheostat in a voltage divider, a zener diode and a 2N3904 or similar transistor in parallel with the one in the existing OT circuit. When the transistor turns on due to the resistance of the thermistor decreasing, it would shut down the heater drive. These parts would easily fit in the available space. There's even a spare hole in the coil form for an additional temperature sensor (at least in the ones I've seen). Since it's only for OT, the sensor can be further from the coil.
Interchanging the Output Adjust and Photodiode inputs to the Error Integrator (A1D) would cause the heater to turn off with no or low optical power. The only difference in functionality is that the laser would lock on the opposite side of the neon gain curve, equivalent to selecting the orthogonal polarization to the photodiode (by rotating the laser head 90 degrees).
Adding a second photodiode for dual polarization stabilization would also be beneficial since the relative intensity of the two modes would be the relevant variable, not the absolute intensity of a single mode. This ratio would still be valid at very low total output power.
Another modification (or complete redesign depending on your point of view!) that would enable the Syncrolase (or any thermally-stabilized laser) to run at the minimum temperature to assure reliable operation would to have a temperature set-point that is based on the ambient temperature of the environment, not a fixed setting. In principle, this can easily be accomplished by counting mode cycles from a cold start. Since each mode cycle represents a precise change in temperature, this would enable the laser to operate at a temperature of ambient plus a constant known to be greater than the heating from the laser tube current. A microcontroller could be used for the implementation, left as an exercise for the student. :)
This of course assumes that the ambient temperature remains relatively constant, but this is often the case with real lab environments. The Zygo metrology lasers with digital controllers compute the number of mode cycles (they call them "mode slews) needed to reach operating temperature based on the actual tube temperature when the laser is switched on, though they may still operate at the temperature required for worst case conditions.
While the Melles Griot version of the Syncrolase is not supposed to melt down due to a fault condition like not powering the laser, it is not known if the design has actually been fundamentally improved (as I've been told) or rather that they are simply depending on careful factory set-up and the reliability of the PCB and components. :) However, I have NOT seen any evidence of thermal damage to the Melles Griot units I've tested.
So, while two data points may not be conclusive, it would seem that that almost any tube that can be stabilized using the conventional heating blanket technique can also be stabilized using the Syncrolase controller if its mirror mount stem will fit inside and extend far enough into the induction heater coil. Where the tip-off is not too long but interferes with the coil assembly, simply removing the plastic cover may gain enough clearance. Of course, if you happen to be friendly with the tip-off person at a HeNe laser tube manufacturer, simply ask them to pinch-off and trim the tip-off closer to the tube! :) For longer higher power tubes, the internal preamp gain would need to be reduced to allow the Output Adjust pot to lock at higher power. Of course, for such tubes, the position on the gain curve over which the output is pure single mode would be reduced.
And flippers will work just fine, thank you. :-) And as noted above, 3 of 3 Aerotech tubes from Syncrolase lasers were flippers, at least when cold!
In fact, the existing optics are almost set up for dual mode stabilization. There are two beam-splitter cubes in the beam path. The first is a polarizing beam-splitter that diverts the unwanted S mode polarization 90 degrees to a beam block. The second one samples a small portion of the P mode output beam for the intensity stabilization feedback. If that beam block were removed, a photodiode could be positioned to sense the S mode.
Based on an examination of the latest locking adapter controller PCB and as my own tests, the hooks are already in place for dual mode (frequency) stabilization, but Melles Griot has either elected not to develop the product to completion, or at least has not yet released it. Even this sample of the Melles Griot 05-STP-910-536 Reference Laser for the Agilent 86122B Multi-Wavelength Meter with a manufacturing date of 2011 and the latest SMT control PCB, and redesigned head mounting and adapter case, still only uses a single mode for locking.
Note the location on the underside of the Melles Griot 05-STP-91x (Syncrolase) Controller Surface Mount PCB for a second photodiode (D5), currently un-populated. The lower (on the photo) pins of both photodiode are connected, but not the upper pins, so they are not simply in parallel. In addition the wiper on trim-pot P3 (which has no effect on normal operation) connects directly to the upper pad of the un-populated photodiode while the wiper on trim-pot P4 (P mode gain) connects to the upper pad on the installed photodiode (which provides the intensity feedback). So, P3 is used to adjust the S mode gain for dual mode stabilization. The only slight problem is that for unfathomable reasons, the position of the second PD doesn't quite line up with the S mode from the polarizing BSC. But moving it a few mm would fix that. The only other change to the optics might be a filter in front of the S mode PD to equalize the S and P mode sensitivities - the diverted S mode beam is much stronger than the sampled P mode beam since most of it exits out the front of the laser.
I have now confirmed that moving the mystery jumper near the edge of the PCB to the opposite position indeed turns on the dual mode option. It both enables light incident on a second PD installed at D5 to affect the lock point AND shifts the set-point of the Output Level trim-pot (P6) so that if P3 and P4 are adjusted properly, the locked output amplitude will be similar when switching between 1 mode (intensity, I) and 2 mode (frequency, F) operation. The first crude test was to use a flashlight to "adjust" the output level. ;-)
So, with some relatively minor surgery, it is possible to modify Melles Griot Syncrolasers with the SMT PCB for dual mode operation and true frequency stabilization! But rather than moving the first BSC, it may be easier to use a smaller photodiode - the existing one has a sensor area about 10 times what it needs to be even accounting for normal laser tube alignment tolerances - and simply install it with its pins bent to shift the sensor area by the required distance. There is a cover over the optics that needs to have a second hole drilled for the S mode PD but this is straightforward and would help to keep the PD in position.
And so it was done! After drilling the hole and squeezing a small photodiode (approximately 2x2 sensor area) from a barcode scanner in there, I may now have the only dual mode Syncrolase on the planet (or at least outside Melles Griot). With the jumper in the 2 mode position, both P3 and P4 have an effect and the lock set-point can be moved over a wide range. It's possible that there should really be a filter in front of the P mode PD as P3 is a rather sensitive adjustment, but the thing does work. And the Output Level trim-pot can still be used to move the lock position. (Though, strictly speaking, with these basic dual mode techniques, best frequency stability requires the P and S mode amplitudes to be equal on either side of the neon gain curve, and that occurs only at one setting.) Since its useful range can be set up to be approximately the same for both intensity and frequency stabilization, an external SPDT switch was added. See to access internal adjustments during normal operation as shown in Melles Griot 05-STP-91x (Syncrolase) Locking Adapter with Dual Mode Option Added. There is ample room to mount the switch in the cover but for this OEM version, there didn't seem to be much point. ;-).
Whether anyone really cares may be another matter. Most users don't have a clue about the difference between intensity (1 mode) and frequency (2 mode) stabilization. Since it's easier to measure output power, intensity stability is something most users can understand and easily measure. Checking frequency stability requires more sophistication such as the use of an ultra-high precision wavemeter or comparison with a frequency reference like an iodine stabilized HeNe laser. The novice user will observe that with frequency stabilization, the output power may change significantly until the tube has reached thermal equilibrium and assume that something is wrong. With intensity stabilization, the change is very small. After full warmup, frequency drift is quite small with either technique.
However, the competition generally implements both frequency and intensity stabilization, and additional expense is quite small, so why not do it! Hear that, Melles Griot? :) In fact, the S mode photodiode active area is way more than it needs to be even accounting for any possible differences in pointing alignment of the laser tube. Using smaller cheaper photodiodes should more than make up for the added parts cost.
Laser head specifications
Laser type: Helium-Neon (HeNe), single frequency.
Maximum output power: 1 mW.
Warm-up time: 15 minutes maximum.
Vacuum wavelength: 632.9907 nm
Wavelength accuracy: +/-0.1 ppm.
Wavelength stability +/-0.002 ppm/hour, +/-0.02 ppm/month.
Beam diameter: 5 mm.
Beam centerline spacing: 11 mm.
Safety classification: Class II.
Power requirements: 50 W at 100 to 240 VAC.
Output signals: λ/2 A-quad-B complimentary line driver
output, λ/2 A-quad-B complimentary
sinusoidal output (0 to +/-1.5 V peak)
"Laser on" and "Laser Ready" (TTL).
Dimensions (LxHxW): 15.38 x 5.50 x 4.04" (390,6 x 139,7 x 102,4 mm)
Weight: 12 lb (5,5 kg).
Operating Temperature: 15 to 40 °C.
Relative Humidity: 0 to 90% non-condensing.
Shock (IEC 68.2.27): 30G, 11 msec.
8 pin DIN
Pin Name Description ------------------------------------------------------------------ 1 SIN A A-quad-B SIN output signal (high active) 2 ~SIN A A-quad-B SIN output signal (low active) 3 COS A A-quad-B COS output signal (high active) 4 ~COS A A-quad-B COS output signal (low active) 5 GND Ground 6 VDC +5 VDC 7 Laser On Signal that corresponds to the "Laser On" LED 8 Stable Signal that corresponds to the "Laser Ready" LED
DB9
Pin Name Description ------------------------------------------------------------------ 1 SIN A-quad-B line driver output signals (high active) 2 ~SIN A-quad-B line driver output signals (low active) 3 COS A-quad-B line driver output signals (high active) 4 ~COS A-quad-B line driver output signals (low active) 5 GND Ground 6 VDC +5 VDC Output Signal 7 Laser On Signal that corresponds to the "Laser On" LED 8 Stable Signal that corresponds to the "Laser Ready" LED 9 GND Ground
The control PCB appears to have essentially the same circuitry as the standard S100 (with the same 5 trim-pots), but uses SMT devices for most ICs and discrete components. It also has the preamps for the optical receiver SIN, COS, and total amplitude channels. (There are 3 photodiodes.) There is also an LCD display with backup battery for a digital runtime meter. Interestingly, that meter appears to have a jumper-selectable option to only run when there is a return beam present to the optical receiver, but I can't confirm that. Perhaps the system could be rented on an hourly basis and only charged against actual use! ;-) A universal switchmode power supply provides 15 VDC for the HeNe laser power supply brick (10 to 15 VDC input, 1,500 to 2,000 V output, adjustable current) and controller PCB.
Several photos of the LZR2000 laser head can be found in the Laser Equipment Gallery (Version 4.06 or higher) under "Aerotech HeNe Lasers".
The LZR2000 I acquired was in excellent physical condition except that many of the screws were a bit rusted and required housekeeping services to clean or replace them before doing anything further! :) Whatever humidity was present appears to have had no effect on anything else. Its manufacturing date is 1996 though there is a second "system" sticker with a date of 1999. When first powered, the tube started sputtering after a couple minutes indicating that it probably was high mileage and the dropout current had increased. Can you believe it?! A slight increase in the tube current allowed it to run stably. The runtime meter backup battery in my sample is quite dead so I have no idea how many hours it has run. The raw output from the tube (before all optics) is almost 0.75 mW after warmup, so that's not so terrible, though probably somewhat less than when new. However, when powered up, it wasn't obvious if the controller was doing anything. The normal behavior for a standard Syncrolase is to immediately initiate a rapid ramp-up of temperature with the induction heater turned on full. With this laser, initial behavior was more like a cross between normal mode sweep and occasional hiccups or pauses. (I guess those glitches should have been a tip-off that something useful was happened!) However, allowing it to warmup for the required 15 minutes did result in the Ready LED coming on solid green. But locking did not occur and the modes then continued to come and go, gradually slowing down as though it was reaching thermal equilibrium. I assumed that either there was a controller hardware problem, or an adjustment was required, possibly due to the tube power being lower than normal. Optimistically assuming the latter, I figured that the only trim-pot that would be safe to adjust would be the one for the power output set-point. I didn't want a repeat of the core melt-down that I had with one of my S100s. Unfortunately, none of the trim-pots are labeled as to function. Although there are the same number of trim-pots associated with the locking circuitry, part numbers don't correspond to the ones on the original Syncrolase S100 PCB and even some of the part values have changed. But there was a clue: All but one of the trim-pots was very well sealed to prevent twiddling. Only the large blue multi-turn trim-pot facing forward had no Loctite™. Turning this definitely had an effect of changing the mode sweep rate dramatically at times. And, if turned too far CCW, the Ready LED turned *red*. This is not documented anywhere. It's supposed to be green or off! I assumed red meant that something really bad was about to take place. However, after some random twiddling, a remarkable thing happened. :) The laser settled down and locked, with the adjustment then having the expected effect of varying the output power. I assume the output level was simply set much too high for this tube that had weakened from long hours slaving at whatever it was doing. Thus, once the tube reached operating temperature, the controller was frantically searching for the lock point, which was impossible to achieve. Now it's set at around 350 µW, which is plenty of power for a metrology laser (especially one that will likely never be used in a real application ever again!). The laser now locks without issues from a cold start, though it sometimes requires much more than the spec'd 15 minutes. This tube's mode sweep profile has a nearly flat top and rather steep sides as shown in Mode Sweep of Aerotech LZR2000 Stabilized HeNe Laser. The top plot shows how the output would look if the laser were allowed to warm up in the normal manner which the bottom plot is of an actual run from a cold start to beyond where locking occurs. To guarantee stability with the single mode intensity stabilization technique, the lock point must be safely on one of the side regions. As such, it would probably lock reliably at well over 400 µW as the peak output power is almost 500 µW. But note the peculiar behavior in several places in the bottom plot. At several locations startup to just beyond the halfway point, there are multiple mode flips where the power in the two modes would be almost precisely equal (if the other one were shown). You can tell I'm not totally surprised, having seen flipper tubes in Aerotech Syncrolase lasers in the past, but it's still an aesthetic problem. ;-)
A pair of anomalies are present once the laser locks. Whether these are problems (or features) associated with this specific laser, I do not know. But how can there be anything wrong with it? :) The following is from one run recorded at 60 samples per second with a data acquisition system:
I do not have the slightest intention of investigating these much further. :)
Some of these lasers are badged Tropel, which is the company that originally developed them. Coherent owned Tropel from 1972 to 1982. According to a former Tropel engineer who was involved with the optics design, the model 200 may have been the first commercial laser to use dual polarized-mode stabilization based the paper: R. Balhorn, H. Kunzmann, F. Lebowsky, "Frequency Stabilization of Internal-Mirror Helium-Neon Lasers", Appl. Opt. 11, 742 (1972).
The HeNe laser head is powered from a standard Laser Drive 6.5 mA, 2,100 V power supply brick via a HV BNC connector. There is no special control or regulation of this supply - it's turned on by the main power switch. But some thoughtful engineer included a high resistance bleeder to discharge the HV caps in the power supply brick after power is removed. :)
The HeNe laser tube itself is a Melles Griot (not made by Coherent!) model, labeled 05-LHR-219-158. It has similar dimemsions to an 05-LHR-120, a common 2 mW (rated) random polarized laser. But, the -158 may mean it has been specially selected to have a well behaved mode sweep cycle (not a flipper!) for this application. It may also be filled with isotopically pure (or at least enriched) gases and an AR-coated HR (to minimize back-reflections from the HR's outer surface). The tube itself puts out more than 2 mW when new - possibly up to 4 mW or even more - but the polarizing and beam sampling optics sucks up some of it. In addition, depending on the particular version, there is either a dielectric filter or polarizing filter in the end-cap. The dielectric filter cuts the output by about half but the this can be varied by 10 percent or so (though I'm not sure if this is intentional or just a byproduct of it being angled). The polarizing filter allows continuous adjustment of output power. (In both cases, the adjustment is done by loosening a set-screw and rotating the end-cap). According to the CDRH sticker, the output beam is supposed to be less than 1 mW. Given the wide swings in output power during warmup (see below), even with 50 percent attenuation, the peak output power may approach 1 mW. But regardless of the type of end-cap, only a single polarization ever exits the laser since the internal beam sampler blocks the other one.
There is a thin film heater attached to a thick rubber jacket between the tube and laser head cylinder. A beam sampler assembly consists of a pair of Beam-Splitter Cubes (BSCs) in series and two photodiodes, each associated with one of the BSCs. The first BSC is a polarizing beam-splitter and reflects the full power of one polarized mode to its photodiode. Thus, the beam that passes through it is linearly polarized with the orthogonal orientation. The second BSC reflects 10 or 20 percent of this mode to its photodiode. So, the output beam from the laser is pure linearly polarized and has slightly less output power than one of the polarized modes of the tube. The controller monitors the lasing modes and maintain cavity length using the heater so that a pair of orthogonally polarized longitudinal modes straddle the gain curve. The beam sensor assembly can be rotated to align the photosensors with the 2 orthogonal lasing modes as this is arbitrary from tube to tube, and orientation within the cylinder, but should remain fixed for the life of the tube.
The controller can be set up to run on various input voltages from 100 VAC to 240 VAC by changing the position of a small PCB that plugs into the AC entrance assembly, and plugging in the appropriate fuse. However, it seems that the HeNe laser power supply always runs on 115 VAC from a tap on the main power transformer so it doesn't need to be capable of 230 VAC operation, even though the one that's in there has that option - the wire for 230 VAC is not used! The output of the HeNe laser power supply is rated 2,100 V at 6.5 mA with no start delay.
The user controls consist of one (1) power switch. There are indicators for AC power and Status. After a warmup period of 20 minutes or so for the laser head to reach operating temperature, the Status indicator will change from Wait (red) to Ready (green). Doing anything that causes lock to be lost will result in a shorter delay of a couple minutes to re-establish it.
The internal circuitry of the controller box is relatively simple and includes a pair of LM3403 quad op-amps, a 741 op-amp, and LM311 voltage comparator, along with a TO5 power transistor on a heatsink to drive the heater.
Here is the pinout of the circular control connector as determined by my measurements. There may be errors.
Pins Wire Color Function Comments -------------------------------------------------------------------------- 1,2 Blk/Wht Heater Power ~22 ohms 3,4 Blk/Red Temp Senseor ~880 ohms at 25 °C, ~1.2K when locked 5,6 Blk/Blu Photodiode 1 Anode is pin 5; Approximately 250 uA max 7,8 Blk/Grn Photodiode 2 Anode is pin 8; Approximately 50 uA max
It would appear that the difference in sensitivities is the way it's supposed to be since this was similar on 3 heads. (However, the readings on an analog VOM for the photodiodes did differ on 2 heads I tested - I'm not sure what, if any significance, that has.) This makes sense given that the sampling is done from the main beam. One polarization orientation is blocked entirely and thus the associated photodiode gets its full intensity. The other mode would then seem to be sampled at about 20 percent intensity. The controller and laser head are normally a matched pair and there is an adjustment inside the controller to equalize the responses.
The heater consists of a serpentine thin file metal pattern on a rubbery backing material that wraps completely once around the tube.
The temperature sensor extends the length of the tube and is buried within the heater backing, technology unknown.
I picked up a controller and 3 laser heads in two separate eBay auctions for a grand total of $22.50 + shipping. The serial number on one of the heads matched that of the controller and while this head was initially hard to start, after running it for awhile on my HeNe laser test supply, it now starts normally.
The controller originally had a dead HeNe laser power supply brick (Laser Drive 314S-2100-6.5-2, 2,100 V at 6.5 mA) which is likely the reason it was taken out of service. I replaced that with an Aerotech LSS-5(6.5) which seems to be happy enough. Using a laser power meter, one of the two modes of the laser (the one present in the output beam) could be seen cycling up and down between about 0.60 and 1.40 mW with the orientation of the beam sensor assembly adjusted for maximum peak power. Each cycle took longer and longer as the tube warmed up to operating temperature, helped along by the heater. After about 15 minutes, it would appear to try to "catch" at certain power levels but couldn't quite remain there. (This behavior may have had nothing to do with the feedback control though.) Then suddenly, after about 20 minutes, the Ready light came on and a few seconds later, it locked rock stable at 0.95 mW. :) A second laser head behaved in a similar manner but with a slightly higher final output power of 1.02 mW. No adjustments were needed inside the controller despite the fact that the second head's serial number didn't match the controller's serial number. Possibly, even better stability or slightly higher stabilized output power could be achieved with some fine tuning. (The 1.02 mW head actually had higher peak power than the 0.95 mW head. The difference is probably in part due to the photodiode sensitivities.) With the fixed filter end-caps installed, the output power dropped to around 0.50 mW. I rather suspect that these are normal power levels for this system. (This was later confirmed when a manual with detailed specifications turned up.) The third head had its cables cut but I finally scrounged a replacement control connector from a box of junk in the garage and jerry-rigged the HV BNC for testing. That laser head now works as well. It also came with an adjustable polarizer in its end-cap. With that installed on either of the other heads, the output power could be varied continuously from near 0 mW to about 1 mW.
Note that the Ready light comes on and then the laser locks in at the proper phase of the next mode cycle. So, basically the pea brain in the controller (no actual CPU of any kind!) decides that conditions are suitable and enables the feedback loop. The final "decision" is based the cycle duration being longer than some magic number (around 1 minute). :) I've also seen the ready light come on even if the laser doesn't start and when one of the previously locked heads was plugged back in after a few minutes of cooling. In the latter case, the laser was indeed locked though it might not have been able to maintain it continuously since the tube was probably no longer really warm enough.
There are actually two feedback loops in the controller. During warmup, the heater is driven to a fixed temperature based on the resistance between pins 3 and 4 of the Control connector. Once the period of the mode cycle exceeds a fixed time (guessing somewhere around 60 seconds), the control loop based on the difference of the photodiode outputs is enabled. The same signal that switches from the temperature feedback to mode feedback turns the Wait indicator goes off and the Ready indicator on. More on this in the next section.
Plot of Coherent Model 200 Stabilized HeNe Laser Head During Warmup and Plot of Coherent Model 200 Stabilized HeNe Laser Head Near End of Warmup show the output power variation due to mode cycling. Note how it seems to "snap" into regulation once the time is right. :) There are roughly 90 mode cycles during warmup prior to lock. The internal optics account for the large variation in output power. The HeNe laser tube itself has a normal mode sweep of only a few percent.
Another Coherent 200 system I have has a fully functional controller but a fully dead laser head. It is very hard start, impossible to run, and way beyond end-of-life. So, that gave me an excuse to go inside.
The Coherent 200 laser head can be disassembled in a reversible manner with fewer individual parts than the Spectra-Physics 117/A or the essentially identical Melles Griot 05-STP-901. However, it doesn't come apart as easily, using a press-fit for the tube/heater sandwich.
As noted above, the tube was found to be way beyond end-of-life. If it could be convinced to start (on a lab power supply), it would not run at any reasonable current and produced no output at all. There was sputtered aluminum coating on the holes near the cathode end-cap and even through holes in the cathode can near the center of the tube. This system had obviously been left on continuously for a large number of years. It was probably not even in use for a good portion of that time, forgotten and lonely in a corner of a lab, wasting its life producing coherent stabilized photons no one was using until there were no more! :) That seems to be the destiny of so many stabilized HeNe lasers. I'll be searching for a suitable replacement tube. The original tube, a 05-LHR-219 (with or without a -158), doesn't show up in any list I've seen) but an 05-LHR-120 has nearly the same dimensions and will run on the same power supply. So, as long as one can be found that is well behaved (non-flipper, wedged HR), it will almost certainly work fine. Other random polarized laser tubes of similar length can also be adapted but may require replacing the HeNe laser power supply and coming up with a creative mounting scheme if diameter is smaller.
An operation manual and application notes for the Coherent 200 can be found at Ajax Electronics Laser/Optics Manuals under "Coherent".
Everything is in Schematic of Coherent Model 200 Stabilized HeNe Laser. Note that most of the part numbering is totally arbitrary as there were *no* part numbers on the PCB except for the PCB connectors (and I only have J2 in the drawing). This is a late revision with PCB artwork dated 1997, though that probably only means that there was a PCB fab run in 1997, since the artwork itself was obviously hand taped. :) I guess some important customer just had to have more of these lasers made well after they would have been considered very obsolete by Coherent. :)
The controller has two feedback loops. The Preheat Loop, which is active while the tube is warming up, drives the heater in the laser head to a fixed temperature (set by a pot). The temperature sensor in the laser head is not a common NTC thermistor, but something that increases in value with increasing temperature. It has a resistance of around 800 to 900 ohms at room temperature, but well over 1K ohms at operating temperature. The preheat loop prevents the mode feedback loop from going active until the temperature is sufficiently high. Only after this occurs, does a timer begin to look at mode changes, and switches from the preheat loop to the mode feedback loop once their period exceeds around 60 seconds. The mode feedback loop uses the difference between the orthogonally polarized A and B modes in a simple PI control loop to drive the heater. Should the laser not stabilize as evidenced by mode changes still occurring, the preheat loop will be switched back on to try again. At least, that seems to be how it's supposed to work. However, a system with a laser tube that doesn't start (or a bad HeNe laser power supply) will likely turn on READY shortly after being powered up even though it is obviously not working correctly. Well, I guess it IS quite stable - dead with a frequency of exactly 0.0000000000 Hz and an output power of exactly 0.0000000000 mW! :)
Note: Some versions of the controller PCB lack the Temperature and Loop Gain pots (R1 and R20, respectively). I'm not surprised about the absence of R20 as it never seemed to do anything useful, but the lack of R1 either means the temperature sensor resistance is fairly consistent from one laser head to the next and a fixed value of R19 could be used, or that R19 was hand selected for each laser head.
Here is the adjustment procedure. A multimeter (preferably an analog VOM, with a needle!) or oscilloscope is required. A 14 pin "DIP Clip" will come in handy, and a laser power meter and temperature probe are desirable but not essential. A hex wrench to set the output polarizer orientation and small flat blade screwdriver to adjust the pots will also be needed.
This should be done from a cold start at an ambient temperature close to where the laser will typically be used. If the laser had been on, it should be turned off and allowed to cool down for a half hour minimum before proceeding.
A printout of the Schematic of Coherent Model 200 Stabilized HeNe Laser will come in handy.
Preparation
Mode A and B adjustment
The balance between the two polarized modes will affect the location of the lasing line on the neon gain curve. The following sets the two mode amplitudes to be equal, which places the modes equidistant on either side of the gain curve. However, it should be possible to offset the modes if desired, if a different location or slightly more output power in Mode A (the output beam) is desired. However, it's not possible to place either mode precisely at the top of the gain curve.
Note: If adjustment of the beam sampler was necessary (or just to double check that it was set correctly), testing with a Scanning Fabry-Perot Interferometer (SFPI) would be desirable. This would allow the undesired Mode B to be virtually totally suppressed. Simply maximizing the Mode A amplitude is not nearly as precise but with care, getting to less than 1 percent of Mode B should be possible. The adjustment using the SFPI can be done at any time, even during warmup, though it's easier once the laser has locked and nothing is changing.
Temperature adjustment
The HeNe laser tube and ballast resistors dissipate almost 12 W (1.8 kV at 6.5 mA). The temperature set-point must be selected such that it is slightly above what would result from the tube and ballast power alone. At an ambient temperature of 18 °C, the required temperature set-point ends up being around 40 °C, a difference of 22 °C. I do not know exactly how this is affected by a change in ambient temperature. If the difference remains constant, the head must run at 62 °C for the maximum allowable operating temperature of 40 °C (from the specifications in the Coherent manual). Such a high operating temperature seems unrealistic.
One way to estimate the value for the temperature set-point is power only the laser HeNe laser tube (not the heater) by disconnecting the Control cable and allow it to reach thermal equilibrium (at least 1/2 hour). Measure its temperature and then reconnect the Control cable and adjust the Temperature set-point to be about 5 °C higher, or so that the mode sweep goes through an additional 15 full cycles.
The following assumes an ambient temperature of 18 °C:
Note that the range of 5 to 10 V is my estimate. The Coherent manual shows a graph with the voltage at 12 V at the time of lock (which would then likely drop down to under 10 V after thermal equilibrium). But there is no description or indication of what ambient temperature was used. Perhaps some key piece of information is missing. While there's no problem adjusting the temperature so the laser locks and is stable at any given ambient temperature or a reasonable range around it like +/-5 °C, I don't see any practical way the laser could be set up to operate over the entire 0 to 40 °C range spec'd in the manual without running excessively hot, especially under typical conditions (below 25 °C). It would make more sense if R2 was a sensor for ambient temperature so that the temperature set-point was an offset from ambient rather than actual temperature, but R2 looks like an ordinary resistor.
If the laser will be used in an environment where the ambient temperature is much different than where it was tested, readjustment may be needed. The official Coherent Adjustment Procedure (CAP) probably sets the temperature so high that this would not be required over the full spec'd temperature range of 0 to 40 °C, but that shouldn't be necessary unless the laser is to be used near in a sauna. :)
Mode feedback gain adjustment
Finally, power off for 1/2 hour and confirm that the laser will then stabilize properly (after the warmup period) when powered back on.
One other thing that's recommended while the case is opened is to check R39 and R40, the third and forth resistors from the right in the first row at the front of the PCB. These are the current limiting resistors for the Wait and Ready indicators, respectively, and were originally 510 and 1K ohms, both apparently 1/4 W (by size and appearance). There are other current limiting resistors inside the indicator packages, but the voltage across R39 and R40 may still be high enough to greatly exceed the 1/4 W ratings of the original resistors. If so, the PCB will probably be darkened beneath them as well. Measure the voltage across R39 and R40 when their respective indicator is lit. If either is more than 20 V and the resistor is only 1/4 W, replacement is highly desirable, especially for R40 which will be stressed possibly for years on end. :) Suitable values are 1K, at least 1/2 W for both. Yes, Ready won't be quite as bright but it will be much happier! Proper replacement will require removing the PCB but this is just five screws and several connectors. Space the new resistors off the PCB a bit to further aid in cooling. The PCB is easily damaged, so use a proper desoldering tool to remove the old resistors and clean up the holes. Or just cut the leads off at the bodies of the old resistors and solder to those.
The first step in tube replacement is to find a suitable tube. Melles Griot probably won't even sell you a tube, and if they did, it would cost $300 to $400! Although a common type, this seems to be harder to find surplus than it would appear. Most of those that turn up on eBay seem to be the 05-LHP-120 - the polarized version - which is useless for this purpose. Once a suitable candidate tube has been found, it needs to be tested for non-flipper behavior. A tube that is a flipper may still be useful if the flipping is consistent, or if it disappears when the tube warms up, but a totally well behaved non-flipper is most desirable.
CO-200 laser head disassembly:
CO-200 laser head reassembly:
I found an old 05-LHR-121 laser head with a good tube, extracted the tube, and spent way too much time installing it in a CO-200 laser head that had a nearly dead tube. This included a liberal application of duct tape and bailing wire. :) But it works. I knew that this particular tube was a flipper and expected to simply pick the proper mode polarity such that it would lock on the opposite side of the gain curve from the one that flipped. However, it turned out to only flip until it warms up for about 4 minutes or 113 half-mode cycles, then abruptly it stops flipping and becomes well behaved. I have an Aerotech tube with similar behavior, cause unknown.
I don't think this is in what might be called original condition, but it does start right up without problems (no hard-start tube!) and has decent power (3.2 mW or more total from the tube). It locks normally with 1.2+ mW in a single mode.
All in all though, much more effort is required to do a tube replacement on the CO-200 than the SP-117/A.
The controller that went along with this laser head also had minor problems. I had to replace the usual toated dropping resistor for the READY LED but also had to totally rebuild the READY LED assembly itself - both LEDs and their current limiting resistors were fried to a crisp. :( :)
The pieces of the system I acquired consist of a CO-200 laser mounted in a semi-enclosed box connected to an interferometer block via am armored single mode fiber-optic cable, and three plastic multi-mode (light pipe) cables for sensing. Based on its design, my assumption is that there is a remote free-space "too" or other device with a cube-corner (retroreflector) on it that can move or change in some way.
The CO-200 appears to be totally standard with a bezel having C-mount threads to which the adjustable fiber port for the input fiber is attached. This fiber may be "polarization maintaining" or simply single mode but there is a polarizer attached to its output end. Since the CO-200 outputs a pure single mode, even If it is not polarization maintaining, its output would still only be a single mode even if not polarization maintaining, but the amplitude would vary.
The interesting part is the interferometer. See Diagram of Strange Fiber-Coupled Interferometer. The optics components within the solid line are all inside a nicely made solid aluminum block. I do not know if this is a totally custom assembly. From appearances, it may not be. The top pair of beams forms a standard linear (Michelson) interferometer. The output of the "White Dot" fiber could be used to sense displacement in the normal way, by counting fringes or portions thereof. However, although I'm assuming that the "Remote Tool" simply uses a retro-reflector, that's just speculation. For the polarized "Green Dot" and "Red Dot" outputs to show anything interesting, it would seem that polarization changes are being sensed somehow. (Their polarizers are redundant as the PBS has separated the polarizations.)
The laser in the system I have is in very good health except that the polarization beam splitter at its output (inside the laser head cylinder) had become delaminated and was not passing *any* light! So, the laser appeared dead. But replacing this with one from another CO-200 resulted in like new performance, How does a prism come apart on its own?
If anyone has more information on this system, please contact me via the Sci.Electronics.Repair FAQ Email Links Page.
Excel had only a single type of laser, the 1001, a Zeeman-split HeNe laser with a split/REF frequency between 1.5 and 3.0 MHz. But there are at least 2 different case styles. The 1001A and 1001F are about the same size as the smaller HP/Agilent lasers with a similar mounting arrangement, and have connectors and signals compatible with the 5501B and 5517, respectively. The 1001B is almost as large as the HP/Agilent 5517A (though its shape is more normal) with a similar mounting arrangement, and its connector and signals are compatible. The specifications for the 1001B appear to be the same as those of the 1001F and internally, it's virtually identical to the 1001F but with wasted space at the front. However, an Excel 1001 laser would be a drop-in replacement for an HP/Agilent laser only if selected for split frequency: 1.5 to 2.0 MHz for the 5501B (1001A) and 5517A (1001B), 1.9 to 2.4 MHz for the 5517B (1001F), or 2.4 to 3.0 MHz for the 5517C (1001F). The default beam size for the 1001s is 5 mm, versus 6 mm for the HP/Agilent lasers, but this is probably of little consequence in most applications.
Although the Excel model 1100B 6DOF Calibration System is based on the 1001 laser technology, it may use only the internal components of the 1001 laser packaged along with additional optics and electronics in a single enclosure.
Here are the specifications for the Excel 1001A/F lasers (mostly from the file linked above). The difference(s), if any, between the 1001A and 1001F are probably only in the connectors (5501B for the 1001A and 5517A for the 1001F) and F1/F2 orientation (rotated 90 degrees). While I haven't seen full specifications for the 1001B, the connector is 5517-compatible and everything below is probably the same except for the case size and possibly for the nominal wavelength, which is listed as 632.99136 nm on the back of one sample (though I doubt the actual wavelength of the laser is any different!). And what's inside the 1001B is essentially identical to what's inside the 1001F!
Here are some observations/comments that apply to all the Excel 1001 lasers unless otherwise noted:
ID Function
---------------------------------------
J1 Polarization mode photodiodes
J2 Tube heater
J3 Not installed, ????
J4 HeNe laser power supply
J5 Input power
J6 Backpanel LEDs
The next three sections have more details on the 1001F, 1001A, and 1001B lasers.
The output power of my 1001F was approximately 270µW, which is a bit low compared to the original value of 335 µW) printed on the backplate but well above the spec'd minimum of 200 µW. And the output power doesn't change by more than a few percent after extended warmup indicating that the tube is relatively healthy - no gas contamination and not end-of-life. The laser was still fully functional with a REF frequency of around 2.22 MHz. After aligning the OC mirror, the output power went to 375 µW and REF to 1.98 MHz. So, it's now equivalent to a 5517B. (Specs for the 5517B: Minimum output power of 180 µW, REF frequency of 1.9 to 2.4 MHz.) However, I have no way of knowing what output power and REF were when new. It might have been deliberately detuned to achieve a higher REF at the expense of output power.
For a summary of the specifications for the 1001F laser, see the previous section.
Several photos of the 1001F laser can be found in the Laser Equipment Gallery (Version 3.00 or higher) under "Excel Precision HeNe Lasers".
The two most interesting ones are:
The unit I acquired had "junk" scribbled in Magic Marker on the case, but appears to work quite well with an output power of over 410 µW (label value is 540 µW) and REF frequency of 1.85 MHz - perfect for a 5501B clone. As with the other Excel lasers, warmup isn't as rapid as with the HP/Agilent lasers, and is constant heating followed by abrupt locking. Interestingly, the date on the backplate - which appears to be original - is 2006. This is rather peculiar given that the Excel Web site hasn't been updated since 1998 and no recent references to the Excel company can be found. :) However, the date code on a tantalum capacitor on the control PCB - the only component with a date code visible - was 1996, which makes more sense. So perhaps it was serviced by a former Excel engineer in 2006.
The output power of my 1001B was approximately 120 µW after full warmup, actually declining from 150 µW just after locking. This is low compared to the original value of 240 µW) printed on the backplate and the spec'd minimum of 200 µW. Other than the low output power, the laser was still fully functional with a REF frequency of around 2.60 MHz. After aligning the OC mirror, the output power increased to 315 µW and REF declined to 1.88 MHz after full warmup. The output power changed only slightly, from 330 µW just after locking. This indicates either a change in alignment or drift in the electronics, not a tired tube. So, it's equivalent to a 5517A. (Specs for the 5517A: Minimum output power of 180 µW, REF frequency of 1.5 to 2.0 MHz.) However, I have no way of knowing what output power and REF were when new. It might have been deliberately detuned to achieve a higher REF at the expense of output power.
For a summary of the specifications for the 1001B laser, see the previous previous section.
Several photos of the 1001B laser can be found in the Laser Equipment Gallery (Version 3.19 or higher) under "Excel Precision HeNe Lasers".
Other than the difference in case size, everything else is virtually identical compared to the 1001F laser.
Excel 1001A reference signal connector
See HP 5501 Reference and Power Rear Panel Connectors for pin location.
Pin Function Socket View
--------------------------------------------- A
A* Accessory +15 VDC fused o
B* +15 VDC return D o o B
C Reference (difference) frequency o
D Complement of pin C C
Excel 1001A power connector
See HP 5501A and 5501B Reference and Power Rear Panel Connectors for pin location.
Pin Function Socket View
---------------------------------------
A +15 VDC input D o o A
B -15 VDC input
C* +5 VDC output (test-point) C o o B
D Power ground
Pins denoted by "*" have these assignments on the HP-5501A/B but they have not been confirmed for the Excel 1001A.
Excel 1001B/F Power and Reference Signal connector
The Excel Power and Reference Cable for the 1001B/F is similar to the HP/Agilent 10791 and the two may be used interchangeably. (This is probably Excel part number 1059A but that hasn't been confirmed.) "Wire Color" is that of the power connections with ring lugs. See HP/Agilent 5517 Laser Rear Panel Connector for the physical arrangement of the pins:
Wire
Pin Color Function
---------------------------------------------------------------------
A* NC ( MEAS signal level on 5508A)
B* NC (~MEAS on 5518A only)
C* NC ( MEAS " "
D* NC (Signal Return for MEAS)
E ~REF (Zeeman beat signal from internal optical
F REF receiver's differential line driver.)
G Black Ground
H Green Ground
J Orange +15 VDC
K Red +15 VDC
L White NC (-15 VDC on HP/Agilent cable)
M +15 VDC
N,P* NC (Cable Shield on HP/Agilent cable)
R Signal Return for REF
S Ground (Optical Receiver)
T +15 VDC (OPtical Receiver)
U* NC (Cable Shield on HP/Agilent cable)
* Connections to pins A,B,C,D,N,P,U are not present on the 1001B/F cable. The wire for Pin L (-15 VDC) is present, so this cable might work with an HP/Agilent 5517 laser but I haven't confirmed that. There is also a blue wire in the cable but it is cut off and hidden under heat-shrink and does not have continuity to the connector.
Having said all that, the Model 100 iodine stabilized laser consists of a massive laser head and separate controller using custom modules built into a Tektronix 5000-series storage scope mainframe. This probably dates the development of this system to the 1970s or early 1980s. While not fully automatic, the controller provides a straightforward way of selecting one of 7 possible I2 absorption peaks and then locking to it. It's based very closely on the original NIST ISHL design first presented in the paper: Howard P. Layer, "A Portable Iodine Stabilized Helium-Neon Laser, "IEEE Trans. on Inst. and Meas, IM-29, pp358-361, 1980. A photo of the same or very similar laser can be found at NIST: Length - Evolution from Measurement Standard to a Fundamental Constant. The slightly shorter resonator shown in the paper uses a different two-Brewster HeNe laser tube than in the two samples I'ave seen (manufacturer unknown) and the scope plug-ins are labeled "National Bureau of Standards" instead of "Frazier Precision Instrument", but everything else appears to be identical. The Frazier controller is in a rack-mount configuration, but it can be converted to the portable upright instrument shown in the paper with a screw-driver and open-end wrench. :)
Specifications of the Model 100 Iodine Stabilized Laser (mostly from the Frazier Web site with interpretation):
That really low power output of 100 µW was bothering me as it seemed as though much more was possible even with the extra I2 cell Brewster windows from a tube capable of at least 1 mW. But the reason is far more fundamental than unavoidable losses. The problem is that for a resonator length of over 30 cm, multiple longitudinal modes would oscillate unless something were done to force single mode operation. The NIST and presumably Frazier ISHLs do this with a really low OC reflectivity of 93 percent, which raises the lasing threshold but also dramatically decreases output power. There would appear to be no fundamental reason why a normal 99% OC could not be used with the addition of an intra-cavity temperature-controlled etalon to select a single longitudinal mode. This might add some complexity to the controller though.
(Other ISHLs have similar anemic specifications for output power. Where higher power is desired, an "offset locked" single frequency HeNe laser is generally added to the system. The difference frequency between the two lasers is phase-locked to a crystal reference, thus retaining essentially the same frequency stability as the I2 laser, but the output power can be 1 to 2 mW as with common stabilized HeNe lasers.)
Iodine Stabilized HeNe Laser Head is a photo of what is almost certainly a Frazier 100. Although there is no manufacturer label, everything is identical down to the pattern of ventilation holes in the cover. The overall appearance is unremarkable with a shutter at the front (the round black thing) and several cables coming out the back (hidden). Leveling "feet" would often be installed be installed in the cast tabs for precise alignment. Although there are no manufacturer labels, this is almost certainly a Frazier 100 laser. But there was an Agilent inventory sticker on the cover, so perhaps this very laser was used to certify HP/Agilent metrology lasers like the 5517A! :) In fact, the base of this laser bears a striking resemblence to the 5517A (though the dimensions don't match). It's a combination of a cast and machined assembly, clearly not made for a one-time research project.
Iodine Stabilized HeNe Laser Head With Cover Removed shows the interior. The glow of the Melles Griot 05-LHB-290 two-Brewster HeNe laser tube can be seen along with the iodine cell within the massive 4 bar Invar resonator structure.
Frazier Model 100 Iodine Stabilized HeNe Laser System shows the laser head on top of the controller. At present it doesn't have a working tube (though as can be seen, the tube does light up), and except for the HeNe laser power supply, the connectors on my old laser head do not mate with the controller.
Frazier Model 100 Iodine Stabilized HeNe Laser Controller shows the front panel. The connectors on the controller are labeled as follows:
The three modules are all discrete circuitry with mostly Burr-Brown parts.
While the Model 100 controller doesn't have fancy auto-magical locking firmware, operation seems relatively straightforward. (This was before the era of cheap silicon!) Paraphrasing from the paper, the front panels of the scope plug-ins have a diagram showing the hyperfine I2 components available at 633 nm along with their precise vacuum wavelengths. In sweep mode, the third derivative of the laser output power is displayed on the (storage) scope screen as the cavity length - and thus laser wavelength - is scanned with a period of 1 second. The desired component can then be centered and expanded using the Sweep (span) and Bias (offset) controls, at which point the system is switched to lock mode. The complete operation summary is printed on the three plug-ins. Who needs the @!%$# manual? ;-)
I wonder how many of these were ever made. My controller appears to be SN 15 and had a DoD (Department of Defense) inventory sticker. It's in mint condition.
General info on iodine stabilized HeNe lasers along with additional photos can be found in the section: Iodine Stabilized HeNe Lasers.
More to come.
The general approach to precision measurement used by all systems based on two-frequency HeNe lasers such as those from HP/Agilent is shown in Interferometer Using Two Frequency HeNe Laser. The capabilities are quite impressive. A typical example is the HP-5501B laser head from the HP-5501A Laser Interferometry Measurement System, which enables a position/distance resolution down to better than 10 nm (that's nanometer as in 0.000000001 meter!). And that's one of the earliest implementations. More information on interferometers based on two frequency lasers including descriptions of the optical components can be found in the section: Interferometers Using Two Frequency Lasers. What follows relates mainly to the laser technology.
Here is a comparison of most of the HP two frequency metrology laser models:
(6)
(4,5) Reference Maximum Beam
Model Case Tuning Frequency Velocity Diam. Comments
-------------------------------------------------------------------------------
5500A Huge :) PZT 1.5-2.0 MHz 0.4 m/s 6 mm (1)
5500B " " " " " " " 6 mm (1)
5500C " " " " " " " 6,9 mm (2)
5501A Small " " " " " " " (3)
5501B " " Thermal " " " " " " (3)
5517A Large " " " " " 6 mm
5517B Small " 1.9-2.4 MHz 0.5 m/s "
5517BL " " " " " " " "
5517C " " " 2.4-3.0 Mhz 0.711 m/s 6,3,9 mm
5517D " " " 3.4-4.0 MHz 1.0 m/s 6,9 mm
5517DL " " " >4.4 MHz 1.3 m/s " "
5517E " " " >5.8 MHz 1.77 m/s 6 mm (8)
5517EL " " " " " " m/s " " (8)
5517F " " " >7.0 MHz 2.15 m/s 6,9 mm (8)
5517FL " " " " " " m/s " " (8)
5517G " " " >7.2 MHz 2.2 m/s 9 mm (8)
5517GL " " " " " " m/s " " (8)
5518A Large " 1.5-2.1 MHz 0.4 m/s 6 mm S/N below 2532A02139 (2)
" " " " " 1.7-2.4 MHz 0.453 m/s " S/N 2532A02139, above (2)
5519A " " " 2.4-3.0 MHz 0.7 m/s " (2)
5519B " " " 3.4-4.0 MHz 1.0 m/s " (2)
Of the lasers listed, as of Winter 2012, only the 5517A/B/C/D (and their variants) and the 5519A/B are listed on the Agilent Web site as still being in general production and "orderable". The 5517E/F/G (and their variants) may be limited editions for special applications or secret Government projects or to satisfy the fantasies of laser jocks :) since there is virtually no information about them anywhere. There are also variations such as higher power or higher REF/split frequency for the above lasers depending on options.
Notes:
The 5518A and 5519A/B have a single optical receiver built-in. And of all the HP metrology lasers, the 5519s are unique is having a built-in DC power supply so they simply plug into the wall and feed their REF and MEAS signals to the associated measurement processor/display.
Like the 5500C, the 5518A or 5519A/B can be used in the normal way (e.g., in a 5528A Laser Measurement System), but are generally intended to be set up stand-alone without any additional optical receivers in a 5529A or 5530 Dynamic Calibrator). For example, the 5519A laser head can be mounted on a cart and aimed through interferometer optics at a cube-corner (retro-reflector) or plane mirror on a tool whose motion needs to be measured precisely.
The polarizing beam-splitter that detects the modes is deliberately oriented so that the separation isn't perfect and a small amount both F1 and F2 are present in each. This results in a beat frequency being generated which is used to produce the reference signal (REF) and to confirm that there is enough beam power to be usable.
Although locking typically occurs in 4 minutes (READY comes on solid), some lasers (perhaps the 5517F and 5517G) may require 20 minutes. My 5517E takes about 9 or 10 minutes. But a fully stable frequency output requires 90 minutes for lasers with a non-vented cover or a vented cover but no fan. Those with a vented cover and a fan require only 45 minutes. (It's not known if the temperature set-point for the tube is lower for these compared to the non-vented variety. But if so, this would explain the "Low heat" option since power dissipation would be reduced by running at a lower temperature.)
From my observations, the frequency oscillates slightly immediately after locking with a period of order of minutes. The amplitude of these oscillations gradually decreases with time and eventually becomes very small. However, the laser likely still meets accuracy specifications during this time.
An internal optical receiver samples part of the output beam and is used to generate the references signal and to confirm that there is enough beam power to be usable.
The 5501B is the only laser to use Pulse Width Modulation (PWM) rather than pure analog to drive the heater inside the laser tube. This was probably done to reduce power dissipation in the electronics, but does result in modulation of the optical frequency by the PWM.
A diagram is shown in Internal Structure of Hewlett Packard 5500C and 5501A Laser Tube Assemblies and a photo of an intact one in HP-5501A Laser Tube Assembly. The naked tube is shown in HP-5501A Laser Tube Removed From Magnet and Output Optics Assembly. And Major Components of HP-5501A HeNe Laser Tube for most of what's inside.
See the sections starting with: Notes on the HP-5500 Two Frequency HeNe Laser for more details.
A diagram is shown in Internal Structure of Hewlett Packard 5517B/C/D Laser Tube Assemblies and a photo of an intact one in Tube Assembly Used in HP-5517B/C/D Two-Frequency HeNe Lasers, and after being taken apart in Major Components of HP/Agilent 5517B/C/D or 5501B Tube Assembly. The tube assemblies in HP-5501Bs and older 5517B/C/Ds may differ very slightly. A diagram is shown in Internal Structure of Hewlett Packard 5501B Laser Tube Assembly. But they are functionally identical. The differences are primarily in the beam expander, the supporting structure for the output optics, and the use of a segmented magnet instead of a single piece magnet. They are functionally identical. The tube assemblies in the 5517A, 5518A, and 5519A/B lasers are quite different, being of cast base metal and also much larger as shown in , and optics themselves are the same as in the other lasers as shown in Tube Assembly Used in HP-5517A, 5518A, and 5519A/B Two-Frequency HeNe Lasers, though the actual glass tube, magnet, waveplates, and optics themselves are the same as in the other lasers as shown in Internal Structure of Hewlett Packard 5517A, 5518A, and 5519A/B Laser Tube Assemblies. (Some really old versions of these tube assemblies had heat sink fins rising about 1/2 inch above the top plate which poking through the outer case. These disappeared quite early with no other apparent changes.) The 5517E, 5517FL, and some 5517DL tubes are a bit shorter. I have not seen any 5517Gs but assume them to be similar. See Internal Structure of Agilent 5517E/F/G Laser Tube Assemblies and a photo of an intact one in Tube Assembly Used in Agilent 5517E/F/G Two-Frequency HeNe Lasers, and after being taken apart in Major Components of Agilent 5517E/F/G Tube Assembly (Coming soon).
See the sections starting with: Notes on the HP/Agilent 5517 Two Frequency HeNe Laser for more details.
There's no way to tell the version (e.g., 5517C) or reference frequency (e.g., 2.3 MHz) of the tube itself by inspection of the assembly or from its label. They don't have that information explicitly, only a part number:
The difference in tube part numbers for same model lasers isn't entirely clear. It may be a combination of the size of the beam optics and other special features like a particularly high REF frequency or high output power option.
As noted, the older 5501A and 5500C use physically identical tubes with PZT tuning. The tubes in the 5500A and 5500B are the same and functionally similar to those in the 5501A and 5500C, but the construction differs enough to make it impractical to substitute for those. None of these tubes are compatible with any of the other lasers. The chart below shows the Physical (P) and Reference frequency (R) compatibility of the most common thermally-tuned HP/Agilent lasers. (This should also apply to the low power versions designated with an "L" following the model number (e.g., 5517DL).
(a) (b)
5 5 5 5 5 5 5 5 5 5 5 5
5 5 5 5 5 5 5 5 5 5 5 5
0 1 1 1 1 1 1 1 1 1 1 1
1 7 7 7 7 7 7 7 8 8 9 9
(c,d) B A B C D E F G A A A B
----------------------------------------------------------
(e) 5501B PR R P P P P* P* P* R
5517A R PR PR P P P
5517B P PR P P P* P* P* R
5517C P P PR P P* P* P* R
5517D P P P PR P* P* P* R
5517E P* P* P* P* PR P P
5517F P* P* P* P* P PR P
5517G P* P* P* P* P P PR
(a) 5518A R PR PR P P P
(b) 5518A P R P PR P P
5519A P R P P PR P
5519B P R P P P PR
Notes:
The "P*" (physically compatible with an asterisk) means that beam height specifications for the 5517E/F/G has changed slightly so shims may be need to be added (or transferred) to be fully compatible if installing a 5517E/F/G tube in a 5501B or 5517B/C/D case. Since the 5517E/F/G lasers appear to come with shim washers under the tube assembly feet that are easily lost and difficult to install, it may make more sense to transfer the other parts like the control PCB into its case rather than the tube assembly into another case. Only when the HeNe laser power supply brick under the tube assembly is found to be faulty would this not be feasible.
Legend:
If anyone has additional info defining what these other options mean, please contact me via the Sci.Electronics.Repair FAQ Email Links Page.
There are photos of various HP/Agilent metrology lasers in the Laser Equipment Gallery (Version 2.32 or higher) under "Hewlett Packard/Agilent HeNe Lasers". These include the 5501A laser head and tube, the 5501B laser head, (which is physically similar to the 5517B/C/D except for the connectors), and the 5517A, 5519A, and 5519A/B laser heads. I think the older 5501A tube looks much cooler than the newer ones. :)
The most significant difference between the various lasers is in the Zeeman split reference frequency. A higher frequency enables a faster slew rate for position and velocity measurements. As of Winter, 2010, all the 5517s and 5519s are current Agilent products. General information, descriptions, and specifications may be found by going to Agilent and searching for "laser positioning laser heads" or a specific model number like "5517C". Some of the specifications from the datasheet:
These sound quite incredible but 1 ppm is a frequency of about 474 MHz (1/1,000,000 of 474 THz, the optical frequency corresponding to a wavelength of 633 nm). Thus 0.1 ppm is 47.4 MHz, 0.02 ppm is 9.5 MHz and 0.002 ppm is 0.95 MHz. So, still impressive, but quite reasonable for a well designed stabilized HeNe laser. However, what is somewhat unique about the 5517 and some of the other HP/Agilent lasers is that this absolute accuracy is achieved without the need for any periodic testing or adjustments, by virtue of the design of the mode sampling and locking electronics.
Various other versions of these lasers also exist which may have variations in REF/split frequency, beam diameter, mounting, and other specifications. For example, I am aware of one that uses a tube assembly resembling the one used in the "small" 5517 lasers (but with a different part number), locks using a 5517 control PCB, has a beam approximately 1 mm in diameter, a REF frequency similar to that of a 5517A or perhaps even lower, works in a normal interferometer setup, and a mounting arrangement that is just incompatible enough to be difficult to adapt to a small laser body. But the vast majority of HP/Agilent lasers are standard products.
There is also an Agilent N1211A Fiber AOM Laser Head. This uses a pair of Acousto-Optic Modulators (AOMs) to generate a much higher difference frequency - from 7.5 MHz to 17 MHz depending on version - than is possible with a Zeeman HeNe laser alone. The N1211A may use a more or less standard 5517 laser tube designed for high power with a relatively low split frequency, perhaps the tube assembly described in the previous paragraph. A polarizing beam-splitter separates the two components which are then passed through a pair of AOMs which shift them by appropriate amounts to achieve the desired difference frequency. They are then sent via a pair of fiber-optic cables to an N1212A or N12122B "Remote Optical Combiner" which generates a free-space beam with a diameter of 6 mm or 9 mm, respectively. More info may be found in the section: Notes on the Agilent N1211A Fiber AOM Laser Head. Everything else below deals only with the normal HP/Agilent Zeeman-split HeNe lasers.
With respect to selecting among the various laser models, if your application has no need for the higher REF frequency (often called the split frequency), there is no advantage to getting a laser like a 5517D as opposed to a 5517B. In fact, the lasers with a lower REF frequency tend to have higher output power and thus may be easier to set up and align especially in multiple-axis configurations. They also tend to be less expensive on the surplus market, though the Agilent price isn't that much different. The only disadvantage of a laser with higher output power is that there can be enough of a detected MEAS signal due to slight angular misalignment of interferometer optics like the 10706A to result in a reading even if the beam to the tool or whatever whose position or velocity is to be measured is blocked or misaligned. The interferometer cube contains a polarizing beamsplitter and if the F1 and F2 orientation are not precisely aligned with the polarizer, there will be a small amount of F1 mixed with F2 and vice-versa even without the reflection from the mirror on tool. With a 400 µW laser and single axis, the required angular accuracy to avoid a false MEAS signal is well under 1 degree with the optical receiver threshold at its default most sensitive setting. And even if the alignment is perfect, polarizing beam splitter and AR coatings are not perfect so there can still be residual mixing. None of this matters once the return beam is aligned since the MEAS signal will be much stronger than the bogus one, but it can be confusing. Increasing the threshold may be desirable to avoid the issue.
And a note about that impressive spec'd lifetime of 50,000 hours - about 6-1/4 years of continuous use. HP lasers used to last a long long time and it wasn't unusual to find an HP laser running fine after 8 years. But I rather suspect this is no longer the case. I've seen many late model (2004 to 2006) Agilent 5517s that were going down hill well before 6 years including at least one that was essentially dead after less than 3 years. These were standard 5517Bs or 5517Cs pulled from semiconductor fabs, either because they failed in normal use, or because they were rejected during preventive maintenance due to low power or the REF frequency going out of spec (which is usually related to the power decline). Thus, even a late manufacturing date is no longer assurance of a healthy laser. Nor would even a close inspection of the HeNe laser tube, as the they appear identical except for the Agilent label - perhaps that's enough! So if you are buying these things new, it probably pays to go for the extended warranty. :)
Note that in this diagram and the others depicting Zeeman-split mode behavior, the magnitude of both normal mode pulling and the mode pulling that produces the split frequency is greatly exaggerated. For example, even for a 5517D with a split/REF frequency of 3.4 to 4.0 MHz, the spacing between F1 and F2 would be only about 0.3% of the longitudinal mode spacing of 1.2 GHz! Without the plots showing F1 and F2 spaced more than 2 orders of magnitude further apart than they really are, they would merge into one line on any reasonable size diagram!
Waveplates at the output of the HeNe laser tube convert the left and right-hand circularly polarized Zeeman split modes to linearly polarized modes that are orthogonal and aligned with the horizontal and vertical axes of the laser. These two modes usually differ in optical frequency by between 1.5 and 4 Mhz (depending on the specific laser). (Some recent versions of the 5517 may actually go to 7 MHz or more.) The X and Y polarizations are sent down different paths in the metrology application. One is generally a reference length and the other is the dimension to be measured or tracked. (It's the change in path length difference that matters so they could both move if desired.) Rather than creating an interference pattern that changes slowly, the two beams are combined together in a detector that outputs a difference (or heterodyne) signal. If the relative distance between the two beam paths changes by one half wavelength of the laser (about 632.8 nm but accurate to many significant digits!), the phase of the difference signal will change by 360 degrees. The laser also generates an electrical signal from beating the signals together internally. This constant reference is compared to the detector signal and an electronics package measaures the phase shift continuously and uses it to determine the distance traveled.
A moderately powerful cylindrical permanent magnet (hundreds of gauss) does the Zeeman splitting resulting in a pair of circularly polarized outputs at two very slightly different frequencies. F1 is designated the lower frequency and F2 is designated the higher frequency. For the 5501A/B, F1 is vertical (perpendicular to the laser base) while F2 is horizontal (parallel to the laser base). For the 5517A/B/C/D/E/F/G, 5518A, and 5519A/B, F1 is horizontal (parallel to the laser base) while F2 is vertical (perpendicular to the laser base). (Exactly why HP switched orientations between the two model series is not clear as there is no benefit to one over the other and it just causes confusion, or perhaps that was the intent!) The difference between F1 and F2 ranges from 1.5 to 4 MHz for most of the HP/Agilent lasers depending on the model (as listed above) and also the specific sample of the laser. The cavity length of the early HP lasers (5500A/B/C and 5501A) is PZT-controlled using feedback based on the amplitude of the two modes. For the 5501B and all later lasers like the various 5517s, cavity length is adjusted by a heating coil wrapped bifilar-style around the bore inside the tube. In all cases, the feedback is used to maintain the position of the lasing lines symmetric on the Zeeman split neon gain curves as shown in Axial Zeeman Split HeNe Laser Mode Behavior. A Quarter WavePlate (QWP) converts the circular polarized output to orthogonal horizontal and vertical polarized components which are used externally. F1 is reflected from whatever is being measured or tested (e.g., disk drive servo writer or wafer stepper) and F2 is reflected from a fixed reference. The difference frequencies (F1-F2) and (F1-F2)+dF1 are then analyzed to determine precise position, velocity, or whatever. The end result is identical in terms of sensitivity to position changes compared to the common single frequency (or homodyne) interferometer, but the two frequency approach has lower noise and greater stability, and is therefore potentially more accurate.
Interestingly, the actual beat or reference frequency (REF) does NOT need to be super stable over the long term. Rather, it is the difference between REF and the return (MEAS) signals that matters and that only depends on the motion of the target reflector, the optical frequency of the meausrement beam, and the speed of light. Thus, although the optical frequency needs to be known to high precision (+/-0.1 ppm for the standard lasers; +/-0.02 ppm for those calibrated to MIL STD-45662), the exact beat frequency of each laser is not precisely controlled or even precisely measured and recorded or used anywhere in the calculations. This is one reason why the listings above include only a range of values. Any given sample will operate somewhere within that range during its expected life, but the exact value is somewhat random depending on the specific characteristics of the tube/magnet assembly and the specific place on the neon gain curve that the lasing line is parked. In fact, REF tends to increase over the life of the laser as the gain and thus output power decline with use. As long as REF remains within the spec'd range for the particular model laser, then the system in which it is installed will work properly. What exactly a machine will do if REF goes out of range is implementation dependent. But one reason for a laser such as this to be replaced is for REF to approach or exceed the high end of the spec'd frequency range, though in many cases, a REF which greatly exceeds the spec'd upper limit can be tolerated.
While one might think that locking the difference frequency to a crystal reference would be superior - and the technique is patented - it's not clear that this would be better and might actually be worse. The difference frequency relative to the mode position can change for any number of reasons. In fact, the REF frequency of HP/Agilent lasers tends to slowly vary by a small amount (typically a fraction of 1 percent) even after locking with the period of the cycle increasing as the tube assembly approaches thermal equilibrium. The cause is probably back-reflections from internal optics and the resulting etalon effect. Despite this, because the amplitude of the two modes is forced to be equal to keep the modes centered on the split neon gain curves, the optical frequency ends up being very stable.
All of the HP lasers use conventional dual polarization mode stabilization to lock the lasing lines to the split neon gain curve. However, the two signals are not from adjacent longitudinal modes as with most common laboratory stabilized HeNe lasers, but are the two Zeeman split sub-modes differing in frequency by a few MHz instead of many 100s of MHz. In fact, both are the same cavity mode but shifted slightly higher and lower than would be predicted by c/2*L. One twist on the implementation is that the 5501B and those below it on the chart use a Liquid Crystal Device (LCD) polarization selector to alternately sample the horizontal and vertical polarized modes. The LCD consists of a large area single pixel LCD (!!) with a linear polarizer bonded to it. Applying a voltage to the LCD rotates the polarization bf the sampled beam by 90 degrees prior to it passing through the polarizer. This sensed output is fed to a subtracting ample-and-hold to compare the amplitudes of the two polarized components in the error amp driving the heater. This is radically different than the polarizing beamsplitter and dual photodiodes used in most other dual polarization mode stabilized lasers including the 5500C and 5501A. The LCD approach does have a sort of elegance as well as practical benefits. Since the same optical path and photodiode are used for both polarization modes, the sensitivity is identical so the mode balance should be perfect assuming the LCD polarization rotation is 90 degrees. Since the intent is to park the modes symmetrically on the split neon gain curve, this is perfect and thus requires no offset adjustment over the life of the laser as the output power of the tube declines. And, the LCD and associated electronics may in fact be cheaper than a high quality polarizing beam splitter. However, it also creates some artifacts as a result of the digital switching, resulting in small cyclical variations in optical frequency over a period of 2 or 3 seconds. These are of no consequence for most metrology applications, but do detract from the elegance of these lasers.
In fact, the thermally tuned lasers have only one adjustment associated with stabilization, and that is for the temperature setpoint at which the controller switches from pre-heating to optical locking. The resistance change of the actual heater coil is used to sense temperature and there are variations from one tube to the next. But this is an extremely non-critical setting and won't affect accuracy, only possibly the temperature range over which the laser will remain locked. (5517s with the Newest Type II PCB may have no adjustments, or at least none that are obvious!)
One oddity with respect to the thermally tuned laser tubes is the patent reference that appears on the label of all newer ones at least: "Licensed by Patlex Corporation Under Patent No, 4,704,583". The title of this patent is: "Light Amplifiers Employing Collisions to Produce a Population Inversion", filed in 1977 but not granted until November of 1987. The most curious thing is that there appears to be very little of relevance in the patent other than its association with laser action! Nothing in the patent diagrams or text has any obvious connection to the tube assembly design. In fact, the exact same text exists on other more mundane things like a Carl Zeiss-badged Siemens LGK 7634, a bog standard 2 mW random polarized HeNe laser head. I've heard that Patlex is actually a bunch of lawyers and I bet they made out or are making out quite well. :)
There are two possible "simple" causes of the lasing frequency shifts resulting from Zeeman splitting: Mode pulling (which tends to attract each lasing mode towards its respective gain center) and magnetically-induced birefringence of the plasma (which results in the effective cavity length differing for each lasing mode's polarization).
Several papers attributes the phenomenon entirely to the birefringence of the plasma for right and left circularly polarized photons. (See for example: T. Baer, F. V. Kowalski, and J. L. Hall, "Frequency stabilization of a 0.633-µm He-Ne longitudinal Zeeman laser", Applied Optics, vol. 19, no. 18, 15 September 1980). While birefringence may be a factor, those researchers were testing at at very low magnetic fields and their results don't explain some observations for lasers like those from HP/Agilent. In addition to cavity loss (or cavity Q) not entering into their equations at all, one fundamental result - that the split frequency is a minimum where the lasing lines are centered on the split gain curves - is exactly opposite from reality for most HP/Agilent and similar lasers! So that leaves mode pulling.
Though I've now lowered the bogosity quotient to 0.1 (down from a much higher value until quite recently), some aspects of the following explanation may still be totally without any basis in fundamental physics. But it has the attractive property that using some reasonable assumptions and not-so-hairy math, it is able to predict the approximate behavior of real HP/Agilent two-frequency lasers. And there is also support for the mode pulling being involved in the book: "Gas Lasers", by Charles Geoffrey Blythe Garrett, McGraw-Hill advanced physics monograph, 1967.
The mechanism for the shift of the Zeeman modes away from the cavity modes is a type of mode pulling - or at least can be modeled that way. There is no need to invoke esoteric effects like plasma or mirror birefringence (though these may alter the results in subtle ways, particularly at low magnetic fields). Mode pulling essentially shifts the positions of the longitudinal modes of a Fabry-Perot laser slightly away from the locations determined by c/2L of the linear cavity, toward the gain center. The basic mode pulling equation for a normal (non-Zeeman) laser (with truly galactic-size gobs of assumptions!) is:
CB
FS = FSR * ---------
GB + CB
Where:
For high-R mirrors with equal reflectivity, Finesse = π*sqrt(R)/(1-R). Where one of the mirrors is HR as is the case with HeNe lasers, Finesse will actually be close to 2*π*sqrt(R)/(1-R). This equation predicts the shift of an adjacent longitudinal mode 1 FSR away from a mode centered on the peak of the neon gain curve toward the peak. As will be seen later, this is where the magnetic field comes into play. What the equation does show is that for GB >> CB, FS is proportional to CB, or equivalently, the lower the finesse of the cavity, the more frequency shift will be present. But at least this should get us into the ball park.
A factor of 2 should be tossed in since what we're interested in is not the shift of a single mode, but the change in distance between the two modes due to each of them shifting in opposite directions. However, this can arguably be partially offset by the fact that the modes are positioned only part of the way down down the gain curves due to the Zeeman splitting magnetic field, not at 1 FSR away and farther down where the mode pulling effect would be greatest. So, it's somewhere in between. (Don't worry, much of this hand waving will go away shortly.) So, we'll use a fixed value in place of FSR selected and for want of something better, let's select it so the REF frequency is reasonable for the 5517B tube whose mirror reflectivity is known.
So, here are 3 laser tubes with OC mirrors whose reflectivity has been measured (dissected conventional Melles Griot 05-LHR-006 barcode scanner tube, HP-5501A tube, and HP-5517B tube), and several others where the reflectivities can be estimated based on difference frequency (REF) specifications (5517A/C/D/E/F/G). There is also a Melles Griot 05-LHR-007 (same as the Spectra-Physics 007), which is about the shortest commercially available barcode scanner tube. The conventional tubes were installed in an HP magnet (model unknown or not remembered) but not locked to the mid-point between the split neon gain curves:
Conventional tubes:
Cavity Cavity OC Mirror Cavity <--- F2-F1 (REF) --->
Tube Type Length FSR Reflectance Finesse Predicted Measured
-----------------------------------------------------------------------------
LHR-006* 139 mm 1.078 GHz 0.990 625 1.34 MHz 1.2-1.6 MHz
LHR-007 110 mm 1.360 GHz " " " 1.70 MHz 1.5-1.7 MHz
HP/Agilent lasers (constant magnetic field):
Cavity Cavity OC Mirror Cavity <--- F2-F1 (REF) --->
Tube Type Length FSR Reflectance Finesse Predicted Specified
------------------------------------------------------------------------------
5501A* 130 mm 1.153 GHz 98.74% 495.5 1.74 MHz 1.5-2.0 MHz
5501B 127 mm 1.180 GHz 98.8% 520.4 1.70 MHz 1.5-2.0 MHz
5517A " " " " " 98.8% 520.4 1.70 MHz " " MHz
5517B* " " " " " 98.5% 415.7 2.13 MHz 1.9-2.4 MHz
5517C " " " " " 98.2% 345.9 2.66 MHz 2.4-3.0 MHz
5517D " " " " " 97.5% 248.2 3.70 MHz 3.4-4.0 MHz
5517E 101.6 mm 1.475 GHz 96.8% 193.2 5.94 MHz >5.8 MHz
5517F " " " " " 96.2% 162.2 7.07 MHz >7.0 MHz
5517G " " " " " 96.1% 157.9 7.26 MHz >7.2 MHz
"*" denotes tubes where the OC mirror reflectance was measured.
Note that the predicted values assume a constant (though not known) magnetic field, but we know that its actual strength also strongly influences F1-F2, roughly in proportion to its strength. And based on measurements of many HP/Agilent lasers, the magnetic field may differ by up to nearly a factor of 2 depending on model and specific sample. (At first, I assumed they were all the same!) Higher REF frequency lasers like the 5517F (nearly the highest split frequency laser available) on average have stronger magnetic fields than lower REF frequency lasers like the 5517A, but there are some notable exceptions. (See the section: HP/Agilent 5517 Laser Construction.) So, a selection process must be involved in mating magnet and tube to achieve a specific split frequency. Nonetheless, the mirror reflectivities for higher split frequency lasers in the chart are almost certainly way too low and would likely result in no lasing at all for these short tubes. Thus the actual mirrors would have somewhat higher reflectivities since their magnetic field strengths tend to be higher as well. For example, the 5517G mirrors may be easily over 97%. While still low for a normal laser (where 99% or more would be typical), it is a much more realistic value for the short cavity of the 5517G where maximum output power is not the objective.
OK, so I kind of picked the reflectivities for the 5517A/C/D/E/F/G mirrors to make the results reasonable. :) With the increasing cavity loss, the output power of lasers with higher REF frequencies will tend to be lower, but the reflectivites listed may simply be too low to lase at all or with useful power on a tube of this length. However, the actual change in discharge length going from the 127 mm to the 101.6 mm cavity is small, perhaps at most 10 mm, so not that much gain is lost. And the bore of the short tubes appears to be narrower and uniform (as opoposed to the stepped bore of the long tubes. This would also result in higher gain, possibly totally making up for the loss in discharge length. Eventually, I will measure the reflectivity of OCs for other HP/Agilent lasers. But I don't have any tubes that I'm willing to take to bits at the present time, partially due to (1) the physical and emotional trauma that would result and (2) the fact that I haven't located the special chants and incantations required for metrology laser tube sacrifice. :) If anyone has done this, or has certifiable 5517 tube bits or a 5517 tube that's already cracked or broken they'd be willing to contribute to the cause, please contact me via the Sci.Electronics.Repair FAQ Email Links Page.
Of course, all these nice results based on numerous assumptions may be wishful coincidence, but they are close to what is observed and don't require delving into esoteric plasma physics. Whew! :)
However, since we do know that the magnetic field is what splits the neon gain curves and moves them apart in proportion to the field strength and this affects the split frequency roughly via a proportionality constant (up to a point where the output power goes to zero!), the magnetic field must be added to the equation. Measurements suggest that the magnets used in the various 5517A/B/C/D/E lasers and 5501B tend to have a higher strength for higher split frequencies, but not always. So, mirror reflectivity alone is probably not used for split frequency selection. And there may in fact be some "mix and match" going on mating tubes with magnets to achieve the desired split (REF) frequency for the specific laser model. For example, the spec'd REF frequency for the 5517B is 1.9 to 2.4 MHz. The completed laser must start out with a REF that is low enough to anticipate the effects of normal tube aging where the beat frequency tends to increase as well as operation at high ambient temperature where the REF frequency tends to decrease (due to higher tube pressure). So, new 5517Bs typically have a REF frequency between 2.1 and 2.3 MHz.
If we assume that the split frequency is proportional to the magnetic field, we can modify the super simplified mode pulling equation as follows:
CB
DF = H * ZSS * k * ---------
GB + CB
Where:
Note that the product of H, ZSS, and k replaces the FSR (mode spacing) in the previous equation. Using 1.5 GHz for the FWHM gain bandwidth of the split Doppler-broadened neon gain curves, H measured in gauss, and assuming that CB is small compared to the GB, this simplifies to:
2.8 MHz CB
DF = H * ---------- * 1.17 * ---------
G 1.5 GHz
(H is in gauss.) And if the Garrett equation is boiled down to its fundamentals, the result is similar. (You really don't want to see the original!) So in the end, it comes down to the difference frequency being proportional to the product of H and CB multiplied by a constant. The value 1.17 for k in the equation above results from the desire for the predicted REF frequency of the 5517B (whose mirror reflectivity and magnetic field have both been determined) to be the nominal value for the 5517B, 2.20 to 2.30 MHz. :) Thus, the result might be called a "phenomenological equation" - based on physical principles but adjusted to fit actual observations. :-) Garrett's equivalent of "k" is 1.00. But it turns out that this can be dealt with as well. :)
As was shown above, the difference frequency is (approximately) inversely proportional to the neon gain bandwidth (GB). All of the calculations so far have assumed the GB to be 1.5 GHz, which is used rather than 1.6 GHz because the GB will be narrowed slightly for each of the split gain curves since they each only have 1/2 the original gain. This helps to boost the difference frequency but also decreases the power output at the lock point! Other factors being equal, the GB for a shorter tube with less gain will be even narrower, further increasing the difference frequency, thus making it easier to obtain the required higher ones with lower magnetic fields and/or higher OC reflectivities. And, since GB gets narrower as the overall gain declines with use, this could also explain at least in part why the split frequency tends to increase over the life of the tube. How all these factors interact is not something I'd want to contemplate, let alone calculate! :)
However, there would appear to be something even more signficant going on. Looking at the bottom diagram in Normal and Zeeman-Split HeNe Laser Mode Power Curves, which is based on actual data, in order for there to be no unwanted "rogue" modes oscillating, it is seen that the effective GB (lasing output power curve which is basically the neon gain curve above threshold) must not be much greater than the FSR of the laser tube, which For the 5517B (long tube), is 1.180 GHz. And it must not extend to or beyond the FSR on either side of the lock point. Or to put it another way, the maximum total shift cannot be greater than 2*(FSR-GB/2) or both the center split mode and the adjacent modes will be above threshold. Thus, the GB cannot be 1.6 or even 1.5 GHz in this case because then there would enough gain for modes on either side of the split lasing mode to lase. In the diagram, the GB is about 1.28 GHz, which is consistent with an actual measurement in SFPI Display of Lasing Mode Power Envelope of Horizontal Polarized Output of Healthy HP/Agilent 5517B Laser. The maximum total shift in this case would be 1.08 GHz. Conventional HeNe laser tubes do not seem to exhibit this effect, or at least not to the same extent.
This brings up the issue again of whether the gases in HP/Agilent tubes are isotopically pure since using only 20Ne or 22Ne would reduce the GB significantly. The shift in the gain center going from one isotope to the other is about 1 GHz! Using only a single isotope could easily account for this narrowing.
Based on some numbers obtained from industry sources, it's possible that the HP/Agilent mix is skewed toward Ne20. The specific conditions under which these numbers were obtained are not known, but they seem reasonable:
Isotope/Mix Vacuum Wavelength HP/Agilent Lasers --------------------------------------------------------------- Pure 22Ne 632.9900840306 nm 50% 20Ne 632.9907522876 nm 90.48% 20Ne 632.9912933099 nm Natural Ratio 95% 20Ne 632.9913537204 nm 5517C/D/E/F/G, 5519A/B 96.5% 20Ne 632.9913724316 nm 5501B, 5517A/B, 5518A Pure 20Ne 632.9914205461 nm
(Another stable isotope of neon is 21Ne, about 0.27%, which is ignored and lumped in with 22Ne.) If these numbers are accurate, the HP/Agilent mix is skewed a bit toward 20Ne compared to the naturally occuring mixture. And if in the interest of cost reduction, HP switched to a mixture with less 20Ne, that would explain the change in wavelength in the early to mid-90s - from 96.4 to 95 percent pure 20Ne. Note that simply using the natural mix would be better than a more equal ratio that may be used to optimize output power by widening the gain bandwidth!
But it turns out that the narrowing actually only occurs at the high magnetic fields present in HP/Agilent lasers. With no magnetic field, the GB of a healthy tube is between 1.5 and 1.6 GHz as expected. So, isotopic purity may not be fundamental to the line narrowing.
Using a Scanning Fabry-Perot Interferometer (SFPI), The GB of a healthy short 5517 tube was measured with a magnet and naked. Don't ask where I found a healthy bare tube! I can only say that no lasers needed to be sacrificed this time. :) Gain and GB will increase as the tube warms up from its bore discharge alone, but for these tests, 8 V was applied to the internal heater to bring it up to approximately the temperature at which the laser would normally operate:
So the high magnetic field in conjunction with mode competition narrows the GB. Whether the narrower GB is really what should be used in the equation isn't known for sure. But it would appear to make sense assuming the mode pulling for the RCP and LCP modes is independent. However, since the effective GB decreases with with increasing magnetic field, the beat frequency may increase faster then would be accounted for simply by the increase in magnetic field and this may help to explain at least in part why theory and practice do not agree - yet!
As noted, the GB narrowing appears to be much more significant in HP/Agilent tubes compared to conventional tubes. Testing of a Melles Griot 05-LHR-007 barcode scanner tube using an SFPI shows a GB of around 1.6 GHz without a field and and at best goes down to 1.5 GHz inside a 5517C magnet. Further, the skewing in the lasing output power profile due to mode competition is much less severe with HP/Agilent tubes. My hypothesis is that conventional tubes are filled for maximum power at a ratio of round 1:1 with 20Ne:22Ne while HP/Agilent tubes are filled with nearly pure 20Ne to narrow the gain bandwidth.
Another phenomenon that isn't addressed at all is the variation in beat frequency during mode sweep. All of these simplified mode pulling equations use only the distance from gain center but not the gain profile, so position of the split lasing mode doesn't matter as long as itis between the gain curves. But we know that depending on the magnetic field, the beat frequency may be a maximum or minimum when the split mode is centered between the gain curves - or it may not change much at all.
And then, there is a second term in the Garrett forumation that has been totally ignored in anything discussed so far and will continue to be ignored for the forseeable future!
But here goes:
2.8 MHz CB
DF = H * --------- * ----------
G 1.28 GHz
(As above, H is in gauss.) This is precisely the simplified form of the Garrett equation. So, all the bogosity terms must have canceled out. ;-)
Plugging in numbers for the reference 5517B:
2.8 MHz 2.84 MHz
DF = 362 G * --------- * ---------- = 2.25 MHz
G 1.28 GHz
Now, this isn't 100.0000000 percent accurate. :) The resulting DF becomes 2.25 MHz to make it work for the reference 5517B. But it was still close, previously 2.21 MHz, and the actual DF for that specific tube in that specific magnet is not known.
The strengths of the magnetic fields for samples of most of the long tube HP lasers (5501A/B and 5517A/B/C/D) have been measured directly. Those of the short tube Agilent lasers (5517E/F/G) have been estimated by measuring the fringe field at the center of the magnet exterior relative to the fringe field of a 5517E, which was the only laser that had a label (363g). Then the interior field was kind of guessed based on measurements of magnets from the long tube lasers. This is the only option available for determining the interior field on intact tube assemblies. Since the consistency among the various magnets on the ratio of outside to inside is not very good, the actual fields may be quite different, most likely lower. More info on the measurements may be found in the section: HP/Agilent 5517 Laser Construction.) But for the few tubes that have been removed from the magnet, the field in the interior could be measured inside. But even so, the field may vary significantly along the length of the tube, so the value used is a sort of more or less average. :) I built a simple gauss meter specifically for making these measurements. See the section: Simple Gauss Meter for Measuring Zeeman Magnet Strength.
Adding the magnetic field into the equation for the 05-LHR-006 and 05-LHR-007 barcode scanner tubes, and what may be a special version Hughes 3121H (unmarked) somewhata longer tube produces results that are quite reasonable. While the OC mirror reflectivity for these tubes has not been measured, 99% is usually a good value for short tubes. The two entries for the LHR-007 are in 5517D and 5517A magnets; the LHR-006, LHR-640, and 3121H are in a 5517C magnet. The range of predicted REF values corresponds to a GB between the 1.6 and 1.28 GHz.
Conventional tubes (measured magnetic field):
Tube Magnet Cavity Cavity OC Mirror Cavity <---- F2-F1 (REF) ----> Type Field Length FSR Reflect. Finesse Predicted Actual ------------------------------------------------------------------------------- LHR-007 250 G 110 mm 1.360 GHz 99.0% 625 0.95-1.19 MHz 1.12 MHz " " 380 G " " " " " " " 1.45-1.81 MHz 1.66 MHz LHR-640 300 G 118 mm 1.272 GHz " " " 1.07-1.33 MHz 1.25 MHz " " 350 G " " " " " " " 1.25-1.56 MHz 1.50 MHz LHR-006 350 G 139 mm 1.078 GHz " " " 1.15-1.43 MHz 1.25 MHz 3121H 350 G 192 mm 0.781 GHz " " " 0.76-0.96 MHz 0.70 MHz
For the 3121H, the measured REF is slightly out of range low but the cause may be that for this longer tube, the magnet covered only about 2/3rds of the active bore.
Note that due to the gas-fill and cavity length, some if not all of these tubes will produce "rogue modes" even when the Zeeman-split mode is centered between the gain curves as it would be when locked in an HP/Agilent laser. (Rogue modes are non-Zeeman lasing lines on the tails of the gain curves.) For example, the 05-LHR-640 in the 350 G 5517C magnet had a pair of rogue modes, each about 5 percent of the total power. Reducing the magnetic field to approximately 300 G eliminated them. Only the 05-LHR-007 with its shorter cavity length may be immune.
The 05-LHR-640 in the chart above was in like-new condition with a power output of around 1.25 mW, well above its 0.5 mW spec. A high mileage 05-LHR-640 with a power output of only 0.6 mW was also tested in the same magnet. It had a split frequency of around 1.75 MHz - well above the predicted range - probably due to the reduced gain and/or increased cavity loss. However, interestingly, rogue modes were still present, and possibly even a bit larger relative to the Zeeman mode amplitude than for the healthy sample.
For the HP/Agilent laser tubes, if we use the measured magnetic field where available, or else the estimated magnetic field, the values for mirror reflectivity become more realistic:
HP/Agilent lasers (measured or estimated magnetic field):
Tube Magnet Cavity Cavity OC Mirror Cavity <--- F2-F1 (REF) ---> Type Field Length FSR Reflect. Finesse Predicted Specified ------------------------------------------------------------------------------ 5501A*+ 371 G 130 mm 1.153 GHz 98.74% 495.5 1.84 MHz 1.5-2.0 MHz 5501B+ 256 G 127 mm 1.180 GHz 98.5% 415.7 1.59 MHz " " MHz 5517A* 259 G " " " " " 98.5% 415.7 1.60 MHz " " MHz 5517B*+ 362 G " " " " " 98.5% 415.7 2.25 MHz 1.9-2.4 MHz 5517C+ 353 G " " " " " 98.2% 345.9 2.61 MHz 2.4-3.0 MHz 5517D+ 380 G " " " " " 97.9% 296.0 3.62 MHz 3.4-4.0 MHz 5517E 490 G 101.6 mm 1.475 GHz 97.7% 270.0 5.82 MHz >5.8 MHz 5517F 515 G " " " " " 97.3% 229.5 7.20 MHz >7.0 MHz 5517G 527 G " " " " " 97.3% 229.5 7.37 MHz >7.2 MHz
The "*" denote lasers where the mirror reflectivity has been measured. The "+" denote lasers where the magnetic field of at least one sample has been measured and this is the value that is used. All the others use the average value of the magnetic field predicted from external measurements (which can be quite unreliable) and mirror reflectivities selected to make the resulting REF frequency be a reasonable value.
So, it looks like the identical tube is used for the 5501B, 5517A, and 5517B, with the higher magnetic field boosting REF for the latter. But the 5517C and 5517D require lower mirror reflectivities to be consistent with the measured field strength.
As a practical matter, 5 different tube types would not be necessary (based on the mirror reflectivity estimates, above) but it's almost certain there were more than one since the range of magnetic fields is not sufficient for both the 5517A and 5517D. However, with careful control of tube design, it's quite possible for there to be only two types of tubes - long and short. All tubes of each type would then be identical - including mirror reflectivities - with REF set solely by magnetic field strength. Perhaps Agilent has a "Dial-A-Field" Magna-Matic-Magnetizer: Insert tube and adjust field strength to obtain the desired laser specifications. ;-) Then, we could have the following:
HP/Agilent lasers (two tube types, selected magnetic field):
Tube Magnet Cavity Cavity OC Mirror Cavity <--- F2-F1 (REF) ---> Type Field Length FSR Reflect. Finesse Predicted Specified ------------------------------------------------------------------------------ 5517A 250 G 127 mm 1.180 GHz 98.3% 366.4 1.76 MHz 1.5-2.0 MHz 5517B 316 G " " " " " " " " " 2.22 MHz 1.9-2.4 MHz 5517C 380 G " " " " " " " " " 2.67 MHz 2.4-3.0 MHz 5517D 261 G 101.6 mm 1.475 GHz 97.0% 206.3 3.65 MHz 3.4-4.0 MHz 5517E 380 G " " " " " " " " " 5.91 MHz >5.8 MHz 5517F 455 G " " " " " " " " " 7.08 MHz >7.0 MHz 5517G 470 G " " " " " " " " " 7.31 MHz >7.2 MHz
The mirror reflectivities were selected to be as high as possible without requiring magnetic fields so strong that they introduce problems of their own. Though Alnico magnets can have a field strength "at the poles" of over 1,500 g, nothing close to this may be achievable inside an actual cylindrical magnet. (In fact, measurements of some of these magnets do show more than 1,000 G at the poles even though the interior field is less than half that.) But more fundamentally, the neon gain curves will be spread so far apart that rogue modes may be generated if the magnetic field exceeds about 385 G for long tubes and 475 G for short tubes. And the limit at which the output power goes to zero because the neon gain curves are spread so far apart that there is no longer any overlap may not be far above 400 g. Thus, the listed fields could be too strong for the 5517F/G.
It would be possible to select the mirror reflectivity for long tubes such that the 5517D could use either type tube. But to keep the field below the "rogue mode limit", would have required a mirror reflectivity of under 98%.
These same tubes are also used in other common HP/Agilent lasers. The 5518A would use the 5517A or 5517B tube depending on serial number, the 5519A would use the 5517C tube, and the 5519B would use the 5517D tube. There are some (not very common) options mostly for the 5517D where a higher REF frequency is specified. These are probably accommodated with magnetic field adjustments.
Finally, there is one other Agilent laser, the N1211 "AOM Laser", which is probably designed for higher power rather than a specific range of REF frequencies since the REF frequency doesn't really matter the way it is being used: An Acousto Optic Modulator (AOM) is used to generate a much higher difference frequency (REF) between F1 and F2. The original REF from the tube adds into that but is small in comparison. Since REF can be lower, the mirror reflectivity could be higher to boost power. Measurements of the external magnetic field of an N1211 laser suggest that it has a magnet like the one in the 5517C and that could indeed work out assuming an OC mirror reflectivity of 99% as in the first entry below:
Tube Magnet Cavity Cavity OC Mirror Cavity <--- F2-F1 (REF) ---> Type Field Length FSR Reflect. Finesse Predicted Measured ------------------------------------------------------------------------------ N1211 350 G 127 mm 1.180 GHz 99.0% 625.2 1.44 MHz 1.60 MHz " " 350 G " " " " " 99.3% 894.5 1.01 MHz 1.60 MHz " " 250 G " " " " " 98.5% 415.7 1.55 MHz 1.60 MHz
That one sample had been removed from service, so its REF frequency may have already crept up some. Or, if the mirror reflectivity were really only 98.9%, then the predicted REF frequency would be 1.59 MHz. However, since the correlation betwee external and actual internal magnetic fields is rather poor, it could simply be a 5517A tube selected for high power, which would be quite likely. Why create yet another tube type for what must have been a very limited production run of these very specialized lasers? I've come across several boxes of Agilent HeNe cavity mirrors undoubtedly intended for use in these type of laser tubes. These were mostly only photos of the label but I do have a few boxes in-hand and was able to measure the diameter and RoC which match mirrors from dissected 5517 tubes within my margin of error. The first were labeled "2% HeNe cc/pl" with a manufacturing date of 2003. The "2%" refers to the mirror transmission, or equivalently, a reflectance of 98%. (I actually measured them to be 2.1%T/97.9%R but we'll go with the labeled values.) There was no further description so they could have been for any of the Agilent lasers. But the 5517C would probably be the most likely. However, it's possible that Agilent was actually adjusting the reflectivities to reduce the number of different tube types. These same mirrors would work easily for 5517As, 5517Bs, and 5517Cs, with somewhat lower magnetic fields than in the table, above. But using them for a "long tube" 5517D would require a very strong magnetic field, possibly resulting in rogue modes:
Tube Magnet Cavity Cavity OC Mirror Cavity <--- F2-F1 (REF) ---> Type Field Length FSR Reflect. Finesse Predicted Specified ------------------------------------------------------------------------------ 5517A 210 G 127 mm 1.180 GHz 98.0% 311.0 1.74 MHz 1.5-2.0 MHz 5517B 275 G " " " " " " " " " 2.28 MHz 1.9-2.4 MHz 5517C 330 G " " " " " " " " " 2.73 MHz 2.4-3.0 MHz 5517D 420 G " " " " " " " " " 3.48 MHz 3.4-4.0 MHz
A similar box of mirrors has a part number with "5518" in it so they may have been intended for 5518A lasers. It listed the same 136 mm RoC but a transmission of 1.6% (98.4%R). This is enough greater than what I measured (1.5%T, 98.5%R) for the 5517A (which should be similar to the early versions of the 5518A) to suggest that Agilent may have been trying to boost output at the expense of REF frequency. And another one listed 2.1% (97.9%R), also with a "5518" in the part number, so that could have been for later versions of the 5518A where the REF frequency is spec'd to be 1.7 to 2.4 MHz, or even for 5519A/Bs (2.4 to 3.0/3.4 to 4.0 MHz). And yet another labeled "3%" (97%R), also with 5518 in the part number, but would more likely be used in a "long tube" 5517D or 5519A/B and is perfect for any of the "short tube" lasers and the default in the chart, above. But now we are into absolutely totally pure speculation. :) And one more box of Agilent mirrors yielded a 13.6 cm RoC OC with a reflectance of 99.3%. That high a reflectance must have been selected for maximum power, which would be near optimal for these relatively short tubes.
Once thing is certin: The wide variation in mirror reflectance/transmission confirms the overall scenario whereby cavity loss is a key parameter in designing HP/lasers to meet REF frequency specifications. And the range found so far agrees with my predictions.
Some of these mirrors have a flat on one side, purpose unknown. At first I thought it might simply be the edge of a larger plate from which these were cut with the optics version of a plug cutter. But it could be for indexing or simply to prevent the mirrors from rotating in their mounts, or to provide a path for the discharge to pass by the mirror if the anode were behind it (possible in the short-tubes). Some of the mirrors were also thinner than the normal ~4 mm. It's possible some or all of these were prototypes or trial runs, I'm not sure how much significance to attach to these details. :)
To summarize:
For the short HP/Agilent tube, there can only be a single split lasing mode present when the laser has locked (READY on solid) centered between the two gain curves, though a second RCP or LCP mode may be present at times during mode sweep before the laser locks. For longer tubes, there may be additional RCP or LCP modes or even multiple split lasing modes. There may then be additional beats present due to the split modes and with RCP or LCP modes.
For an HP/Agilent tubes with its strong magnetic field, the maximum beat frequency usually occurs when the split lasing mode is centered. With weaker magnetic fields, the minimum frequency may be at that location. This is also probably related to the profile of the neon gain curves and how mode competition changes their effective shape. The maximum beat frequency for HP/Agilent lasers varies from 1.5 MHz to over 7 MHz depending on the model.
To reiterate, the only direct consequence of the Zeeman effect is to split the discharge spectral lines and shift their position up and down in optical frequency. Since the neon gain curves are related to the emission spectra, they then also get shifted. Lasing which takes place will then be RCP or LCP depending on which gain curve applies. The Zeeman effect does NOT produce the beat frequency directly. A spectrometer or interferometer with sufficiently high resolution would show a split of the spectral lines regardless of whether the gas is in a laser or not. Thus the Zeeman effect also has nothing directly to do with creating the lasing lines except that both LCP and RCP components will be present if the gain curves overlap at the cavity mode location. The lasing optical frequencies - the RCP and LCP lines - are then spread apart slightly by mode pulling depending on where they are relative to their respective gain centers affecting both the difference frequency and its magnitude.
For example, in a typical HP/Agilent 5517B laser, the Zeeman shift of the two neon gain curves may total around 1 GHz. But mode pulling shifts the RCP and LCP components of a single split lasing mode apart by only 2.2 MHz (1 part in 455). This is the difference (or beat or split or REF) frequency. Increasing the magnetic field would increase the Zeeman shift and corresponding difference frequency proportionally until the point where the neon gain curves are so far apart that the gain for a mode centered between them is below threshold and there is no lasing at all.
HP-5517C HeNe Laser Tube Mode Sweep Versus Magnetic Field has several plots for very weak to normal field strengths. More on this can be found in the section: Axial Zeeman Experiments Using Variable Magnetic Field. Until the numrical discrpencies have been resolved, consider this as a qualitative demonstration of the effect of changing the magnetic field, but don't take the exact field values too seriously.
Unresolved issues:
Now back to your regularly scheduled programming. :)
The first is by far the easiest and generally sufficient except for extremely dedicated academic types. :)
There is a YouTube video claiming to show Zeeman splitting of the HeNe lasing line using stone knives and bear skins, well almost. :) However, while what is shown is due to the Zeeman effect, it is NOT the split lasing line, but rather rogue lasing modes on the tails of the split gain curves. See Longitudinal Zeeman splitting with a HeNe laser.
Rogue modes originate when the laser cavity is so long that both the (normally desired) Zeeman-split lasing mode AND 1 or more normal modes fit within the gain bandwidth of the split neon gain curve as shown in: Avoiding Unwanted "Rogue" Modes in Zeeman-Split HeNe Lasers. Where the tube is too long and thus its longitudinal mode spacing (also known as FSR) is too small, there will be significant gain one or more mode spacings away from the desired Zeeman-split lasing mode. Where this is above the lasing threshold, additional rogue lasing modes will appear. And with a large enough magnetic field, the central split mode may be totally non-existent (though that is unlikely with the field from a home-built electromagnet).
The laser tube in the video appears to be 9-10 inches long, with a mode spacing of around 700 MHz. Thus the split rings in the video are actually the rogue outer modes in the second diagram, above, about 1.4 GHz apart. Like the central Zeeman-split mode, they would also be left and right circular polarized. I estimate the FSR of the etalon in the video to be around 10 Ghz, consistent with the split being around 10 percent of the ring spacing. It all makes sense! Too bad. :( :)
I do not have any full life data for HP/Agilent lasers. But in general it would appear that the trend is for output power to decline, usually accompanied by an increase in REF frequency, especially near end-of-life. But not always - and that's somewhat of a mystery. Since HP/Agilent lasers have mirrors whose alignment is permanently fixed by glass-glass contact, changes in alignment cannot be the cause for either power decline or REF frequency increase. (There is one exception to this: If the tube experiences a significant physical shock, due to the slight difference in bore and mirror diamaeter, the mirrors may move relative to the bore changing alignment.) However, the gas-fill He:Ne ratio and pressure do change with use affecting gain. And, optics can degrade increasing cavity loss. The strength of the permanent magnetic field also affects output power and REF frequency, but if it were to change at all, it would most likely decrease, resulting in an opposite effect compared to what's in evidence here. The one electronic adjustment in these lasers - the tube temperature set-point - has a modest effect, but rarely changes on its own. Cavity length is fixed for any given laser tube but is included here as one of the relevant parameters. And while there may be a very small effect from external losses in the optics due to changes in beam expander alignment and contamination over the years, it's almost always insignificant.
Parameter Output REF
(Increases) Power Frequency Comments
----------------------------------------------------------------------------
Gain Increases Decreases Declines with use
Cavity Loss Decreases Increases Mostly due to mirror transmission
Magnetic Field Decreases Increases Generally higher for higher REF
Temperature Increases Decreases Set by electronic adjustment
Cavity Length Unchanged Decreases 127, 102, or 100 mm
External Loss Decreases Unchanged Windows, beam expander, WPs, etc.
Of course, these aren't independent. For example, effective gain increases with temperature and tubes with a longer cavity also generally have higher gain due to a longer discharge length.
The following table shows how the output power and REF/split frequency changes with use for some recent vintage (2004-2006) Agilent 5517B lasers:
Laser Label PWR Warm PWR W-PWR/ Label REF Warm REF W-REF/
ID (µW) (µW) L-PWR (MHz) (MHz) L-REF
-----------------------------------------------------------------
1 355 395 1.11 2.27 2.17 0.96
2 580 525 0.91 2.29 2.39 1.04
5 560 510 0.91 2.29 2.29 1.00
6 607 525 0.86 2.22 2.27 1.02
7 580 516 0.89 2.22 2.32 1.05
9 680 530 0.78 2.20 2.47 1.12
12 251 222 0.88 2.19 2.17 0.99
13 560 305 0.54 2.23 2.75 1.23
16 654 519 0.79 2.23 2.51 1.13
17 660 574 0.87 2.24 2.35 1.05
18 628 387 0.62 2.24 2.67 1.19
20 625 543 0.89 2.24 2.36 1.05
21 642 436 0.68 2.26 2.44 1.08
23 580 500 0.86 2.22 2.35 1.06
25 628 536 0.85 2.25 2.31 1.03
27 692 637 0.84 2.28 2.35 1.05
28 568 495 0.87 2.26 2.36 1.04
29 562 473 0.84 2.30 2.43 1.06
30 564 512 0.91 2.27 2.30 1.01
31 457 375 0.82 2.29 2.49 1.09
(The missing laser IDs did not have label values for output power and/or REF Frequency.)
Though the actual number of on-time hours is not known, these lasers were all labeled by Agilent with the original output power and REF frequency. ID #1 is believed to either be grossly mislabeled, or an unused spare, which thus has like-new performance. All the others have probably been run for thousands of hours. There appears to be no correlation between the manufacturing date and relative performance compared to the label values, so the date is not listed here. And for this reason, it is believed that these lasers were replaced after a specific number of on-time hours or for performance reasons, not when a fab shut down. Several have REF frequencies that have climbed to near or beyond the upper limit for the 5517B (2.4 MHz) and may pulled from service for that reason. (However, they would now meet all specifications for the 5517C!) But on some others, while the output power has declined significantly, the REF frequency is unchanged or even lower than the labeled value. So, it's possible that there is indeed more than one mechanism accounting for the changes in output power and REF frequency. With gain and cavity loss being critical, each of these will degrade at different rates.
After reconnecting the control PCB so the laser would lock normally, I used a hair dryer to confirm that heating the overall tube also affected the split frequency and output power with the cavity length maintained constant by the feedback loop. However, the correlation between split frequency and output power was not quite the same as when using the internal heater. So, the temperature and pressure of the gas inside the tube is a factor but not the whole story. But both the split frequency and output power changes could be caused by increased Doppler broadening at higher temperatures. This would both reduce the mode pulling effect thought to cause the Zeeman mode splitting (and thus reduce the split frequency), and increase the gain at the "valley" between the Zeeman-split gain curves (and thus increase the output power).
The typical 5517 laser goes through a total of about 70 mode sweep cycles from power-on until READY starts flashing. But because the heater resistance is only sampled every 25 seconds or so, the temperature can overshoot by a large amount. The controller then supposedly backs off and zeros in on the set-point. Supposedly.
But HP lasers from day 1 (the original 5500A around 1969 even before it had an official model number) and Agilent lasers to the present have all had both a QWP and HWP, with the basic design unchanged over more than 40 years. Further, both waveplates are in mounts that allow the tilt of each one to be adjusted around one of its principle axes. Why? Some possibilities as to the reasons for this more complex setup are as follows:
Real HeNe laser tubes exhibit some small random amount of birefringence both from the fine structure of the mirror coatings as well as from unavoidable geometric asymmetry in their construction. Without a magnetic field or explicit polarization control measures such as a Brewster window or plate, these tend to lock the polarization of the longitudinal modes to a fixed orientation about the tubes optical axis, and 90 degrees from it. Adjacent modes will almost always be orthogonally polarized. In a HeNe laser with an axial magnetic field such as one from HP/Agilent, this will result in the Zeeman modes being slightly (or not so slightly) elliptically polarized rather than pure circularly polarized. So, the orientation of the QWP will matter and only certain orientations (2 or 4) will convert these to orthogonal linearly polarized modes. But the resulting linealry polarized modes in general won't be aligned with the system's X and Y axes, so the HWP is then required to rotate them to match. (The magnetic field will also never be perfectly symmetric or uniform, though I don't know whether this is ever a significant factor in affecting the mode polarization.)
The adjustable tilt allows the exact retardation of each waveplate to be altered slightly. I find it somewhat hard to believe that the reason is simply to be able to use cheapo waveplates that might not always be exactly 1/4 or 1/2 wave! However, this explanation can't be entirely discounted since the accuracy of the retardaion is critical to producing F1 and F2 modes that are purely linear and precisely orthogonally as required for the metrology applications. And, the waveplates are made from what looks like optical-grade mica whose discrete layers preclude the ability to select the exact retardation by controlling thickness. And whether mica waveplates were originally selected based on low cost or zero order or temperature stability or being very thin to avoid significantly shifting the beam when tilted or being what the designers had laying around is not known either. But there might be another reason for this "feature" - namely to further compensate for some deficiency in the modes coming out of the tube, again related to deviation from being purely circular. Or something. ;)
Thus, there are in fact 4 degrees of freedom, though clearly the tilt has a much smaller effect than the rotation. And without a full understanding (possibly including hairy math!), it's difficult to really come up with an adjustment procedure that will work in general. That's the bad news.
The good news is that from experience, swapping the entire waveplate assembly between HP/Agilent tubes is likely to result in acceptable performance without any adjustments other than making sure that the overall orientation is the same. Therefore, it would appear that these errors are usually small. But it does seem that it is sometimes necessary to touch up the tilt of the QWP in order to produce the best mode purity in addition to optimizing the waveplate orientations. Adjusting a waveplate assembly from scratch (random orientation and tilt) may be much more challenging though since it would first be necessary to match the orientations of each of the waveplates to a waveplate assembly to at least allow the laser to lock. Without a stable output, going further would be virtually impossible.
In addition, while the theory predicts that rotating the output polarization by 90 degrees when installing a 5517B tube in a 5501B would simply require rotating the HWP by 45 degrees, this doesn't appear to guarantee optimal F1/F2 orthogonality even with a fair amount of fiddling. Swapping in the waveplate assembly from a (dead) 5501B tube and then touching up its adjustments sometimes works better, though not perfectly (or at least not easily in finite time). Thus, these must not be set up to be pure 1/4 wave or 1/2 wave for the typical tube.
So much for hand waving. :-) One might think that a good source for information relating to Hewlett Packard's metrology laser technology would be patents. But so far, patent searches have turned up almost nothing of relevance. If anyone has knowledge or references related to the waveplate issue or anything else of relevance, please contact me via the Sci.Electronics.Repair FAQ Email Links Page.
The only actual waveplate fine tuning procedure to maximize the F1/F2 separation and achieve optimal orientation I've ever seen for one of the early HP-5500 lasers with built-in interferometer optics. The goal then is to minimize any return of the component that is supposed to be transmitted. It simply says something like:
I believe the reference to "the outer plate" is with respect to the 1/4 waveplate. But there is no explanation as to why this procedure is the way it is. And they don't mention the tilt adjustment on the 1/2 waveplate at all.
I did tests of WavePlate (WP) sets from three HP-5517 lasers using a linearly polarized HeNe laser. The results are as follows:
<--------- Orientation ---------->
ID# Laser Input QWP HWP Output
---------------------------------------------------------
1 5517C 9mm +20° +20° +32.5° +45°
2 5517B 6mm -20° -20° +12.5° +45°
3 5517B 6mm +20° +20° +32.5° +45°
(I doubt that the specific type of 5517 laser or its beam diameter makes any difference. The accuracy of my measurements on orientation is within +/-2° for input and +/-1° for output, though the latter at +45° is probably quite precise based on theory.)
The Input is the orientation required for the polarized HeNe laser to produce a pure linearly polarized beam at the output of the WPs, and thus also of the orientation of the optical axes of the 1/4 WP. (Only if aligned with the slow or fast axis of a 1/4 WP will the polarization remain linear, a requirement for these tests.) As expected, the output orientation is the same in all cases since the desired output will be rotated by 45° to align with the X and Y axes. This is a result of the conversion from circular to linear polarization by the 1/4 WP, at 45° with respect to its optical axes. The orientation of the 1/2 WP was inferred from the transfer function from input to output. Now, it's quite possible that the orientation of the 1/4 WP was chosen at random and not actually determined for the specific tube unless it was found that the adjusting the 1/2 WP alone would not meet specifications. However, I have found that it may be necessary to iteratively tweak the 1/4 WP and 1/2 WP to achieve best purety of the F1/F2 modes - the same result could not be obtained by adjusting only the 1/2 WP. Indeed, with a genuine original HP/Agilent tube that has its WPs optimally adjusted, the purety of the F1/F2 modes is nearly perfect. So, perhaps they start at 0° (or +/-20°!) for the 1/4 WP orientation and go from there.
With the Zeeman tube producing a pure two-frequency left and right circularly polarized beam, all of these WP assemblies would result in pure orthogonal linearly polarized outputs oriented along the X and Y axes. This was confirmed by placing a true (non-HP) QWP in the linearly polarized HeNe laser's beam at +45 or -45 degrees to produce pure right and left circularly polarized inputs to the HP WPs. The results were linearly polarized outputs oriented along X and Y. In neither WP assembly, was there any indication that the tilt of either WP was adjusted for other than pure 1/4 or 1/2 wave as the extinction when set for optimal linearly polarized outputs was nearly perfect.
It should be noted that even original unmodified HP/Agilent lasers don't necessarily produce perfection, though it is quite close. In addition, the X and Y extinction ratios may not be the same indicating some residual elliptical polarization in at least one of the Zeeman modes.
Interestingly, Excel 1001A/B/F two-frequency HeNe lasers - which are physically and functionally equivalent to some HP/Agilent lasers but are based on HeNe laser tubes of conventional design - only use a single QWP in a similar rotate/tilt mount. Thus, Excel apparently determined this to be acceptable - or never saw an HP laser! From my limited experience, it may be nearly as good as the HP approach, but possibly not quite the same. On two sample lasers (a 1001B and 1001F), I was unable to achieve quite the same degree of orthogonality with the single waveplate as HP/Agilent lasers usually have. But that was after I had aligned the laser tube mirrors for maximum output power, which may not have been where they were set originally. (Mirror alignment provides a way of trading off output power and split/REF frequency by varying cavity loss.)
Laser |<------- Wavelength ------>| Optical Type Vacuum Air Frequency -------------------------------------------------------- 5500A 632.99???? nm 632.81???? nm 473.6122?? THz 5500B 632.99???? nm 632.81???? nm 473.6122?? THz 5500C 632.99???? nm 632.81???? nm 473.6122?? THz 5501A 632.99???? nm 632.81???? nm 473.6122?? THz 5501B 632.991372 nm 632.816759 nm 473.612234 THz 5517A 632.991372 nm 632.816759 nm 473.612234 THz 5517B 632.991372 nm 632.816759 nm 473.612234 THz 5517BL 632.991372 nm 632.816759 nm 473.612234 THz 5517C 632.991354 nm 632.816741 nm 473.612248 THz 5517D 632.991354 nm 632.816741 nm 473.612248 THz 5517DL 632.991354 nm 632.816741 nm 473.612248 THz 5517E 632.99???? nm 632.81???? nm 473.6122?? THz 5517FL 632.991354 nm 632.816741 nm 473.612248 THz 5517G 632.99???? nm 632.81???? nm 473.6122?? THz 5518A 632.991372 nm 632.816759 nm 473.612234 THz 5519A 632.991354 nm 632.816741 nm 473.612248 THz 5519B 632.991354 nm 632.816741 nm 473.612248 THz
I'm not sure what accounts for the two different wavelengths (and thus optical frequencies) among these lasers. There are no obvious physical differences to account for it. The tubes, beam samplers, and relevant portions of the control electronics are all identical. So, it's possible there was a change in isotopic gas-fill or pressure or something else between 5517A/5517B/5518A/5501B lasers and those that came after them. The difference of approximately 12 MHz is still way lower than the commercial-grade error spec of +/-0.1 ppm (roughly +/-47 MHz), so it really doesn't matter. For the Military-grade lasers, the exact optical frequency is measured and included in the calibration report. But a report for one laser I saw had the optical frequency over 10 MHz away from the spec'd value anyhow. My contact at NIST doesn't even know whether it's an actual change in wavelength/optical frequency or simply an upgrade to the calibration in the measurement electronics! Or a change in the speed of light. :)
I've compared the optical frequency of multiple 5517 and 5501B lasers and have found no evidence of any real difference, let alone one averaging 12 MHz as shown in the specifications above. Here are some data. These lasers are listed in more or less the order in which they were tested:
Locked REF/ Balanced
Laser Laser Output Split Frequency
ID Type Power Freq. Difference Notes/Comments
-------------------------------------------------------------------------------
1 5517B 660 µW 2.3 MHz -2.30 MHz Faulty beam sampler was replaced
2-0 5517B 480 µW 2.4 MHz -1.44 MHz Laser 2 with its (new) beam sampler
2-1 " " " " " " -1.35 MHz Laser 2 with beam sampler 1
2-2 " " " " " " -6.75 MHz Laser 2 with beam sampler 2
3 5501B 450 µW 1.9 MHz 0.00 Mhz Other lasers referenced to Laser 3
4 5517E 120 µW 6.3 MHz -2.10 MHz Only laser with Type II Control PCB
5 5517D 120 µW 3.6 MHz -8.25 MHz
6 5517C 260 µW 2.7 MHz -15.60 MHz Tube run at 4.0 mA (not 3.5 mA)
7 5517C 240 µW 2.9 MHz -9.00 MHz " "
8 5517C 210 µW 2.7 MHz -8.58 MHz " "
9-0 5517D 80 µW 3.7 MHz -9.10 MHz Laser 9 with its beam sampler
9-1 " " " " " " -7.10 MHz Laser 9 with beam sampler 1
9-2 " " " " " " -11.10 MHz Laser 9 with beam sampler 2
10-0 5517A 550 µW 1.7 MHz -7.63 MHz Laser 10 with its beam sampler
10-1 " " " " " " -2.13 MHz Laser 10 with beam sampler 1
10-2 " " " " " " -9.33 MHz Laser 10 with beam sampler 2
10-3 " " " " " " -3.44 MHz Laser 10 with beam sampler 3
11 5501B 220 µW 2.1 MHz -10.70 Mhz
12 5501B 150 µW 1.8 MHz -8.65 Mhz Tube run at 4.0 mA (not 3.5 mA)
13 5501A 100 µW 2.0 MHz -23.48 MHz Really high mileage!
14 5501A 50 µW 2.1 MHz -25.42 MHz " "
15 5501A 35 µW 2.0 MHz -34.47 MHz " "
16 5517A 410 µW 1.6 MHz -0.34 MHz Tube run at 4.0 mA (not 3.5 mA)
17 5517D 405 µW 3.7 MHz +75 MHz Rebuilt with non-HP/Agilent tube
Diagram of Test Setup for HP/Agilent Laser Optical Frequency Comparison shows the way these measurements were made. Photo of Test Setup for HP/Agilent Laser Optical Frequency Comparison shows how ugly it really is, but the scope on the right is displaying the actual beat signal of a pair of 5517Bs, around 1.5 MHz. For later measurements - and to be able to display that nice clean scope trace - two external HeNe laser power supplies (the white boxes and the Variac) were added to eliminate the FM introduced by the switching noise/ripple of the internal HeNe laser power supplies. But for the tests of the difference frequency, they weren't used and wouldn't have affected the results since the high frequency FM would be averaged out. Older HP lasers were really terrible in this regard; newer HP/Agilent lasers with the VMI 217 or VMI 373 power supplies are much better since they include built-in ripple reduction. But my linear supplies have a beefed up filter capacitor bank along with the normal active current regulation.
The 5517 lasers were enclosed in standard cases (non-vented for all except the 5517E) and allowed to reach equilibrium (2 hours minimum). (Removing the cover may significantly change the optical frequency once equilibrium is reached.) The lasers in the photo don't have any cloths on, but, well, that's another matter! :) The 5501As were run naked so that the "Photodiode Offset" adjustment could be performed. (More below.) It might be best to do this via a hole in the cover as they do drift significantly with the cover in place. But I wasn't *that* enthusiastic!
Laser 3, the first and healthiest 5501B to be tested, was chosen arbitrarily to be the reference for optical frequency. The Balanced Frequency Difference is the frequency of the mid-point between F1 and F2 for the subject (ID) laser minus the frequency of the mid-point between F1 and F2 for Laser 3. There's still a +/-1 MHz or more uncertainty due to variations in the specific lock point of the two lasers being compared during any given run.
The three 5501As are very well used and weak, but it was easy to obtain a beat between them and the 5517A, Laser 10 (which happened to be the 5517 laser tested and thus conveniently left in place!). 5501As have a "Photodiode Offset" adjustment, which moves the lasing line on the split gain curve. It's the square pot (R4) on the Lock Reference PCB, clockwise rotation decreases optical frequency. I could have set them all to have a 0 MHz difference frequency, but this would have resulted in grossly unbalanced mode amplitudes for these high mileage lasers. So, they were adjusted according to the HP procedure - maximizing the F1-F2 REF frequency, which centers the lasing line between the split neon gain curve. Before doing this, it wasn't even possible to see the difference frequency with Laser 13 likely because it was too high for my instrumentation. This was probably because parts of Laser 13 had been swapped, including the tube, without making any adjustments. The Photodiode Offset adjustments on the other 2 5501As were quite close to optimal. However, this is a single turn pot which adjusts the mode ratio from 1:2 to 3:2, and thus the optical frequency varies significantly with very small changes in its position - possibly 50 MHz or more end-to-end. Going only by the REF frequency - which isn't perfectly stable - it's quite likely that there will be an uncertainty of 5 or 10 MHz. So, best would be to adjust this pot (make it a 10 turn pot!) through a hole in the cover after the laser has reached thermal equilibrium. And if what you want is a precise optical frequency and don't mind some possible mode imbalance, adjust it with respect to a reference laser like an iodine stabilized HeNe laser instead of for maximum REF frequency! :) And, although the optical frequency changes with the cover installed, the Photodiode Offset adjustment could still be optimal if the change is due to the tube temperature, and thus the gas pressure increasing. That would still maintain the same mode balance.
Laser #17 was rebuilt by another company other than HP/Agilent. (I'm also not at liberty to reveal the company name.) It had its original laser tube removed from the magnet/optics assembly and replaced with a laser tube that is not from Agilent. This explains the large offset in optical frequency, which could result from any number of factors but is quite consistent with a tube uning non-isotopically pure neon. (See below.) The offset is probably of little practical consequence as long as it remains relatively constant.
Aside from laser #17, based on these data, there is really no consistent difference in average optical frequency based on laser type and if anything, it goes the wrong way! And note the change resulting from the swap of the beam sampler. Beam Sampler 1 was originally on Laser 2 and was resulting in the optical frequency dancing around, then swapped with Beam Sampler 2 which resulted in a large frequency offset, then with a third Beam Sampler which was finally well behaved and now remains in that laser. I have no reason to suspect anything is wrong with either Beam Sampler 1 or 2 and did test them for basic functionality with a voltage source and polarizer. All beam sampler assemblies I've checked regardless of what laser they came from have exactly the same part number though it's possible that the optics inside differ in some subtle way depending on laser type. There are at least two versions of the housing - one with a small aperture for 6 mm optics and another with a large aperture for 9 mm optics, but beyond that I don't know of any differences. Lasers 3 through 7 definitely have their original beam samplers. Though I don't have minimum specs for the 5517E, Laser 4 is probably relatively high mileage. And lasers 5 and 9 are high mileage as evidenced by their low (below spec) output power. Even though the output power of Lasers 6, 7, and 8 is well within spec, they are also definitely high mileage lasers being extremely slow start and unable to run on the normal 3.5 mA discharge current. This in itself shouldn't have a large effect on optical frequency unless the tube actually runs hotter (in which case the optical frequency should increase, more below). But the lock point temperature adjustment has not been changed on these lasers, so the equilibrium bore temperature should be similar to that of the others though the equilibrium laser tube envelope and laser temperature will be slightly higher.
So, the actual optical frequency may be dominated by the amount of use (number of hours on the tube) which also tends to correlate with a decline in output power. This may overwhelm any real or fictitious optical frequency offset found in the specifications. While I don't know what the original output power was for most of these lasers, those with 400 µW or more start very quickly or instantly and are likely relatively young (usage-wise). Laser 1 is known to have been taken out of service due to a bad LCD in the beam sampler, so it could have seen relatively little use.
There has been research showing that the neon gain center frequency tends to decline with use due to a drop in tube pressure and other factors. Helium has an effect on lasing center frequency of about +22 MHz/Torr, so a loss of He due to gas entrapment on the tube walls or cathode, leading to a drop in its partial pressure, can easily account for these large frequency differences. (Loss of He due to diffusion through the tube walls would also result in a decline in its partial pressure, but this loss mechanism should be minimal.) Major factors include:
Cause Sensitivity Comments
------------------------------------------------------------------------------
Helium Pressure +22 MHz/Torr Pressure of He decreases with use
Neon Pressure -25 MHz/Torr Pressure of Ne decreases with use
Neon Isotopic Ratio +10 MHz/% of 22Ne Ratio of 22Ne:20Ne Decreases with use
Temperature +280 kHz/°C Affected by specific lock point
Both He and Ne partial pressures descrease over the life of the tube but because the fill ratio is between 5:1 to 9:1 of He:Ne, the decrease in He pressure dominates and a frequency drift downward of several MHz/year is quite reasonable. If not filled with a pure Ne isotope, the Ne isotope ratio also will change slightly as the 22Ne will be trapped at a slightly a higher rate than 20Ne. Note the strong dependence on the Ne isotope ratio, a 1 GHz range! So, just over a 1 percent change in the ratio at the time of manufacture could account for the 12 MHz difference in nominal frequency specifications. Natural neon contains approximately 9.25% 22Ne. And, for any given measurement, there is uncertainty in the actual lock point as the laser warms up but that's probably only a maximum of +/-1 MHz or so. The 280 kHz/°C is for a tube about 8-1/2 inches long - similar in length to most of the HP/Agilent tubes. However, note that for the HP/Agilent lasers, the temperature of the tube envelope is not controlled, only the mirror spacing rod for the 5501B and 5517s, and not at all for the 5501A. So, the temperature of the interior of the laser may have a significant impact on the optical frequency. This differs from many other stabilized HeNe lasers where a large portion of the tube is wrapped in a heater.
The above has been distilled from the paper: "Frequency stability measurements on polarization-stabilized He-Ne lasers", T. M. Niebauer, James E. Faller, H. M. Godwin, John L. Hall, and R. L. Barger, Applied Optics, vol. 27, no. 7, 1 April 1988, pp. 1285-1289. However, a later paper states the contribution from the Ne isotope ratio as being 8.75 MHz rather than 10 MHz per percent of 22Ne.
So, it's quite possible that any differences in the optical frequency of these lasers when they were new is totally swamped by changes due to use. For example, if a laser has been run 24/7 for 3 years (middle age for these lasers!), its optical frequency could have gone down by 10 to 15 MHz due to the decline in gas pressure and isotope ratio changes. But the ultimate conclusions may be that (1) it's not worthwhile to assume anything about the nominal optical frequency on used HP/Agilent lasers, but if the optical frequency can be measured (or compared to that of a new laser), (2) the frequency shift may be a means of estimating how many hours or years they've been on! :-)
Despite all these potential source of variability, for an application requiring an accurate stable optical frequency reference like calibration of a wavementer, a healthy 5517 laser (any version) is probably a better choice than a laboratory stabilized HeNe laser like a Spectra-Physics 117A. The reason is that the design of the 5517 inherently locks to a balanced mode state, with no adjustments and little in the electronics to drift with age to change this. Lasers like the SP-117A have separate photodiodes and pre-amps for the two mode signals as well internal adjustments that can affect the lock point. Furthermore, the optical frequency specifications of all HP/Agilent lasers are known (even if there is an unexplained discrepancy of 12 MHz going from the 5517B to 5517C). This is not the case for many other stabilized HeNe lasers. And, if needed, the laser can easily be packed up and sent to NIST or elsewhere to have its optical frequency measured precisely without fear of it changing either from a few bumps during shipment, or over time if turned on periodically rather being run 24/7. I wouldn't recommend other HP lasers like the 5501B simply because healthy ones are becoming harder and harder to find. And the 5501A uses a different locking design which is similar to that of the other (non-HP/Agilent) lasers. However, a healthy 5518A or 5519A/B would also be suitable, using the same design as the 5517.
Also see the section: Comparing the Optical Frequencies.
So, if you're salivating for an HP/Agilent laser and can't live without one, they cost somewhere between $8,000 and $12,000 new depending on options! Used (or "previously owned" - which would be classier!) HP/Agilent lasers can be had much cheaper but caveat emptor. The only ones most people can afford for personal use would be found on eBay. But most of these lasers that end up being resold are taken out of service because they have an end-of-life tube. Interferometry lasers used for metrology are often run 24/7 from the day they are installed until they die. Even though the lifetime of the special HeNe laser tubes used in these lasers may be 50,000 hours, that's still only about 6-1/4 years. And guess where they then end up? :) If the seller hasn't powered the laser head (or doesn't admit to it) and lists the laser "as-is" with no returns, chances are excellent that it will serve as a nice doorstop but not much else, at least not without some effort. Unfortunately, except for the 5519A which plugs into a standard wall socket, these lasers require +/-15 VDC for power with a Military-style connector that you won't find at Radio Shack. It's easy to "hot wire" power from inside, but unless the seller is familiar with this sort of thing or has the mating power supply and cable, it may be better to just get a DOA warranty in writing and accept that you may have to pay shipping both ways if the laser is only good as a doorstop.
These lasers also show up at surplus dealers but they tend to ask higher prices than would be considered acceptable for basic tinkering and many seem content to simply have the laser gathering dust than to let it go for a realistic price if untested. But I've also heard of at least one instance where such a laser was found at a garage sale. That price was almost certainly right!
Also, note that when looking for lasers like this on eBay or elsewhere, the clothes these lasers wear are of little importance. Newer Agilent OEM 5517 lasers (which are mostly what show up as late model surplus in 2010) tend to have a thin cheaply made gold-ish (alodined) aluminum shroud rather than the beige or gray two-piece case of most older HP lasers. It has a feeble attempt at a rubber gasket all around to seal it but this really doesn't work well and only makes reassembly a royal pain. (The gasket is easily removed if desired.) What's inside is the same, though lasers built after around 2003 will likely have the Type II Control PCB with mostly SMT components rather than the older Type I Control PCB with all through-hole components, but they are otherwise identical and functionally equivalent. And in the trivial triviality department, Agilent's only concession to style seems to be in the color of the front and back plates: Beige for 5517Bs, silver for 5517Cs, and gold for 5517Ds! :) But this is only true of some samples and there doesn't appear to be any way to predict which ones.
Regardless of who is selling the laser, if they are able to power it, the three most important things to ask of them would be:
"Yes" and "a few seconds or less" means the HeNe laser tube and power supply are probably good and happy working together. A laser that takes awhile to start may still be fully functional, but it can be annoying to wait 10 minutes for a beam, and associated equipment may expect the laser to be ready within a fixed amount of time. However, even a minute or could still be acceptable.
Note that the 5501B is the one exception where the laser tube isn't turned on until near the end of the locking process. Thus a beam that doesn't appear immediately on the 5501B is a feature, not a bug. :) This was probably done in the design to minimize the DC current consumption during warmup to be backward compatible with the 5501A. My preference, where this isn't an issue, is to bypass the transistor switch and enable the tube immediately as with all the other lasers.
A "yes" answer to this question alone is usually sufficient to confirm proper operation with usable power for many purposes. However, cold start to READY on solid may be over 10 minutes - even up to 20 minutes for a few lasers like some versions of the 5517E, 5517FL, or 5517G, and some other 5517s using the Newest Type III PCB, but these lasers are almost non-existent surplus. Nearly all those found on eBay even in 2012 should come ready in the typical 4 minutes. A slightly longer time like 5 minutes is of no consequence, but if a pre-2000 laser takes several minutes longer, it's probably very low power and requires the extra time for the power to increase enough as the tube warms up for the laser to be convinced there is enough power available. However, such a laser could still be useful. It's also possible that someone twiddled the one internal adjustment - the temperature set-point trim-pot - without knowing what they were doing!
However, some types of data processing systems like the HP-5508A Measurement Display will produce a hard error if the laser takes more than 10 minutes to become ready. (Power cycling the 5508A once the laser is ready will generally get around this even if it powers the laser, as it's likely to become ready much quicker once warmed up.)
For many applications, much less than spec'd minimum power is quite sufficient. Even if the seller is unable to measure the output power, as long as READY comes on solid, it is probably at least 80 µW for all the lasers except the 5501A, which will lock at much lower power - down to 40 µW or less. Even this may be sufficient for a single axis system.
Where the laser passes these tests, it will probably be more than adequate for an experimental, demo, test, educational, or research system which doesn't have many measurement axes and isn't run continuously for years. However, before considering such a laser for installation in a semiconductor wafer stepper producing next generation multi-core processors, many additional tests would need to be performed to determine its present health and life expectancy. In some installations, the laser is swapped out after a fixed number of hours, like fluorescent lamps! :) While in others, they are replaced at the point where their output power or REF frequency have changed by a certain percentage or fail to meet HP/Agilent specs. Either of these approaches makes sense where the cost of down time is extremely high. So, for example, even though they may start instantly, run reliably, and have decent output power much greater than the HP/Agilent minimum, if their REF frequency is found to be at or above the range for that model laser, they may be flagged for replacement during preventive maintenance. (REF frequency tends to increase with use and is related to the decline in output power.) An example would be a 5517B outputting 500 µW with a REF frequency of 2.5 MHz. (The spec'd REF frequency range of a 5517B is 1.9 to 2.4 MHz.) Details are beyond the scope of this presentation, but there may be a writeup in the future. Stay tuned. However, most tools can tolerate a REF frequency way above the spec'd maximum, and certainly for the vast majority of exprimenters, this is totally irrelevant. A laser outputting 500 µW is generally darn healthy. :)
(If anyone has an HP/Agilent laser with a record of the output power and REF frequency when new either from measurements, the label, or original paperwork, and what they are now, and if possible, an estimate of how much it has been run, please contact me via the Sci.Electronics.Repair FAQ Email Links Page. This will aid in my attempt to more accurately estimate previous use and life expectancy for these lasers.)
Even if the laser plays dead, it could just be a bad HeNe laser power supply brick, or something else that's easily and (relatively) inexpensively repaired. Or, it could be *really* slow start. And on really old lasers, there's a service switch inside - someone may have left it in the wrong position! :)
For details, see the section: Common Problems with HP/Agilent 5517 Lasers.
And even if the HP/Agilent laser tube is certifiably dead, it is possible to install an inexpensive barcode scanner tube in its place that results in a usable system, at least for experimentation or demos. This isn't for the casual user, but if you're up to a modest challenge and have some basic mechanical, electronic, and optical skills, see the section: Installing a Common HeNe Laser Tube in an HP-5517 or 5501B. You can then say you built a $10,000 laser for $3.87. :)
For more information on alternatives to purchasing new HP/Agilent lasers and critical issues in their selection and testing, see the companion document: Considerations in Evaluating Used or Rebuilt Hewlett Packard/Agilent Metrology Lasers.
And a note on shipping: While these are not really particularly fragile as lasers go, a sharp physical shock can misalign the laser tube internally resulting in the power declining dramatically, possibly to zero. (Most common lasers would be destroyed by this treatment.) Careful tapping with a wood block may get at least some of the power back, but this is for the advanced course. :) So, impress upon the seller the importance of careful packing with at least 2 inches of non-collapsible padding on all sides in a sturdy oversize box.
And, if you do come across one of these lasers at a garage sale, just splurge, pay the $2 they're asking, and take the risk. :-)
So, in short, the laser itself won't function any better if run continuously compared to being turned on at most 90 minutes before needed as long as it doesn't affect the environment in such a way as to change the calibration. (90 minutes is HP/Agilent's spec for warmup to full accuracy on an unvented laser, only 45 minutes on one with forced air cooling). And for less critical applications, simply waiting until READY comes on solid may be adequate. It should be possible to test for the overall effect by making a measurements of a known length in each axis when the laser comes READY, after 90 (or 45) minutes, and after 24 hours. If any differences found are acceptable, there is nothing to be gained by continuous operation.
Where the laser might be used for a few hours a week, as in a diamond turning machine at a custom optics house, this should effectively extend the life of the laser to infinity.
(The following deals with retrofitting systems using 5501A or 5501B lasers. For really old systems using 5500A/B/C lasers, a few more issues are present since the 5505A Measurement Display is more tightly coupled to the laser and somewhat more is involved to keep it happy.)
Replacing 5501B laser with 5517 laser:
The preferred approach is to install a more modern 5517 laser in place of a 5501A or 5501B. 5517 lasers are still in production and used working units are also readily available at very reasonable cost. This may require no modifications to the 5517 laser so if a replacement is required at a later time, it can be a drop-in.
Only a few relatively minor differences need to be accommodated to substitute a 5517B for a 5501A or 5501B. With a bit of resourcefulness, the total cost (excluding the laser and labor) for this type of conversion will be under $100 in most cases:
5517 5501
(Male) (Female) Function
------------------------------------------
1. J1-J,K,M,T J2-A +15 VDC
1. J1-J,K,M,T J2-A +15 VDC
2. J1-L J2-B -15 VDC
3. J1-G,H,S J2-D Power Ground
J1-B " "
4. J1-F J1-C REF
5. J1-E J1-D ~REF
This assumes that the adapter cable is short (e.g., a foot or so) so that shielding isn't an issue. I'd still recommend twisting the REF and ~REF wires together with a pitch of about an inch. Once completed, double check with a multimeter!
Installing 5517 tube in 5501B body:
If the electronics in the 5501B are in good condition or can be repaired, it's also possible to install a new or used healthy 5517B tube assembly. This is simpler from the point of view of the user and preserves the external appearance of the Tool so Field Service doesn't get upset about unauthorized modifications. :) (A 5517C or 5517D could also be used if the higher REF frequency isn't an issue.) Physically and electrically, 5517B/C/D tubes are drop-in replacements for the 5501B as long as the beam diameter is the same. (Else, the beam expander will need to be replaced.) So that leaves F1/F2 orientation, REF frequency, and the temperature set-point adjustment:
If you don't want to risk messing with the waveplate assembly on the good tube, transfer the one from the dead 5501B. It's settings may be close enough but slight adjustments of both the orientation and tilt may be required.
Note that even if everything is done perfectly a couple of factors may conspire to slightly degrade the REF/MEAS signal quality compared to that of a newer 5517:
However, neither of these is likely affect system performance in any detectable way, only the appearance of the signals on an oscilloscope. So it's probably not worth losing too much sleep over them.
Modifying 5517 laser:
Where the 5501B electronics are faulty, a usable 5501B is not available, or it is desired to upgrade to a 5517 laser but the F1/F2 orientation must match that of the 5501B, there is a hybrid approach that will also work. Two additional things need to be done beyond what's required to use a 5517 laser without this change:
It may be possible to do either of these same modifications to the Newer Control PCBs but it's likely to be more complex as the relevant signals may not exist outside of an FPGA. Just find a 5517B/C/D laser with a dead laser tube but good Type I Control PCB to modify. Note that swapping signals to the LCD device itself is NOT equivalent and will actually have no effect.
And doing any or all of this will void the warranty. :) But the result will be functionally indistinguishable from an original 5501B laser.
However, even a tube deemed to be dead by Agilent due to low power or an inability to stay lit, may often be made usable for many applications (especially where only 1 or 2 measurement axes are required), for a test or educational system, or as an emergency spare, with at most some relatively minor low cost modifications to the laser, or possibly even simply an adjustment. But if the output power is so low that the beam actually disappears periodically while warming up, there won't even be a beat signal and such a tube is only good as a high tech paperweight with built-in magnetic paper clip holder. :)
Assuming the tube is usable, except for late model Agilent 5517 lasers based on Type II or Type III Control PCBs, all of these lasers are very serviceable as far as the electronics are concerned. Pre-2004 lasers - most of what's found surplus even in 2010 - will almost certainly have the older Type I Control PCBs. (This is definitely true for pre-2000 lasers.) For these, even most of the HP house-numbered ICs have standard equivalents available from major electronics distributors, and none of the other electronic parts are special. Operation and service manuals are available which include detailed adjustment and troubleshooting information and complete schematics. And parts units can be obtained on eBay at low cost. Except for a blown fuse of my own doing, dried up electrolytic capacitors on really old lasers, a blown line driver chip, and bad REF photodiode, I've yet to see an Type I Control PCB with any serious problems including defective proprietary HP ICs. However, on a 5501B, the heater driver transistors and main fuses were blown as a result of dried up electronic capacitors on the Connector PCB. So, for 5501B lasers, it's probably good preventive maintenence to replace all 4 large electrolytic capacitors on the connector PCB on a laser more than 10 or 15 years old as a precaution. This is the only situation I know of where a high ESR/low uF capacitor will result in actual damage to other components in these lasers. For more on the 5517 laser in general and the Type II PCBs in particular, see the section: HP/Agilent 5517 Laser Construction.
Now, if you're independently wealthy and would like to have Agilent repair your laser, I've heard that an evaluation is about $500. Essentially, they confirm that it's an HP or Agilent laser and then tell you how much it will cost to repair, if they are willing to repair it at all. For a single failure, the cost is a flat rate between $1,500 and $2,000, but the evaluation fee will be applied toward that, thank goodness. :) A "single failure" probably includes a blown fuse, broken resistor, dried up capacitor, or bad IC. I don't know whether something like a degraded LCD in the beam sampler or blown HeNe laser power supply would qualify as a single failure, or if two dried up capacitors would be charged (no pun....) for separately. And, it's almost certain that if you read the fine print, the flat rate would exclude a weak or dead HeNe laser tube that required replacing the tube assembly even though it is technically a "single failure". In that case Agilent would simply return the laser after collecting their evaluation fee.
For amusement, go to Find-A-Part: Agilent's Test and Measurement Parts Catalog and enter a laser model like "5517D". If you're not independently wealthy, you better be sitting down when viewing the prices. For example, (in 2009) the cover is $344, the Control PCB is $1075 (not known which version), the HeNe laser power supply is $496, and a small screw is a bargain at $1.24 each. However, prices for the operation and service manuals are not totally ridiculous - $28.44 for the 5517A and $42.67 for the 5517B/C. But the exact parts available for each model laser seem to be somewhat random and forget about even being able to order a new tube assembly (or parts). They are listed as: "Not orderable, contact Agilent for repair service". Right. :-)
And before doing something silly, getting inside HP/Agilent lasers is trivial. On the large lasers (5517A, 5518A, 5518A/B) it's just a matter of removing the 4 tiny screws on top and gently levering up the cover using a knife blade. On the small lasers (5501A/B, 5517B/C/D), rotate the front turret so the large hole is at the bottom. That will expose a slotted head screw - a 1/4 turn fastener. Push in and rotate 1/4 turn counter-clockwise and the front plate will pop off. The covers or shroud can then be removed. The only reason I've gone to this level of detail is that I had an academic type ask me if that screw was for tuning the laser frequency! :)
Also see the section: Common Problems with HP/Agilent 5517 Lasers (which applies to other lasers like the 5501B as well). For operation and service manuals, see the section: Additional HP/Agilent Resources.
I also have backup copies of the same PDFs at Sam's Bakcup of Agilent Laser and Optics System Design Manual.
Note that the file for Chapter 7Y does not exist on the Agilent Web site and I haven't been able to find it elsewhere. I've left the link in place should it magically appear.
Part# Description
------------------------------------------------------------------------------
05500-60025 5500A/B/C to 5505A cable
05505-60048 Rack Mount Kit for 5505A
05508-60021 Remote Control Unit (5528A)
5500A Laser Transducer (w/interferometer, optical receiver, 0.4 m/s)
5500B Laser Transducer (w/interferometer, optical receiver, 0.4 m/s)
5500C Laser Transducer (w/optical receiver, 0.4 m/s)
5060-0049 Extender Board, 15 pin
5060-0630 Extender Board, 22 pin
5501A Laser Transducer (0.4 m/s)
5501B Laser Transducer (0.4 m/s)
5505A Measurement Display (5526A)
K01-5505A Extender Board (XA-14), 52 pin
5507A Electronics
5508A Measurement Display (5528A)
5510A Automatic Compensator (5525A/5526A)
H01-5510A High Accuracy Automatic Compensator (5525A/5526A)
K15-5510A Multiplexer for 5510A
5517A Laser Transducer (0.4 m/s)
5517B/BL Laser Transducer (0.5 m/s)
5517C Laser Transducer (0.7 m/s)
5517D Laser Transducer (1.0 m/s)
5517DL Laser Transducer (1.1 m/s)
5517E Laser Transducer (1.6 m/s)
5517EL Laser Transducer (1.77 m/s)
5517F Laser Transducer (1.7 m/s)
5517FL Laser Transducer (2.15 m/s)
5517G/GL Laser Transducer (2.2 m/s)
5518A Laser Transducer (w/optical receiver, 0.4 m/s), <SN2532A02139)
5518A Laser Transducer (w/optical receiver, 0.453 m/s, >=SN2532A02139)
5519A Laser Transducer (w/optical receiver, 0.7 m/s)
5519B Laser Transducer (w/optical receiver, 1.0 m/s)
5525A Laser Measurement System
5526A Laser Measurement System
5527A/B Laser Position Transducer System
5528A Laser Measurement System
5529A Dynamic Calibrator
5530 Dynamic Calibrator
9211-1586 Transit Case for 5500A/B/C
9211-1587 Transit Case for 5505A
9211-1738 Transit Case for 5510A
10550A Reflector
10550B Retroreflector
10551A Plane Mirror Convertor
10552A Resolution Extender
10555A Remote Interferometer
10556A Retroreflector
10557A Turning Mirror
10558A Beam Bender
10559A Reflector Mount
10560A Barometer
10562A Single Beam Interferometer
10563A Material Temperature Sensor
H01-10563A High Accuracy Material Temperature Sensor
10564A Air Temperature Sensor
10565A Remote Interferometer
10567A Dual Beam Splitter
10579A Straightness Adapter (Resolution Extender And Optics)
10580A Laser Tripod (5500C)
10581A Plane Mirror Converter (5526A)
10585A Metrology Program Package (5526A)
10690A Straightness Interferometer
10691A Straightness Interferometer
10692A Penta-Prism
10692B Optical Square
10693A Vertical Straightness Adapter
10700A 33% Beam splitter
10700B 4% Beam splitter
10700C 15% Beam splitter
10701A 50% Beam splitter
10702A Linear Interferometer
10703A Linear Retroreflector
10704A Single Beam Retroreflector
10705A Single Beam Interferometer
10705A-080 Fiber Optic Receiver Adapter
10706A Plane Mirror Interferometer
10706A-080 Fiber Optic Receiver Adapter
10706B High Stability Plane Mirror Interferometer
10707A Beam Bender
10708A Power Supply (May Not Apply)
10710A/B Adjustable Base (Small, Beam Bender, etc.)
10711A/B Adjustable Base (Large, Linear Interferometer, etc.)
10713B/C/D Cube Corner
10715A Differential Interferometer (DI)
10715A-001 DI (turned configuration)
10716A High Resolution Plane Mirror Interferometer (PMI)
10716A-001 High Resolution PMI (turned configuration)
10717A Wavelength Tracker
10719A One-Axis Differential Interferometer (DI)
10719A-C02 One-Axis DI (low thermal drift)
10721A Two Axis Differential Interferometer
10721A-C02 Two Axis DI (low thermal drift)
10722A Plane Mirror Converter (5501A)
10723A High Stability Adapter
10724A Plane Mirror Reflector
10725A 50% Beam Splitter
10725B 4% Beam Splitter
10725C 15% Beam Splitter
10726A Beam Bender
10728A Plane Mirror
10735A Three-Axis Interferometer
10736A Three-Axis Interferometer
10736A-001 Three-Axis Interferometer/Beam Bender
10737L/R Compact Three-Axis Interferometer
10740A Coupler (5501A)
10741A Laser Transducer Interface (10740A card)
10742A Laser Transducer Counter (10740A card)
10743A Extender Board (10740A)
10745A HP-IB Interface (10740A card)
10746A Binary Interface (10740A card)
10751A/B Air Sensor (5528A)
10751-60209 Laser Interferometer Cable
10753A Laser Tripod (5518A)
10755A Compensation Interface
10756A Manual Compensator
10757A/B/C Material Temperature Sensor (5528A)
10757-60306 Laser Interferometer Cable
10759A Foot Spacing Kit
10760A Counter (10740A card)
10761A Multiplier (10740A card)
10762A Comparator (10740A card)
10763A English/Metric Output (10740A card)
10764A/B Fast Pulse Converter (10740A card)
10764-60005 Laser Interferometer Cable Assembly
10766A Linear Interferometer
10767A Linear Retroreflector
10767B Lightweight Retroreflector
10768A Diagonal Measurement Kit
10769A Beam Steering Mirror
10770A Angular Interferometer
10771A Angular Reflector
10772A Turning Mirror
10773A Flatness Mirror
10774A Short Range Straightness Optics
10775A Long Range Straightness Optics
10776A Straightness Accessory Kit
10777A Optical Square
10778A/B/C Laser Power Cable (5501A/B)
10779A/B/C Reference Cable (5501A/B)
10780A/B/C Optical Receiver (Free Space)
10780F/U Optical Receiver (Fiber-Coupled)
10781A Pulse Converter
10781-60003 Cable assembly
10782A Service Kit without Laser Assembly (5501A)
10782AOP001 Laser Assembly (5501A) only
10783A Numeric Display
10784A Interferometer Base
10785A Height Adjuster and Post
10786A Linear Measurement Transit Case
10787A Straightness And Squareness Transit Case
10790A/B/C Receiver Cable (4 pin BNC plug both ends)
10790A-C10 Laser Interferometer Cable
10791A/B/C Laser Head Cable (5517, spade lugs, 4 pin BNC REF)
10793A/B/C Laser Head Cable (5517A to 5507A and 5518A to 5508A)
10880A/B/C Receiver Cable (4 pin BNC to LEMO)
10881A/B/C Laser Head Cable (5517, DIN for power, LEMO for REF)
10881D/E/F Laser Head Cable (5517, spade lugs for power, LEMO for REF)
10882A/B/C Laser Head Cable (5519A/B To 10887P)
10883A/B/C Laser Head Cable (5518A, DIN for power, LEMO to 10887A)
10884A Power Supply (5517 lasers, universal switchmode, DIN)
10884B Power Supply (5517 lasers, universal switchmode, DIN)
10885A PC Axis Board
10886A PC Compensation Board
10887A/B PC Calibrator Board (5518A or 5519A/B)
10887P PC Programmable Calibrator Board (5519A/B)
10887-60202 Laser Interferometer Cable
10888A Remote Control
10889A/B PC Servo Axis Board
19895A VME Laser Axis Board
10897B VME Laser Axis Board
10898A VME Dual Laser Axis Board
55280A Linear Measurement Kit
55281A Angular Optics Kit
55282A Flatness Accessory Kit
55283A Straight Measurement Kit
C05-59995A Reference Cable (5501A/B)
C07-59995A Power Cable (5501A/B)
C08-59995A Diagnostic Cable (5501A)
C39-59995A Laser Head Cable (5517A to 5507A and 5518A to 5508A, 1 meter)
E1203C Precision Beam Translator
E1204C Precision Horizontal Beam Bender
E1705A Fiber Optic Cable
E1706A Remote Sensor
E1207C Precision Vertical Beam Bender
E1208C 33% Bare Beam Splitter
E1208D 40% Bare Beam Splitter
E1208E 50% Bare Beam Splitter
E1208F 66% Bare Beam Splitter
E1208G 60% Bare Beam Splitter
E1250A/B High Performance Receiver Cable
E1251A/B High Performance Laser Head Cable
E1708A Remote Dynamic Receiver
E1709A Remote High Performance Optical Receiver
E1713A Scale Servo Axis Board for E1720A.
E1826E/F/G One-Axis Plane Mirror Interferometer
E1827A Two-Axis Vertical Beam Interferometer
E1833C 15% Bare Beam Splitter
E1833E 33% Bare Beam Splitter
E1833G 50% Bare Beam Splitter
E1833J 67% Bare Beam Splitter
E1833M 100% Bare Beam Splitter (Beam Bender)
E1837A Two-Axis Vertical Beam Interferometer
ET-319283 Interferometer Adapter Cable
N1211A Fiber AOM Laser Head (15 to 17 MHz split frequency)
N1211A-001 RoC Cable, 6.7 m
N1212A/B Remote Optical Combiner, 6 mm/9 mm
N1250A/B High Performance Optical Receiver Cable
Z4399A Three-Axis Interferometer
Z4420B Five-Axis Interferometer
Z4379G-A08 Polarizing Beam Splitter (with fiber adapters?)
To get inside the 5500A/B requires removing 4 screws - 1 on each side front and back. There may be an interlock that turns off the laser tube when the cover is removed.
The HeNe laser tube in the 5500A/B is generally similar to the one in the 5500C and 5501A, but isn't quite identical and thus is not interchangeable, at least not without some work. The original patent for the 5500A/B laser tube is: U.S. Patent #3,771,066: Gas Laser. The most notable obvious differences between the 5500A/B tube and the one in 5500C and 5501A are in the PZT connector at the rear which is a ring (rather than a center terminal) that allows the waste beam from the HR mirror to escape, and the optics assembly at the front of the tube assembly which only has the beam expander - the waveplates are mounted externally (though strictly speaking these aren't part of the tube itself). And the glass tube is simply clamped to the mounting feet, which are not part of the tube assembly.
Rather than using a portion of the main beam for feedback, there's a shielded can with a photodiode behind a Quarter WavePlate (QWP) and motor driven rotating polarizer that samples the waste beam from the back of the tube. The photodiode signal is used in a feedback loop to lock the laser so the modes are of equal amplitude. (See: U.S. Patent #3,701,042: D.C. Motor Circuit for Rotating a Polarizer and Providing a Detector Synchronizer Signal for a Laser Stabilizing System.) Ironically, this is actually closer in function to the LCD optical switch of the 5501B and later lasers, than the polarizing beam samplers of the 5500C and 5501A that followed the 5500A. Since the 5500C/5501A tube has no waste beam exiting the laser tube, duplicating this function would be a bit of a challenge.
The 5500A is in the same size case as that of the 5500C. The main difference between the 5500A and 5500C is what's at the front of the laser. The 5500A has interferometer optics and detectors for both REF and MEAS within the case. The 5500C has two channels of optical receivers but no interferometer optics. However, it was possible to install linear interferometer optics inside the 5500C to give it 5500A functionality.
The 5500A is also unique among HP lasers since it is the only one with a run-time (hour) meter!
There are photos of a 5500A in the Laser Equipment Gallery (Version 2.49 or higher) under "Hewlett Packard HeNe Lasers".
For several original articles introducing HP's interferometer-based measurement system using the 5500A, see the Hewlett Packard Journal, August 1970.
Also see Dave Meier's HP Laser Interferometer Evolution Page which includes a links to the early HP catalog pages.
I have a 5500A laser (see gallery pages, above) which appears to be from around 1970 based on the date code found on a 74H10 TTL IC in the optical receiver. Except for the shape of the beam expander mount and color of the ballast resistor cover, my 5500A appears identical to the laser shown on the last page of the August 1970 HP Journal. An external HeNe laser power supply was used to perform initial tested before being connected to a 5505A Measurement Display. The laser tube starts and runs flawlessly with a raw output (after the beam expander but before the waveplates) of at least 370 µW and possibly as high as 450 µW. (The power varies with temperature as the tube warms up if not feedback stabilized and I didn't run it long enough by itself to determine the actual maximum power.) Even the low end of 370 µW would be considered excellent power for a much newer 5501A tube. The output power of the laser is between 106 and 150 µW (again depending on the temperature as it's not locked). If the locked output is anywhere near the higher end of this range, then it's basically like it was when it was last serviced. There is a note inside the laser saying: "120 µW August 1978". Perhaps the tube was also replaced at that time. The reason for the large difference between tube output power and laser output power is that the waveplates cut the power by 15 to 20 percent, and the internal interferometer optics suck up approximately half of the remainder since most of the F1 frequency component doesn't exit the laser.
When first attached to a 5505A, the laser powered up and locked instantly, and within a couple minutes, I was able to make sub-micron measurements! But, then at some point while my back was turned, the original HeNe laser power supply inside the laser head failed. Hard to believe! Not like the thing has probably been turned on for the first time in 20+ years! :) I don't know if the failure was in the two transistor driver, or inside the potted HV module, which is beautifully made in clear semi-flexible plastic with no obvious damage, the remains of which (after salvaging the HV wire) are shown in HP-5500A HeNe Laser High Voltage Assembly. But there could be a shorted turn in the inverter transformer or a capacitor breaking down. The driver transistors passed ohmmeter tests and were getting equally warm, but the output was only going to around 1 kV and then dropping to 0 V, never lighting the tube. So, I replaced it with a small brick power supply from a barcode scanner, installed inside the original aluminum can to preserve authenticity. Unless one knew exactly where to look, there would be no way to tell that it wasn't totally original.
One thing that's probably only of curiosity value is that both the HeNe laser HV power supply and the PZT HV power supply are driven from a common oscillator which must be running for the PZT tuning to work. Without tuning, the 5505A readout may still function, but the RESET button will keep flashing. Newer versions of the 5500C, as well as the 5501A use independent self-oscillating inverters in ugly bricks made of hard tan potting compound for these two power supplies. The earliest 5500Cs are probably similar to the 5500A.
It's extremely easy to align the interferometer with my home-built authentic replica of the retroreflector mount shown in the 1970 HP Journal article. As long as it adjusted so the return beam enters the optical receiver aperture or even the tiny alignment holes in the laser head turret, the system is happy.
And here is the genuine imitation authentic setup hot off my time machine:
More information and photos from early HP manuals and brochures, and elsewhere can be found at Dave Meier's HP Laser Interferometer Evolution Page.
The cable wiring is given in the next section since it is the same for the 5500A and 5500C.
There are photos of a 5500C in the Laser Equipment Gallery (Version 2.48 or higher) under "Hewlett Packard/Agilent HeNe Lasers".
Also see Dave Meier's HP Laser Interferometer Evolution Page which includes a link to the early HP catalog pages.
To get inside the 5500C requires removing 4 screws - 1 on each side front and back. There may be an interlock that turns off the laser tube when the cover is removed.
The 5500C uses a HeNe laser tube with PZT tuning that appears identical to the one in the 5501A, though the part number differs. See the section below on the 5501A for detailed descriptions and and photos. (The 5500A has a very similar, though not identical tube. See the description and patent reference in the previous section.) The beam sampler for the feedback stabilization is of the common modern polarizing beam-splitter variety with the control loop driving the PZT of the laser tube to adjust cavity length. But, unlike the 5501A which only requires DC power supplies, the 5500 requires the mating 5505A Measurement Display to even turn on and stabilize since its HeNe laser power supply and PZT power supply are controlled by the 5505A. Although the HeNe laser power supply could be run open loop with a variable DC voltage, this would not provide current regulation. However, the PZT power supply of later 5500Cs which appears to be a potted module inside more potting, can be used as a stand-alone PZT, PMT, or other variable HV low current power supply since its output is fairly linear with respect to input from 0 to 15 V, which is multiplied approximately by somewhere between 100 and 200 to produce the output voltage. (Although I have not seen it specifically stated, the PZT power supply appears to be capable of more than 2 kV based on the 5501A schematics.) Both HV Control and PZT Control are really just the power input to a self oscillating inverter. (Very early versions of the 5500C and the 5500A have the inverter transformers and other high voltage components potted inside metal cans with the driver circuitry on separate PCBs fed from a common oscillator.)
The pinout for the self contained PZT power supply module is:
5500A/B/C and 5505A connector pinout
Pin Function
------------------------
A Gnd
B DOPPLER (A)
C +5V
D LOCK (A)
E HV CON
F REF TRIP
G -15V
H BEAM AL
J PZT MON
K REF (A)
L GND
M REF (B)
N DOPPLER (B)
P NC
R LOCK (B)
S LASER I
T +15V
U PZT CON
If constructing your own cable, the wires to pins B and N should be shielded twisted pair, shield to pin A, and the wires to pins K and M should be shielded twisted pair, shield to pin L. The shield probably isn't critial for relatively short cables, but use the twisted pair. Size the voltage (+5, +15, -15) and Gnd wires to handle a couple amps. HV Control will also need to supply some current.
On most (probably later) versions, the HeNe laser tube can be powered with a variable DC power supply. The two connections are:
If the cover is removed, there may be an interlock (microswitch) in series with the power to the tube. So that would need to be defeated. The useful range for the tube to turn on is from around 15 V to 30 V between these pins, with the tube operating at the optimal current at around 20 to 25 V. But it's best to start at 0 V and work up. ;-) As soon as the tube starts, reduce current to just above where it stays lit without flickering. To safely measure tube current, put a 1K resistor between the cathode terminal (on the side of the tube) and its connecting wire. Then measure voltage (V) across the resistor. The current is then I = V(mA). The optimal current (when new at least) is usually marked on the tube and is typically in the 3 to 3.5 mA range. If the optimal current isn't labeled, a rule of thumb is to set the current 0.5 mA above the point where the discharge drops out and starts flickering, or 3 mA, whichever is higher.
However, it appears as though very early 5500Cs may have a transistor in the circuitry leading to the HeNe laser power supply. So, if testing it as described above results in no output beam and current being drawn from your DC power supply, it will be necessary to go inside and connect directly to the HeNe laser power supply brick on the PCB under the laser.
The 5501B is a functional replacement for the 5501A. Locking of the 5501B typically takes 5 to 9 minutes compared to 10 seconds or so for the 5501A, but this is of no consequence for machines that are run for hours or years. In terms of optical characteristics, and power requirements and reference signals (including connector pinouts), they are equivalent. However, the 5501B lacks the Diagnostic (J3) connector of the 5501A, so other system components may not be happy and some substitutes may need to be provided. Going the other way doesn't have this issue, but if a 5501A is installed in place of a 5501B, it may be necessary to press the Retune button from time-to-time whereas there is no such button or need on the 5501B! This may be anywhere from a few hours to never, but it would be a good idea to do this periodically at convenient times between measurement runs, at least until the system has reached thermal equilibrium. Performing a Retune cycle does not compromise the accuracy in any way. Once the Retune LED goes out, it's ready to go again. Even from a cold start, a laser may go 12 hours or more without requiring a Retune. After that, once a day may be more than sufficient.
Operation and service manuals for the HP-5501A and HP-5501B may be found on the Hewlett Packard/Agilent Metrology Laser/Interferomter Page.
Compared to the 5500C, the 5501A is in a much smaller lighter case, similar to the later 5501B and 5517B/C/D lasers. It also has simplified optics and totally different electronics. See Interior of the HP-5501A Laser Head - Left Side and Interior of the HP-5501A Laser Head - Right Side. The HeNe laser tube dominates the interior space in both views. The high voltage piezo driver power supply brick is visible under the magnets at the center of the tube. The HeNe laser power supply brick is underneath the output end of the tube. The piezo driver electronics circuit board at the far right end of the right side view. The optical sensor circuit board is at the far left of the left side view.
HP-5501A Laser Tube Assembly shows a 5501A tube by itself. The naked tube is shown in HP-5501A Laser Tube Removed From Magnet and Output Optics Assembly. The normally enclosed part is really just a very thick-walled fine-ground bore inside an outer glass envelope. A spring (visible through the glass at the left) at the rear holds the PZT, HR mirror, bore, and OC mirror in place. No adjustment is possible. There are distinct multiple spots on the card because the output window is at a slight angle and not AR coated.
See Major Components of HP-5501A HeNe Laser Tube for an official autopsy photo of one that was end-of-life and had it's tip-off broken in shipping. Only minimal sacrifices to the gods of dead lasers were required since it was already deceased. :)
The top photo includes an intact sample of an HP-5501A tube assembly with the waveplates and beam expander. Then below from left to right:
The inset photo at the lower left shows the HR mirror, the two tiny spring contacts that pass through it to the PZT, the PZT disk, and the HR-end of the Zerodur bore.
The inset photo at the lower right shows the OC-end of the Zerodur bore and concave OC mirror (which magnifies the printing on the Fragile sticker).
The Zerodur bore is precision ground at both ends to form the laser resonator with no adjustments.
Both the HeNe laser power supply and piezo power supply run off the -15 VDC power supply. An interlock switch (easily defeated) prevents disables operation with the cover removed. In the 5500A and 5500C, these power supplies are regulated by the 5505A Measurement Display. In the 5501A, the potted power supply bricks have no inputs other than power. Rather, current and voltage regulation are accomplished by controlling the input current. For the HeNe laser power supply in the 5501A, as well as later versions of the 5500C, while the passive HV components are buried in potting compound, the two 2N5192 driver transistors are mounted on the outside of the brick and are replaceable. However, from my experience, when the transistors blow, there is probably a fault in the potted section so replacing them doesn't help, I've successfully replaced the 5501A HeNe laser power supply with a common barcode scanner brick, the Laser Drive model 103-23. This has an input rance of 21 to 31 VDC at less than 0.5 A, and an output of 1.1 to 1.5 kV at 3.5 mA (fixed). The 3.5 mA is a bit higher than the labeled current on most 5501A tubes, but seems to be acceptable and actually beneficial for some high mileage tubes that like to run at a slightly higher current. But, adjustable versions of these supplies are readily available. I connected the supply between the HV Control (white/green wire) and -15 VDC (purple wire) with the pot set fully CCW (max current). This assures that the 5501A current regulator will not attempt to compete with the brick's internal regulator. However, with some HeNe laser power supplies, it may be possible to use the 5501A's regulator to *reduce* the current in a stable manner. This is left as an exercise for the student as it may not work in general.
The output of the laser tube is passed through a Quarter WavePlate (QWP) to convert the circular polarization to orthogonal linear polarization components, and then through a Half WavePlate (HWP) to rotate the linear polarization by an arbitrary, but fixed angle to line the two linearly polarized components up with subsequent optics. These waveplates are adjustable with respect to orientation around the optical axis of the laser as expected. But the angle of each waveplate along one of its principle axes with respect to the optical axis of the laser is also adjustable - presumably to optimize the QWP or HWP performance, but could also be required to adjust them so they are not quite perfect to compensate for imperfect polarization purity in the raw beam - or something. :) They are both very thin and may be zero order waveplates, possibly made of optical grade mica. The beam is then expanded and collimated and passed through an angled partially reflecting plate located just beyond the collimating lens on the laser tube assembly. This deflects about 20 percent of the beam to a polarizing beamsplitter which sends each component to its own photosensor to provide the frequency control feedback. A control loop uses these signals to adjust the PZT, and thus resonator length, so that the two signals are of equal amplitude. The difference of the two signals is the frequency/phase reference.
The laser stabilization control algorithm is actually dirt simple: The voltages from the photodiodes corresponding to the two polarization components are compared in an integrator which maintains the PZT voltage at a level so they are equal. (There is an adjustment to compensate for slight differences in amplitude resulting from beamsplitter ratio and photodiode sensitivity.) While crude and simple to implement, this approach is adequate to achieve the needed stability. The electronic reference signal is derived from the slight residual difference frequency present in one of the polarization components.
While the spacer rod has a very low coefficient of thermal expansion, it isn't exactly zero, so as the system heats up (over hours), the cavity length will still change slightly. Eventually, the PZT voltage may be unable to compensate. The PZT voltage is compared with fixed upper and lower limits which are well within the range over which locking is assured. When either limit is passed, the "Tune Fault" flag is set turning on the "Retune" LED and asserting the "Retune_Status" signal. The laser may be retuned via a pushbutton or external TTL signal). This clamps the PZT control voltage at its lowest value for a short time and then releases it to ramp up to the lock point. Requiring external intervention (whether manually or by computer) assures that a measurement will never be made when the laser isn't stable, nor will one in progress be interrupted due to the laser relocking unexpectedly.
When testing, continuous monitoring of the amplitudes of the F1 and F2 modes is recommended, or at least periodic checking to assure that they are still approximately equal. All of these lasers show some drift in both total power (which tends to increase) and the relative mode amplitudes. The latter is likely due to etalons effects from several uncoated optical surfaces between the tube's output mirror and the F1/F2 photodiodes. Error checking in the laser is not very comprehensive, so it's possible for a failure in the locking circuitry to go undetected even though F1 and F2 differ by a large amount. For example, if the integrator is unable to reach the upper or lower detection thresholds, F1/F2 could become very unbalanced without flagging an error.
The 5501A laser head requires +15 VDC and -15 VDC for power. (There is also a +5 VDC pin but it is an output according to the manual.) The two voltages (and common) are all that is needed to operate the laser head but an interlock switch (on the right side at the rear of the case) must be depressed to turn on the laser tube. I haven't yet looked at the output with a photodiode or scanning Fabry-Perot interferometer but after a few seconds, the "Retune" LED goes off, similar to if the "Retune" button is pressed. And then there is a stable reference signal. I have since acquired an operation and service manual for the HP-5501A laser head which confirms the information above.
HP-5501A reference connector J1
See HP 5501A and 5501B Reference and Power Rear Panel Connectors for pin location.
Pin Function Socket View
--------------------------------------------- A
A Accessory +15 VDC fused o
B +15 VDC return D o o B
C Reference (difference) frequency o
D Complement of J1-C C
HP-5501A power connector J2
See HP 5501A and 5501B Reference and Power Rear Panel Connectors for pin location.
Pin Function Socket View
---------------------------------------
A +15 VDC input D o o A
B -15 VDC input
C +5 VDC output (test-point) C o o B
D Power ground
HP-5501A diagnostic connector J3
See HP 5501A Diagnostic Rear Panel Connector for pin location.
Pin Function I/O Comments
------------------------------------------------------------------------------
A +15 VDC TEST O Sample for diagnostics
B -15 VDC TEST O Sample for diagnostics
C +5 VDC TEST O Sample for diagnostics
D SYS COM - Ground/return
E Retune_CMD- I Active low to initiate PZT tune/check cycle.
F Retune_Failure O Active high output indicates failure of PZT
tune/check cycle.
J Retune_Status O Active high when tune/check cycle is in progress.
K Laser_Cur_Err O Active high indicates laser tube current is
outside acceptable limits.
L Error O Logical OR of J3-J, J3-K, and PZT voltage outside
of specifications.
M L I Mon Test O Laser current sample for diagnostics.
N PZT Mon Test O PZT voltage sample for diagnostics.
P Ref OK Status O Active low diagnostic signal indicates laser
is properly tuned.
Like the 5501A, the 5501B also requires only +/-15 VDC to power up. There is no case interlock on most of these, but some really old versions had one to disable the laser tube from being powered if the covers were removed, and a "Service" switch to override this. :) Both of the switches have long since been eliminated, though the PCB pads and wiring for them are still present, but bypassed. It's worth removing both switches on lasers that have them and adding the required jumper diagonally between the center pads that are closest together on both switches. (Do NOT just add the jumper - the switches must be removed!)
When power is first applied, only the +/-15 VDC power LEDs come on. (The laser tube is not powered.) After 3 to 6 minutes, the "READY" LED begins to flash at about a 1 second rate. After another 1.5 minutes or so, the "Laser On" LED comes on and the beam appears. Finally, a minute or so after that, the READY LED comes on solid and remains that way. Note that this method of turning on the laser only after the temperature set-point has been reached is unique to the 5501B, probably to maintain backward compatibility with respect to the maximum DC current on the +/-15 VDC power supplies. But it creates problems of its own on high mileage lasers. (More on this below.) All the other HP/Agilent lasers turn on the laser with the application of DC power.
Specific times for one test beginning from a cold start at an ambient temperature of about 65 °F were: (min:sec) 3:15, 1:35, and 0:48. The first of the times is called "preheat" and is determined by how long it takes for what HP calls the "laser rod" to reach operating temperature. The laser rod is the large glass bore of the laser tube to which the mirrors are clamped at either end. It thus controls cavity length. The temperature is sensed by disabling the heater drive and measuring the resistance of the heater coil every 25.6 seconds. The warmup is much shorter if the laser is restarted after having been running: 1:00, 1:20, and 0:50. Only after the READY LED is on solid, do the reference signals appear. The 5501B adjusts the cavity length so that the two polarized components of the beam (the Zeeman split longitudinal modes) have equal power. Interestingly, there is only one photodiode sensor which is alternatively switched between beams using a liquid crystal polarization rotator. A sample-and-hold then outputs to the error amplifier of the optical mode control feedback loop. (This is the same scheme used in all later HP/Agilent lasers.)
There are two outputs of about 5 to 6 V p-p (centered about 0 V), 180 degrees out of phase. For this laser, the reference frequency is about 1.80 MHz. There is no need for a "Retune" button as with the PZT based system of the 5501A. Also unlike the 5501A, there are no other signals to or from the 5501B (no large connector), only the +5 VDC output on the power connector, and a fused +15 VDC output on the reference connector.
Although the control board inside the 5501B looks similar to that of the "small" 5517 lasers, it is NOT interchangeable with them as some functions like the heater drive are located on the small "connector PCB" at the back-end of the case, which is also unique to the 5501B.
When swapping tubes in 5501B, the only adjustment that needs to be performed is for the temperature set-point, which is the same for the 5517. See the section: HP/Agilent 5517 Temperature Set-Point Adjustment.
For basic testing to see if the laser tube works at all, there are two ways to force it to come on immediately. Either of these should be done before applying DC power:
The laser should turn on immediately with DC power, though high mileage tubes may take awhile to start. On the 5501B, these tend to be somewhat problematic even if they have decent output power as the power transient when the tube finally starts may reset the state machine on the control PCB. I've seen this on many 5501Bs, so it's not something unique to a single sample. The result is that the laser may require a few extra minutes to finally lock as it repeats portions of the warmup sequence. However, where DC power supply capacity is adequate, I recommend leaving option (1) in place permanently, by soldering the closest two pins of the power transistor together. Warmup will then be similar to that of 5517 lasers.
When done, remove DC power, wait a few seconds for the DC voltages to decay to 0 V, then remove the jumper on the power transistor or move the jumpers to their NORM (far left) position.
There were two problems with the first 5501B I acquired that I had to deal with. The first was that the tube wouldn't stay on stably at the 3.5 mA setting (fixed) of the power supply but works fine at 4 mA. Such a condition is usually due to the tube having been run for a long time, which wouldn't be surprising with a surplus 5501B laser head. Since the existing power supply has no current adjustment, I needed to find a similar size HeNe laser power supply brick (1"x1.5"x4" or smaller) that will run on 15 VDC to replace it that can be set for 4 to 4.5 mA. The tube seemed healthy enough otherwise. I installed one that runs the tube at 4 mA but draws more DC input current than the original, and possibly for that reason, the controller aborts and resets after about 1 second when it turns the laser on. For now, to get around this, I have connected the HeNe laser power supply directly to the raw -15 VDC and added a transistor to drive its enable input when the original laser power turns on. That appeared to work fine. But after replacing the cover, the laser tube wouldn't come on. :( I discovered that it needed the room light to start! I had thought this to be a relatively rare malady for HeNe laser tubes, but more common for neon lamps and glow-tube fluorescent lamp starters. However, it turns out that a decent percentage of HP/Agilent HeNe lasers start more quickly when illuminated. So, there is now a decorative red LED shining on the back of the tube which is lit when the laser is powered. An HeNe laser power supply with a higher starting voltage would probably make this kludge, oops, feature, unnecessary. But no one will ever know about it. :) While many of these higher mileage HP/Agilent lasers can benefit from this addition, since the 5501B turns the laser on and expects it to come on quickly, it is more critical than with the other lasers like the 5517s that really don't care whether the laser is outputting a beam or not, until they actually try to lock. However, in either case, if the laser takes too long to lock, associated equipment like the 5508A Measurement Display may flag it as a failure.
HP-5501B reference connector J1
See HP 5501 Reference and Power Rear Panel Connectors for pin location. This connector is labeled as REFERENCE on the rear panel but shown as J1 in the installation instructions. It is J6 on the Connector PCB schematic.
Pin Function Socket View
--------------------------------------------- A
A Accessory +15 VDC fused o
B +15 VDC return D o o B
C Reference (difference) frequency o
D Complement of J1-C C
HP-5501B power connector J5
See HP 5501 Reference and Power Rear Panel Connectors for pin location. This connector is only labeled as POWER on the rear panel but shown as J2 in the installation instructions. It is J5 on the Connector PCB schematic.
Pin Function Socket View
---------------------------------------
A +15 VDC input D o o A
B -15 VDC input
C +5 VDC output (test-point) C o o B
D Power ground
The HP-5525A was used in the original HP interferometer introduced around 1970 and includes the HP-5505A Measurement Display and the HP-5500A two-frequency HeNe laser head. The 5500A laser has the interferometer optics built-in and thus only requires an external retroreflector (cube-corner) on the moving part to be measured. The HP-5525B upgraded to the 5500C laser head which requires external interferometer optics but allows for two axis measurements (with a pair of 5505As!). The 5526A seems to have added a variety of options and but it's not clear how it really differs from the 5525B.
The 5525A/B and 5526A can be set up in the field with relative ease with a minimum number of individual components and no need for a control computer as its basic functions are built-in to the HP-5505A. It provides for the stand-alone precise measurement of position and velocity. But straightness and angle are not directly supported.
The 5505A implementation of the display function is all done in MSI TTL logic with a pair of 36 bit counter/registers for REF and DOPPLER (same as MEAS for other HP lasers), with a decimal adder/subtractor to generate the result. This is all on multiple PCBs and while there is one labeled "Program", there is no actual microprocessor controlling the system.
The 5525A, 5525B, and 5526 all require the 5505A display but differ in the laser and options. (There may be some minor changes required to convert an older 5505A to be used in a 5526A system.) The following is from the N4MW HP 5526A Documentation Page which also has links to the actual HP catalog pages for each system.
5500Cs have also been showing up with internal linear interferometers like the 5500A. I haven't seen any reference to this as a standard product though. I wonder if they were retrofits for customers who found their original 5525A configuration adequate or whined when their 5500As went bad and wanted an exact replacement.
The HP-5525A/B and HP-5526A are very obsolete, but many are still in use. 5505As show up on eBay, often for next to nothing. To non-interferometer geeks, the set of Nixie tubes is probably more valuable than a working unit! However, being so old, they often have problems, and at least some of the ICs like the Nixie tube drivers are proprietary parts and no longer available.
For info (or lack thereof):
The laser connector on the back of the 5505A is the same type and has the same pinout as that on the 5500A and 5500C heads. The 5508A supplies +/-15 VDC power for the laser head. It also controls both HeNe laser power supply current regulation and PZT laser tuning.
To use the 5505A with a 5500A, all that's required is a 05500-60025 cable and a retroreflector (cube-corner) as shown in Original HP-5525A with HP-5500A, HP-5505A, and Retroreflector - View 1. (Additonal photos can be found in the section above on the 5500A laser.) It's straightforward to make a cable. The connectors are standard and everything is wired 1:1 at both ends. To use the 5505A with the HP-5500C laser also requires external interferometers optics. All of the standard configurations that have separate outgoing and return beams should work.
To use the 5508A with other HP laser heads will require a custom cable and possibly a separate optical receiver which can be any version of the 10780 (A, B, C, F, U). However, some circuitry may need to be added to the 5505A to keep it happy by making it think it still has control of PZT tuning.
FWIW in the "well that's interesting department", here is the board set from another 5505A. This is a rather vintage sample, S/N: 2016A01966, which puts its manufacturing date around 1970:
Slot Name Part # Additional Markings ----------------------------------------------------------------- A1 Analog Board 05505-60001 Series 1920 03L A2 Clock Board 05505-60002 B3 Series 952-2 03F A3 Accumulator Board 05505-60034 Series 1920 01403F A4 Accumulator Board 05505-60034 Series 1920 01403F A5 Adder Board 05505-60005 Series 952 03F A6 Algorithm Board 05505-60006 Series 952 00203F A7 Program Board 05505-60007 Series 2240 23103F A8 Function Board 05505-60058 Series 1920 23903F A9 Multiplier Board 05505-60049 Series 1948 23103F A10 D-Register 05505-60010 03F A11 Display Board 05505-60011 Series 1324 03F A12 Power Suppy Board 05505-60012 Series 1940 01503F
The HP-5528A includes the HP-5508A Measurement Display, an HP-5518A two-frequency HeNe laser head on heavy duty tripod, a variety of interferometer optics, and optional environmental sensors, and other stuff. :) So, this system can be set up in the field with relative ease with a minimum number of individual components and no need for a control computer as its basic functions are built-in to the HP-5508A. It provides for the stand-alone precise measurement of position, velocity, angle, and straightness by using the appropriate interferometer assemblies.
The 5508A implementation of the display function consists of X16 frequency multipliers for REF and MEAS, which are then applied to separate 16 bit up counters. These initiate a non-maskable interrupt to the microcontroller when either exceeds the half way point (the MSB gets set). They are then stopped while separate small "swallow counters" absorb pulses occurring while the interrupt is processed and the position is updated. The microcontroller is kept rather busy, but since it doesn't have all that much else do do, should be quite happy. :)
Although the HP-5528A is considered obsolete by Agilent, it's still very useful and surplus systems or components are now much cheaper. The Agilent 5529A (now superceded by the 5530) Dynamic Calibrator is the replacement for the 5528A. Rather than a dedicated display, it requires a PC (not included). But aside from the slightly higher REF frequency of the 5529A laser head (generally irrelevant in these types of typically slow speed applications), the precision is no better than that of the 5528A.
(Both of these are also accessible directly from Agilent. Search for "5528A".)
The laser connector on the back of the 5508A is the same type and has the same pinout as that on the 5517 and 5518A laser heads. The one "No Connect" pin on the 5517 connector (pin A) is used to drive the MEAS beam level indicator on the front of the 5508A. The meter reading seems to be proportional to the current flowing out of this pin, from an internal +5 VDC source, with approximately -2 mA being full scale. Pins B and C that are also unused on the 5517 lasers now get ~MEAS and MEAS. (They are connected to line drivers on the 5517 lasers but only used for testing.) The 5508A supplies +/-15 VDC power for the 5518A laser head.
Note that the +15 VDC power supply in the 5508A uses remote sensing (pin J) for fine regulation. If the 10793A/B/C or an equivalent cable that directly connects the 5518A and 5508A is used, there is no problem. But if a custom cable is made, "+15 Sense" should be a separate wire run between pin J at both ends. And if a combination of a standard HP cable like a 10791A/B/C and a custom cable is made, the +15 VDC may end up being slightly low due to the uncompensated voltage drop between where all the wires are tied together in the standard cable, likely where it attaches to power. This will then be maintained very close to +15 VDC, but the voltage will be lower at the laser head. This is usually not an issue but something to be aware of should strange problems be encountered. However, if possible, it may be best to disconnect pin J entirely at the 5508A. The +15 VDC output will then rise to about 15.6 VDC at the 5508A which after accounting for typical voltage drops in the wiring, is likely to end up within spec. But this should be confirmed with a voltage measurement at the laser head.
There are several other connectors on the rear of the 5508A for various environmental sensors (temperature, pressure, etc.) and even a remote control. (I'd like to see that!) There is also a IEEE-488/GPIB/HP-IB interface for control and data acquisition.
To use the 5508A with a 5518A, all that's required is a 10793A/B/C cable, which is wired 1:1 at both ends. To use the 5508A with other HP laser heads will require a custom cable, and possibly a separate optical receiver which can be any version of the 10780 (A, B, C, F, U). Except for ~MEAS and MEAS, all signals are wired 1:1 for 5517s. However, I'm not sure whether all versions of the 5508A will work over all velocities with any of these except the 5517A. The 5517B/C/D have REF frequencies, and result in possible MEAS frequencies, that may be too high, at least under some conditions. It is also possible to use the 5508A with the 5501A/B lasers, but the connectors are totally different. For the 5519A/B laser, these same issues apply. The connector of the 5519A/B provides only the REF and MEAS signals as this laser head has an internal power supply that plugs into the AC line. To use the 5508A with 5500A/C lasers would require additional circuitry to provide the HeNe laser current control and PZT tuning.
One way to get around the REF frequency issue is to build a divide-by-two circuit for REF and MEAS that goes between the laser and 5508A. This is simply a dual differential line receiver, a pair of D flip-flops, and a dual differential line driver. Add a switch to select straight through or divide-by-two if desired. Of course, measurements values will now be halved unless a plane mirror interferometer is used, in which its doubling will be exactly offset by the halving in the divider!
The only thing that won't work when using laser heads other than the 5518A without additional effort is the beam level meter on the 5508A front panel, fed from pin A of the laser connector. This seems to require a current of 0 to 2 mA to Ground from an internal +5 VDC source. The test-point on the outside of all 10780 receivers generates a voltage related to signal level, but simple voltage to current converter circuit (1 transistor and a few resistors) is then needed to interface to the meter input. If you're not a purist, this can just be ignored as it is not used anywhere and its only purpose is to aid in optical alignment and confirmation of adequate signal. But the 10780 test-point services the same function.
In summary:
I've attached my 5508A to my measurement test setup and initially have been using a 5517D laser head with it. I'm a bit surprised that this even works with the 5508A as it has a REF frequency almost double the maximum of even the later versions of the 5518A laser. I don't know if it will run at full velocity, but for modest speeds, the readings seem to be fine. But I intend to add the divide-by-two circuit eventually as insurance. I've installed a DB9 disconnect to make this easier. It has the following pinout (for my own reference!):
Pin Function ------------------- 1 +15 VDC 2 Power GND 3 ~MEAS 4 MEAS Return 5 MEAS 6 NC 7 ~REF 8 REF Return 9 REF
The module would have an option for a gain of 1 or 5 and divide ratios of 1, 2, or 4. The higher gain is needed for some late model Agilent 5517s which have the Type III Control PCB and appear to output lower amplitude REF and ~REF signals. So far, I've seen this on a 5517D and 5517E. With these, the 5508A would never acknowledge "LASr UP" even though the laser itself came READY, and would eventually time out with "LASr FAIL" even though my home-built measurement display was perfectly happy.
When warming up, the difference frequency only appears for 5 to 20 percent of the time during mode sweep - only when the Zeeman modes are near equal amplitude on the split neon gain curves. And this percentage tends to be lower for higher REF-frequency lasers. The difference frequency is maximum and the output power is minimum at the center of this region, which is also where it will eventually lock. This is normal behavior for these lasers based on what is shown in Axial Zeeman Split HeNe Laser Mode Behavior. Note that while there may be another longitudinal mode present for part of mode sweep, there will be no beat except from the main pair, and then only when relatively close to being positioned symetrically on the Zeeman-split neon gain curves and only the main F1/F2 mode is present when locked. While other "rogue" modes would not produce any beat signal, they could result in problems in the interferometer and possible transient errors.
The only functional difference among 5517 models (and the laser part of the 5518A and 5519A/B) is in the spec'd range for the REF frequency. With suitable processing electronics, any 5517 that's physically compatible (e.g., same case style and beam diameter) can stand in for any other 5517 subject to the maximum velocity limitation for its REF frequency. The measured displacement, velocity, etc., will be the same. In fact, since the REF frequency tends to increase as the laser is used, it's not unusual for a mid-life 5517B (REF range of 1.9 to 2.4 MHz) to actually meet all 5517C specs (REF range of 2.4 to 3.0 MHz)!
The heart of all these lasers is the HeNe laser tube assembly (henceforth often referred to as simply "the tube"). This consists of the actual glass HeNe laser tube (much more below) mounted inside the Zeeman magnet and a cast or machined structure which also includes the output optics. A typical tube assembly from a 5517B is shown in Tube Assembly Used in HP-5501B and HP-5517B/C/D Two-Frequency HeNe Lasers. Except for nutcases like me, these tube assemblies are considered to be non-repairable as disassembly is virtually impossible. Much more below.
The output optics consists of a beam expander/collimator (the black object just to the right of the central aluminum cylinder) and an additional optical assembly to the right of this whose front and rear halves contain what appear to be AR coated optical quality mica pelicles oriented at slight, but different angles. The front and rear sections can be rotated independently and they were sealed with blue paint once the perfect orientations were found. The two mica (or whatever) pieces of the optics assembly (just after the beam expander) are adjustable waveplates. The first one is a Quarter-Wave Plate (QWP) to convert the circular polarization of the Zeeman split output of the HeNe laser tube to linear polarization and the second one is a Half WavePlate (HWP) to rotate the resulting linearly polarized components to be aligned along the horizontal and vertical axes. These can then be separated out with a polarizing beamsplitter at the detectors.
The locked F1/F2 amplitudes from these lasers are usually not quite equal. This is due in part to the beam sampler not being perfectly non-polarizing, so the horizontal polarization experiences less loss than the vertical polarization. But in addition, although electronics-induced imbalance should be very small, the LCD switch device may not be ideal. And the locked beat frequency varies a bit after locking and does not stay at its maximum value as would be expected if the stabilization was optimal. This is not a quirk of one particular laser I've use for these experiments as I've tested dozens with similar behavior - some worse than others. The cause may be various back-reflections from the multiple optical surfaces outside the laser cavity. On a common cherry-flavored HeNe laser these would not produce any detectable effects, but when dealing with small differences in very large numbers like the optical frequency, they become very evident. So, perhaps these lasers aren't as perfect as we might hope! :)
The HP/Agilent lasers do not employ any sophisticated method of stabilization such as locking the Zeeman beat frequency (which changes slightly depending on where the modes are on the neon gain curve) to a crystal reference. They simply use the amplitudes of two orthogonally polarized signals in an analog feedback circuit as is common with most other stabilized HeNe lasers. However, here, the two polarizations are of the two Zeeman split components of the single oscillating mode rather than two separate longitudinal modes. The error signal is the difference between their amplitudes, which is forced to zero by temperature tuning of the cavity. And, in fact, there is no real need to have the frequency be precisely known or even constant over the long term, as long as it is stable over the short term. More below.
The warmup/locking algorithm is straightforward, though just a bit different than used in many other stabilized lasers. When the laser is first turned on, it is in "Warmup Mode" and the heater, which is wrapped around the internal bore of the laser tube, is driven to reach a fixed temperature (set by the only pot on the electronics PCB). The temperature is sensed by periodically measuring the heater's resistance. This is done by disabling the heater driver, passing a small fixed current through heater wire (for 2.56 seconds out of each 25.6 second period), and storing the resulting voltage in a sample-and-hold. Since the heater wire changes resistance with temperature, this eliminates the need for a separate temperature sensor inside the tube. Once the temperature set-point is reached (the voltage from the pot approaches the voltage on the sample-and-hold), the feedback switches to Optimal Mode and alternately samples the two polarized Zeeman split sub-mode signals with their voltage difference being the error signal in the feedback loop, which is driven to zero by adjusting the temperature, and thus cavity length. In fact, from the relative shapes of the red and blue mode cycles, it can be seen that from about the last half dozen mode cycles till just before locking, the tube is actully steadily cooling rather than heating. With the heater located inside the laser tube, the time from power on to a locked condition is typically only about 4 minutes and should also be less susceptible to ambient conditions. In fact, from the relative shapes of the red and blue mode cycles, it can be seen that during most of the time from power on (a cold start) to lock, the laser tube is heating (about 75 cycles), but it switches to steady cooling (about 6 cycles) just before locking.
Later versions of the 5517 lasers in the small case have a totally redesigned electronics board using surface mount technology with a single Xilinx FPLD containing most of the digital circuitry. I don't know exactly when this changeover took place to this Type II Control PCB, but it appears to be sometime in late 2003. The original Type I Control PCB was becoming rather dated as to parts availability so perhaps Agilent was not simply reinventing the wheel. :) The Type II Control PCB is functionally equivalent to the Type I Control PCB. There is also a much more complex Type III Control PCB, which appears in a few late model lasers, reason unknown. More on the control PCBs and locking schemes below.
Photos of virtually all 5517 laser models may be found in the Laser Equipment Gallery (Version 2.42 or higher) under "Hewlett Packard/Agilent HeNe Lasers.
Converting a 5517A into a 5518A is simply a matter of installing the internal optical receiver PCB and replacing or removing the shutter assembly on the front of the laser. (Replacement with a shutter assembly from a 5518A laser is necessary if there is a desire to use the modified laser for straightness measurements since it has a separate setting for these.) Of all the 5517s, the 5517A (as well as the 5518A and 5519A/B) are the only ones to have a tube assembly that might appear to be of lower manufacturing quality as shown in Tube Assembly Used in HP-5517A, 5518A, and 5519A/B Two-Frequency HeNe Lasers, and is larger than the the tube assembly in the others (and the in the 5501B). But the real reason may be that it is cast with precise locating pegs so that a tube can be swapped without requiring even minimal alignment. The actual glass laser tube is physically the same for all models except the 5517E/F/G, and some later versions of the 5517D/DL, which have shorter tubes. Compare Internal Structure of Hewlett Packard 5517A, 5518A, and 5519A/B Laser Tube Assemblies and Internal Structure of Hewlett Packard 5501B and 5517B/C/D Laser Tube Assemblies. And the 5517A tube assembly is physically interchangeable with the 5518A. For installation in a 5519A/B, there is a small piece of metal in older tubes that needs to be cut away from older 5517A tubes to provide clearance for the 5519's internal DC power supply. More below. Interior of the 5517A Laser Head shows the major laser/optics components of the Hewlett-Packard 5517A laser. The actual glass HeNe laser tube is inside the gray cylindrical housing which also has the cylindrical magnet for Zeeman splitting the HeNe laser lines to create the difference reference frequency in the interferometer application. See the previous sections for more information on these two-frequency lasers.
There's no reason that a version of the 5517A couldn't be made available in the smaller style case but there was probably never any demand.
Photos of virtually all 5517 laser models may be found in the Laser Equipment Gallery (Version 2.42 or higher) under "Hewlett Packard/Agilent HeNe Lasers.
With minor exceptions, any 5517 tube assembly, HeNe laser power supply brick, and control PCB may be installed in any 5517 case, requiring only a single adjustment of the lock temperature set-point to be done if the tube/control PCB combination was changed.
The 5517E/F/G are to only significant variation on the 5517 theme to have been introduced by Agilent. They may have been a last desperate attempt to push the basic Zeeman-split two-frequency laser concept to its limits and compete with the Zygo 7701/2 and other lasers, with their 20 MHz REF/split frequency. Based on how often these have either appeared on eBay :) or from requests for repair, it's fairly obvious that they never caught on. This is likely for two reasons: Due to the limitations of Zeeman-split HeNe physics, the output power spec is significantly lower for the 5517E/F/G (believed to be 65 µW) compared to the other 5517 lasers (and even lower when compared to the Zygo lasers). This significantly limits the number of axes that can be controlled from a single laser, as well as reducing the lifetime of the laser since the power doesn't need to decline very far to be unusable. And at least as significant, even the 5517G doesn't provide a REF/split frequency that comes anywhere near that of the Zygo 7701/2 lasers - the maximum being 7.2 MHz for the 5517G. The only Agilent laser that comes close is the N1211 "Fiber AOM Laser". While the N1211 starts with a laser tube assembly similar to that of the 5517s, its REF frequency is more or less irrelevant as a pair of AOMs shifts the optical frequencies apart by an arbitrary amount. See the section: Notes on the Agilent N1211A Fiber AOM Laser Head.
My 5517E is in the gold case, so perhaps it is only available as an OEM or "military calibrated" product? The tube also has no label, so perhaps it was an early prototype. It runs at 6.3 MHz so the listed range above of 5.5 to 6.5 MHz is just a guess based on the only spec I have for the 5517E - 1.6 m/s maximum velocity. This is much higher than the 5517D's maximum of 4 MHz resulting in over a 50 percent greater velocity measurement capability. The textbook party line had been that axial Zeeman HeNe lasers above about 4 MHz were simply not viable. The requirement for a higher REF/split frequency likely means that the magnetic field is stronger and thus the total extent of the Zeeman-split neon gain curve is wider (necessitating a shorter cavity length with the larger FSR to suppress rogue modes, which as a side effect, also increases REF/split frequency at the expense of output power). An informal measurement of the magnetic field of a 5517E did show it to be 5 to 10 percent stronger than that of any other 5517 laser, though there was quite a bit of variability even for the same models (e.g., 5517B). The tube assembly looks almost identical to all the others except that it is about 0.75" shorter up front beyond the section with the magnet and the glass tube itself is only 6.3" compared to 8" for the long tubes. In fact, in order to work at all at these high REF frequencies require a careful selection of mirror reflectivity, magnetic field, cavity length, and no doubt many other parameters of the laser design. The 5517E/F/G lasers are all operating on the hairy edge of what's possible with Zeeman-split HeNe laser technology, balancing desired high REF/split frequency, acceptable output power, and avoidance of rogue modes in the output. These lasers have a very low spec'd minimum output power (65 µW for the 5517FL, and probably the other 5517E/F/Gs as well) compared to the 5517A/B/C/Ds (180 µW). A new "lively" 5517E may produce 120 µW compared to over 600 µW for many 5517Bs, and even more for some 5517As. A photo of a 5517E I saw recently had a backplate output power of 110 µW and REF frequency of 6.1 MHz. The physical design of the tube squeezes almost every last drop of performance out of it, and even then, the power is low. It's really a last gasp on Agilent's part to retain customers needing higher performance who might have otherwise switched to a Zygo laser with its 20 MHz REF frequency. (Zygo lasers use an AOM to split the frequency rather than the Zeeman effect and has no problems with output power.)
Although I haven't totally dissected a 5517E/F/G tube, the cavity length can be estimated by measuring the longitudinal mode spacing using a Scanning Fabry Perot Interferometer (SFPI). During warmup, two longitudinal modes are present over a portion of the mode sweep cycle, so their spacing compared to the FSR of the SFPI provides a good estimate of the FSR of the laser, and thus its cavity length. For the 5517E and 5517FL, the mode spacing/FSR is approximately 1.5 GHz implying a cavity length of 10 cm (~4 inches). This is about 20 percent shorter than the cavity length of the 5517A/B/C/D lasers. (Though newer 5517Ds may also use the shorter tube.)
Unfortunately my X-ray vision is somewhat limited. Of what is visible, the most obvious difference is that the HR-end of the tube has a metal cap on it instead of the glass with spring affair of all the other 5517 lasers. Presumably, this allows the mirror to be closer to the start of the discharge and increases the discharge length and thus available output power. At the OC-end of the tube, reducing the space between the mirror and discharge escape hole would decrease the cavity length by a sufficient amount, but wouldn't require any major redesign. So, at first (before even seeing the ends) I assumed the design would be similar to that of the 5517A/B/C/D tubes, only shorter. But when I finally was able to remove the beam expander on one for inspection, that was found to not be the case at all, with the mirror spacing rod also terminated with a metal cap holding the OC mirror in place (which might even screw on based on three ribs that could be there to improve grip), but the entire affair is apparently unsupported at the front end. Strange.
Another obvious difference is that the bifilar heater coil has roughly 1/2 the number of turns compared to that of the long tubes, thus covering a smaller section of the mirror spacing rod. (And it is wrapped from approximately the center of the bore back toward the HR - opposite that of the longer tubes.) The shorter heater means that if the lock temperature is the same (as is likely), the number of full mode sweep cycles from ambient to the nominal lock temperature will be smaller. The heater resistance in production short tubes is the same as for the long tubes (about 8 ohms at room temperature), but the heater resistance of my 5517E tube is only around 4 ohms, with an external resistor to make up the difference.
The glass envelope at the front of the tube has the same shape as the others and extends the same distance out, but is filled with a lot of empty space! Well, OK, it isn't quite empty but provides a larger gas reservoir than would otherwise be present if the tube were even shorter. But the internal construction of both mirror mounts as depicted in the diagram is mostly pure speculation. The only way to get accurate information would be to see the engineering drawings or to totally rip apart a tube. And neither is likely to happen!
Several views of a naked 5517E laser are shown in Agilent 5517E Laser Head With Cover Removed.
This 5517E has the most incredibly complicated Control PCB of any HP/Agilent laser I had see before finding this laser, even compared to the Type II Control PCB (see below). (I've since found a similar control PCB on a few other late model 5517 lasers, but they are quire rare.) It includes a SHARC DSP, two Lattice FPLDs, and a lot of other digital circuitry, purpose unknown. They also seem to have gone back to PWM for the heater drive since there is no power transistor on a heatsink, as with the original Type I and the updated Type II Control PCBs. However, that collection of inductors visible in the lower left of the photo may be there to clean up the drive to the heater and remove the high frequency switching noise. Since the locking should be basically the same as for the other lasers, this level of complexity is perplexing unless this particular unit was designed to have much better stability - perhaps the "military calibrated" version. Unfortunately, the Type III Control PCB lacks all the familiar jumpers and the temperature set-point pot, and adds a couple of micro DIP-switches and connectors, purpose also unknown. Aside from the unknowns, everything else is obvious. :)
Here are several closeup photos:
High resolution scans of the front and back of a similar digital control PCB can be found linked from the section: HP/Agilent 5517 Laser Construction.
There is also an additional resistor in series with the tube heater in the wire bundle, apparently as an afterthought since it is part of a cable extension. The heater and resistor each measure just under 5 ohms cold. The heater of other 5517 tubes measures about 8 ohms cold, so at the same current, this shorter tube which must have a shorter heater gets only slightly over half the heater power. I originally thought that this might be why it takes longer to stabilize, but then found that this was true of "normal" 5517 tubes with the fancy Type III Control PCB. That shroud above the tube would make tube swaps much more tedious, as the Control PCB on the opposite side would need to come off to remove it. Then, the cable ties would have to be cut to free the wiring and tube. But at least the connector PCB is identical to the ones in the 5517B/C/D lasers, and even includes the usual appendix - the HeNe laser current adjust pot that is no longer used!
With no label on the tube assembly and that unusual plastic rear cover, for awhile I was suspecting that this might not even an HP/Agilent tube. But that style of glasswork at the back is clearly HP/Agilent even if it does differ slightly from the normal design. And everything else is normal HP/Agilent including the beam expander, HeNe laser power supply wiring/ballast, and the Newest Type II PCB, which, as noted, has turned up on a late model 5517D. However, there were some mica washers under the tube presumably as shims to fine tune the vertical position of the beam, reason unknown. Other 5517E/F/G lasers all seem to have similar washers, which are darn pain to reinstall after removing the tube for inspection. :( :) While there is no manufacturing date on the laser, date codes on the ICs suggest that is from around 2003. Rework in the area of the REF out circuitry may mean this was an early version of the Type III Control PCB. This is not present in my other sample.
The back-end is normally enclosed in a removable plastic cover rather than being filled entirely with the usual rubbery potting compound, most likely because the metal cap on the back of the tube which is also the anode terminal is poking too far out for (shocking) confort.
Another anomoly for this specific sample is a total lack of any label on the tube, so Agilent can deny any knowledge of its existence. :)
And a further note about disassembly: To get this photo required almost totally removing the tube since the rear plastic cap was held in place by three screws with nuts and it would have been almost impossible to replace the nuts without being able to access behind and under the magnet assembly. That's when I discovered the shims. Hopefully, I got them back more or less in the proper locations. Must maintain specifications! :) Later I realized that to gain access to the tube, it would be easier to remove the sheet metal shroud by loosening the 6 setscrews along its bottom edge, leaving the tube assembly attached to the baseplate.
Functionally, the 5517E behaves more or less like the other 5517 lasers. The user LEDs are the same but there are 4 LEDs on the control board that I'm sure provide a wealth of information if one knows how to interpret them. My sample takes over 5 minutes for READY to start flashing. READY also stops flashing once or twice for a couple minutes, before it resumes flashing, and then locks after about 9 minutes. Whether these long times and peculiar flashing behavior are normal or indicate some problem, is also unknown. However, with a similar control PCB and heslthy 5517B tube, the behavior is similarly strange. More on this in the sections: HP/Agilent 5517 Laser Control PCBs and Locking Sequence and Agilent 5517 Laser RS232 Communications.
The tube is somewhat low power compared to what's normal for other 5517s - about 120 µW locked. But I have no specs on 5517E minimum output power, so with the shorter tube and likely stronger magnetic field, that might be acceptable. In fact, the minimum power spec for the 5517FL is only 65 µW, so the 5517E may be similar. And given that other sample with 110 µW as the value when new, 120 µW on a used laser may in fact be absolutely wonderful. :-) Once locked, it's quite stable with minimal drift in REF frequency. Given the huge amount of computation power available, it may count mode sweep cycles instead of using a temperature set-point (or in addition), and might also adapt automagically to a replacement tube - or require a factory upload of tube parameters via one of those unlabeled connectors!
The high REF frequency of 6.3 MHz works fine with my home-built SG-MD1 measurement display, but comes up as "Laser Fail" on a 5508A. This isn't all that surprising since 6.3 MHz is almost twice the maximum REF frequency of the 5518A for which the 5508A was intended. However, the same type of control PCB with a 5517B tube locks fine but also fails keep the the 5508A happy, so it is more likely due to wimpy line drivers or something like that. :)
This montage of Agilent 5517FL Laser and Components shows views of a 5517FL in various stages of disassembly. (Sorry about the photo quality - I do not have this laser.) It had a listed output power of 160 µW and REF/split frequency of 7.12 MHz. (The minimum specs are 65 µW and 7.0 MHz, respectively.) The overall construction is similar to that of the 5517E including the overhead-mounted ballast resistor, though the HeNe laser power supply brick is in a fully shielded metal box. The portion of the tube assembly housing the beam expander is longer than the one in the 5517E and the same as that of the 5517B/C/Ds, but of no significance since it doesn't affect anything beyond looks. The design and size of the tube is also similar except that it's a bit more polished with a real Agilent label! But, from the photos, it appears as though the heater resistance adapter found in my 5517E is not used, so the tube heater resistance must be higher. However, this unit had the normal (for recent Agilent lasers) Type II Control PCB rather than the fancier one found in the 5517E. Another 5517FL had the fancy Type III Control PCB, and indeed, no heater resistance adapter.
Additional photos of the 5517E and 5517FL (and other 5517s) may be found in the Laser Equipment Gallery (Version 3.18 or higher) under "Hewlett Packard/Agilent HeNe Lasers".
The appropriate DC power supplies and laser head cable will be required for all except the 5519A/B, which plugs into the AC line. (Testing of a 5501B laser is similar except that by design, the laser beam doesn't appear for several minutes into warmup.) Aim the laser at white card or wall to view the beam.
Assuming a beam does appear, it should stay on without any flickering or sputtering. Continue watching it for the next several minutes. Power down if it does not stay lit - damage to the tube and/or HeNe laser power supply may occur if it continues to drop out and restart. Note that on high mileage lasers, there may be a significant periodic variation in beam intensity during warmup due to normal mode sweep. This should not be confused with flickering or sputtering. The smooth variation probably means the output power is relatively low but the laser may still be usable. However, a slight variation may be present even on a new laser.
A very few versions of the 5517 may take 10 minutes or longer for these two steps to occur. But it's highly unlikely you'll ever run across one of those. However, occasionally, a laser with marginal output power will take somewhat longer than the normal 4 to 5 minutes to lock as the laser power gradually increases after full warmup.
Once READY is on solid, the laser is locked and usable. But to have any confidence in its true condition, additional tests need to be performed. The most important are to measure the laser output power and REF frequency to compare with either values on the laser's backplate (if present) or specifications for the specific model. However, it is now known that the laser is most likely good for more than a doorstop. :)
If the laser beam appears and remains on but the laser doesn't lock, either the laser output power is too low (typically less than 80 to 120 µW) or there is a problem elsewhere in the laser. If the beam does not come on or does not remain on, the tube or HeNe laser power supply may be bad. See the sections below for more information.
And for much more than you probably want (or need) to know, see the companion document: Considerations in Evaluating Used or Rebuilt Hewlett Packard/Agilent Metrology Lasers.
The rather large discrepency between the diameter of the OC mirror (9.32 mm) and Mirror Spacing Rod (9.55 mm) is surprising and suggests that perhaps a sufficient physical shock could change alignment at least slightly. Also, since the mirror spacing rod floats between springs within an outer tube and there is some friction, a shock from one end may mis-position it axially as well. There is some anecdotal evidence to suggest that various types of whacking works both ways:
Perhaps that's how these tubes are aligned after being assembled - by whacking with a BIG hammer! :) Steinway && Sons has a "Pounding Room" where new pianos are broken in. Perhaps Agilent has a "Whacking Area" where new lasers are aligned. :)
Actually doing the calculations for this cavity using bwlss (see Handy Little Programs) results in the following:
Input: Enter RoC for mirror 1 (mm) (0 for planar): 0 Enter RoC for mirror 2 (mm) (0 for planar): 140 Enter distance between mirrors (mm): 127 Enter wavelength (nm): 633 Output: g1 = 1.000. g2 = 0.093. Spot diameter at beam waist = 180.96 um. Beam waist location relative to mirror 1 = 0.00 mm. Beam waist location relative to mirror 2 = -127.00 mm. Spot diameter at mirror 1 = 180.96 um. Spot diameter at mirror 2 = 593.86 um.
So, it looks like the limiting dimension would be the 1 mm diameter of the capillary at the output mirror. The mirror could be up to ~0.2 mm offset from the capillary and still not directly cut off the intra-cavity mode but diffraction losses would appear much earlier. Although the possible offset is only ~0.115 mm, it's possible that misalignment could have at least some impact (no pun....). And, as noted above, this is apparently easier to do than might be thought based on the general construction of the tube.
Unlike most other internal mirror HeNe laser tubes, there are no mirror adjustments. The mirrors are held in place against the thick glass bore (or mirror spacing rod as HP calls it) by spring pressure alone. So, the ends of the bore and mirrors must be ground to a precision sufficient for alignment to be near perfect. The distance between the mirrors is 127 mm in the 5517B corresponding to an FSR or longitudinal mode spacing of about 1.18 GHz. The purpose of having multiple glass backing disks behind the OC mirror is not really known. But along with the appropriate length spacing rod, these would provide a means of selecting cavity length by using a variable number of them without having to manufacture multiple (glass) tube designs. So, everything could be identical except for the spacing rod and number of backing disks installed behind the HR mirror and in front of the OC mirror, and perhaps the length of the associated spring(s). Having at least one backing disk may be desirable to reduce stress on the mirrors from the springs, but some tubes don't have any behind the HR mirror. So backing disks may simply be present to accomodate different spring sizes so the bore is properly positioned. Visual examination of 5501B, 5517A, 5517B, 5517D, and 5519B tube assemblies from the front after removing the beam expanders and waveplates has been inconclusive. And the glass tubes from 5517A, 5517B, and 5517D lasers appear physically identical in all dimensions that matter from the outside, which includes the cavity length. I was expecting the low REF frequency 5517A to have some obvious difference compared to the high REF frequency 5517D, but this does not appear to be the case. So, if they are all the same, what is the purpose of the extra space in front with 5 backing disks and the extra length spring? I have also checked a working 5517D, as well as a 5517E and 5517FL using a Scanning Fabry Perot Interferometer (SFPI) to measure the mode spacing during warmup (and thus cavity length based on c/2f) and confirmed what I already knew from physical examination - that the 5517D is the same as the 5517B (1.18 GHz, 127 mm). But the 5517E and 5517FL were found to be around 1.5 GHz implying a cavity length of around 100 mm. At this point, I assume the cavity length is also 127 mm in the 5517C, 5518A, and 5519A/B, but 100 mm in the 5517G. If anyone has taken more of these tubes to bits or X-rayed them, or has measured the longitudinal mode spacing either with an SFPI or a high speed photodiode and RF spectrum analyzer, and can thus shed some more coherent light on this topic :), please contact me via the Sci.Electronics.Repair FAQ Email Links Page. Also of particular interest would be the reflectivity of the OC mirror, which is also critical in determining the REF frequency. (See the section: Speculation on Axial Zeeman HeNe Laser Behavior.)
In terms of cavity length, the original "long" 5517 tubes are very similar to typical 1 mW barcode scanner HeNe laser tubes. Their relatively low power - always under 1 mW and often much less - may be accounted for by the combination of the magnetic field, precisely specified OC mirror reflectance, and possibly isotopically pure helium and neon and fill pressure, to achieve the desired Zeeman beat frequency range for each model of the 5517. This is consistent with the generally lower output power for higher REF frequency lasers - 5517Ds almost always have much lower power than 5517As when new. The heater connections with red and purple wire stubs sticking out can be seen at the left of the tube. The purple one also attaches to the cathode via a piece of springy sheet metal, no welds. The anode connection goes through the outer glass envelope but there does not seem be a glass seal into the bore, but simply a hole drilled in it to coincide with the anode location. Even after totally disassembling a tube, it's not clear what prevents the discharge from bypassing the bore unless there was some other sealant present that got lost in the process. But there are no traces of any, nor anything visible in two other tubes. The fit between the bore and surrounding glass cylinder is quite close but even this wouldn't normally guarantee that the discharge goes through the bore, especially on hard-start tubes. However, I've heard that there is some good E/M field explanation of why this works. :) The magnet is a single piece cylinder of Alnico with an inner diameter a few mm larger than the tube. It extends at the left and right ends of the tube to exactly where the discharge begins and ends, at least with the longer tubes. The shorter ones appear to use the same magnet even though the discharge ends far inside it.
I have measured the tube+ballast voltage on one 5517B laser to be about 1,675 V at 3.5 mA. This may be a bit greater than typical since the tube was high mileage, though with normal start and run behavior. The ballast was 100K so the actual voltage across the tube was 1,325 V, This is still somewhat greater than expected for a laser with a power output of less than 1 mW. Part of the reason may be that the magnet probably slightly increases the tube voltage.
Here are the long and short 5517 HeNe laser tube parameters as best I have been able to determine them so far, compared with the tube from a 5501A and a typical short barcode scanner HeNe laser tube. Except for tube voltage (which I haven't measured and am only estimating) and tube current (which is always spec'd to be 3.5 mA), the parameters for the 5517B are for one very healthy sample. The total length, planar HR mirror, divergence (without beam expander), and beam diameter are similar or identical for other models. The parameters for the barcode scanner tube are for a range of typical models. The divergence for a particular model barcode scanner tube is usually achieved by either the specific curvature of the outer surface of the OC mirror glass or with an external lens glued to it, but the cavity design including the OC RoC (radius of curvature) is the same for all. These are all random polarized tubes (and must be for the Zeeman splitting to work properly).
5517B 5517E Barcode
Parameter 5501A "Long" "Short" Scanner
------------------------------------------------------------------------------
Output Power (From tube) 0.5-1 mW 0.5-1 mW 0.5-1 mW 0.5-1 mW
Total Length (Tube only) 170 mm 194 mm 160 mm 125-155 mm
Cavity Length 130 mm 127 mm 100 mm 115-150 mm
Cavity FSR 1.153 GHz 1.180 GHz 1.5 GHz 1.3-1.0 GHz
HR Mirror RoC Planar Planar Planar Planar
OC Mirror RoC 132 mm 136 mm ???? 250-300 mm
OC Mirror Reflectance 98.74% 98.5% ???? 99.0%-99.5%
Beam diameter at output ~1 mm ~1 mm ~0.4 mm 0.4-0.6 mm
Beam divergence ~10 mR ~10 mR ~2.0 mR 1.7,2.7,8 mR
Discharge Length 102 mm 100 mm ???? 60-75 mm
Operating Current 3-4 mA 3.5 mA 3.5 mA 3-4 mA
Operating Voltage (tube) 1.4 kV 1.3 kV 1.1 kV 0.7-1.1 kV
Anode Ballast 136K 100K 100K 75K-100K
Of note here is the relatively low OC reflectance of 98.7 and 98.5 percent for the 5501A and 5517B tubes, respectively. This is one of the primary parameters that determins the REF frequency. All other factors being equal, reducing the reflectance increases cavity loss which increases the REF frequency. 5517s with higher REF frequencies (e.g., 5517D) may have even lower OC reflectivities, though magnet strength can also be used (up to a point) to adjust REF frequency.
Some other changes appear to have been made including setting the divergence closer to the diffraction-limited value rather than the much wider divergence of the previous tubes, probably a byproduct of the use of a larger OC mirror RoC and elimination of the stepped bore. The smaller divergence (or mostly the smaller beam diameter that comes with it) means that a new beam expander is required in the short tube lasers to achieve the same beam diameter.
And gone are the funky springs and backing disks, so the bore now appears to be rigidly mounted (and possibly fused) to the HR mirror mount assembly, with a cap stuck on the other end to hold the OC mirror in place.
The tube voltage is about 200 V lower but even with the same 100K ohm ballast, they should run happily on the same power supplies, though late model lasers do have yet another VMI PSU brick - the PS 253 in a shielded case.
The outer glass tube is shorter for the short tube :), but not as short as it could be. This was done either out of manufacturing or mounting considerations, or possibly to maintain an adequate gas reservoir volume.
Since Agilent no longer manufactures the 5501B, 5517A, and 5518A with their low REF/split frequencies, it's possible that all of the laser models in production now use only two types of tubes: long and short, where all of each type are identical. There would no longer be any need to select OC mirror reflectivity based on model (assuming this was actually ever the case). The required variation in the REF/split frequencies could be achieved solely by selecting a long or short tube and the appropriate magnet field strength. This would certainly simplify inventory control!
Plot of Hewlett Packard Model 5517C Stabilized Laser During Warmup shows how a typical 5517 laser behaves. Note that the entire warmup period from laser on to locked is only around 3.5 minutes because of the internal location of the heater for the active mode as noted above. A laser with the more common external heater could take 20 minutes or more to lock. The control algorithm is a bit more sophisticated than used on some other stabilized lasers, checking periodically for the status, and switching from "Warmup Mode" to "Optical Mode" about half way through the warmup period, at which point the READY LED starts flashing. A short while after it locks is when the READY LED comes on solid.
Plot of Hewlett Packard Model 5517C Stabilized Laser Near End of Warmup shows the 5 mode cycles just before locking and the final transition to the locked state. The peculiar shape of these Zeeman-split modes is clearly evident in this expanded view. Part of this is due to the locking algorithm switching between heating and cooling, but mostly it's a result of the effects of the magnetic field. More below.
The beat frequency is shown for the last 5 cycles and after locking in both these plots. This is the actual measured frequency captured along with the vertically and horizontally polarized modes and total output power. (Showing the frequency plot earlier would be a mess.) The beat only appears for a small percentage of the mode cycles with some variation during the time it is present, peaking when the F1 and F2 amplitudes are equal, and only when F1 is rising with increasing temperature. There is no beat when F1 and F2 are equal but F2 is rising with increasing temperature. The reason for this becomes evident from the simplified diagram in Axial Zeeman Split HeNe Laser Mode Behavior, or more accurately in HP-5517 Zeeman Split HeNe Laser Mode Behavior. The second diagram has been specifically crafted based on the mode plots, above, as well as SFPI Display of Lasing Mode Power Envelope of Horizontal Polarized Output of Healthy HP/Agilent 5517B Laser, of mode sweep on a storage scope as the laser warmed up. (The envelope of the vertical polarized output would be a mirror image of this one but I don't have a color digital scope to view them at the same time. The single peak visible within the envelope is really a pair of Zeeman-split modes but the resolution of the SFPI is over an order of magnitude to small to resolve them.) Thus, the plots in the second diagram more accurately represent the actual behavior of the 5517. And the clutter of the gain curves and associated junk has been removed. :) Both diagrams show snapshots of most of a mode sweep cycle starting with the cavity being 1/4 wavelength too short and ending with it being 1/8 wavlength too long. (The case of 1/4 wavelength too long would be the same as the first, 1/4 wavelength too short). Only when the longitudinal mode is near the center of the Zeeman-split neon gain curves will there be a beat. In addition, the mode amplitudes are changing rapidly as the cavity expands at those high slope locations on the gain curves. When the cavity length changes (longer or shorter) by 1/4 of the lasing wavelength of approximately 633 nm, the amplitudes are again equal, but the two separate longitudinal modes are oscillating far apart and there is no beat. Note that the red and blue plots include the F1 and F2 amplitudes, but also may have contributions from another longitudinal mode derived from the same split gain curve which will thus have the same original circular polarization. But when centered and locked, only the desired Zeeman-split modes are oscillating.
Note that as the tube ages with use, the gain declines and the width of each gain curve that is above the lasing threshold decreases. Eventually, with a really high mileage tube, there may be no overlap at all and the beam will probably disappear for a part of the mode sweep cycle. But it is exactly at that point where the Zeeman beat would be generated, so it will also disappear entirely. Lasers are generally taken out of service long before this happens, but I recently found one whose output power was so low that this behavior was present - or absent depending on your point of view!
The second diagram above, HP-5517 Zeeman Split HeNe Laser Mode Behavior would likely apply most accurately to a nearly new 5517 since that's what it was more or less based on. As expected, when the split mode is centered, there are no other modes oscillating. But if slightly off-center, there is a strong mode at a distance of 1 longitudinal mode spacing from it. Normal and Zeeman-Split HeNe Laser Mode Power Curves. Compares a "normal" (common cherry flavored HeNe laser), and two 5517s. One has seen a fair amount of use while the other is close to new. Note that the similarity in the general shape of the "hat" - the top portion of the lasing output power curves, but the new laser has the added "skirt" below, which has a similar amplitude. The skirt is present in the region where there are two longitudinal modes lasing with one of them being a Zeeman-split mode. Thus, when and if a skirt will be present, and its height relative to the hat region, will be affected by the cavity mode spacing and magnetic field strength. As far as the mode sweep is concerned, the skirt mainly adds an offset to the total output power. Two other 5517Bs in various stages of life show similar skirts. A relatively low mileage unit (but not quite as new as the one in the diagram) looked much the same but with a slightly higher ratio of hat:skirt height. :) And one that had been really high mileage whose magnetic field was reduced to bring down REF had a 3:1 ratio of hat:skirt. With the reduced field, the central region is wider, but the hat is otherwise similar. So there are 3 lasing regions in these Zeeman-split mode plots as shown in Mode Competition in HP/Agilent 5517 Zeeman-Split HeNe Laser:
Here is how the article "An Instant-On Laser for Length Measurement" by Glenn M .Burgwald and William P. Krugein describes the operation of the laser tube in the Hewlett-Parckard Journal, Aug., 1970.
"If an axial magnetic field is applied to a laser which is free from polarization anisotropy in either the mirrors or the plasma tube, the output splits into two frequencies of left and right circular polarization. First-order theory predicts that the frequency splitting is proportional to magnetic field strength and to the ratio of line Q to cavity Q. In the new laser, magnetic field strength is adjusted for a difference frequency of about 2.0 MHz. Line center is virtually midway between the displaced lines, so proper cavity tuning can be assured by adjusting for equal intensities of the lines."
This was written with respect to the earliest HP metrology lasers but the principles are the same for the 5517s (as well as the 5501B). And they show a gain curve diagram even simpler than the one above. See that article for more details. The first order theory is consistent with my measurements and speculation where I use "cavity loss" instead of "the ratio of line Q to cavity Q" but they are equivalent.
The peculiar shape of the real mode plots almost certainly due to mode competition between the pair of Zeeman modes, and at times between the Zeeman modes and a normal mode that may also be present, and between two normal modes if only they are present. But so far I have found no references anywhere. The split gain profiles need to be asymmetric to account for it, and this has been confirmed by testing several 5517 lasers on a Scanning Fabry-Perot Interferometer (SFPI). The simpified explanation of Zeeman splitting rarely takes into account what happens in the real World which distorts the gain profiles in these lasers as a result of mode competition for the same pool of excited atoms. This happens in short normal HeNe lasers as well, but it isn't as dramatic. (More on this below.) So, drawing a pair of nice bell-shaped gain curves really isn't accurate. The net effect is depicted in HP-5517 Zeeman Split HeNe Laser Mode Behavior. Here, the lasing mode power curves have been modified so that the results would be roughly similar to what was in the plots, above. And HP-5517 Zeeman Split HeNe Laser Mode Behavior Versus Mode Position on Gain Curve shows one complete mode cycle along with little split lasing mode power curves.
Also note the second longitudinal mode (in addition to a Zeeman-split mode) present for a part of the mode sweep cycle. Extra "rogue" modes should never be present when HP/Agilent lasers are locked, though one may appear at times as in the diagram when warming up and the Zeeman-split modes are not centered on the split neon gain curves. If any are present when locked and they align with the X and Y axes, then the only effect will be to slightly decrease the MEAS or detected REF signal level with respect to laser output power since any difference frequency is way outside the passband of any electronics. However, if they are not aligned with the X and Y axes (e.g., at 30 degrees), they will cause level changes in the envelope of the signal from the optical receiver's photodiode due to self-interference in the interferometer. This is similar to what would happen if the primary Zeeman modes were misaligned, or not pure. The consequences could be transient position errors but only during motion. The end-points would be accurate since the optical receivers only respond to AC. There's a fine balance between the desire for a high split frequency (which extends the split gain curves) and the desire to suppress these "rogue" modes. So, for example, increasing the magnetic field to boost split frequency may produce rogue modes if the cavity length isn't also decreased.
Normal and Zeeman-Split HeNe Laser Mode Power Curves compares behavior quite close to what's actually observed. I can't guarantee that these are to scale, but they do show the general shape with and without a magnetic field. The lasing mode placement in the diagrams is such that there are two equal amplitude normal modes without the magnetic field and two equal amplitude Zeeman modes with the magnetic field, with a cavity length selected to just suppress rogue modes in the latter case.
I was curious (actually quite curious) to see what the mode behavior of a typical HP laser would be without a magnetic field. One must be quite curious - in fact quite quite quite curious - to do this as it requires removing the glass tube from the magnet intact - which is literally an all day (well all morning) affair using knives, dental tools, and other instruments of torture to dig out the rubbery potting compound securing the tube inside the magnet/optics assembly. With the strong axial magnetic field doing all sorts of wonderful things and based on the effects of Zeeman splitting, the plots of mode sweep with and without the magnet should be dramatically different. Although I had already removed several tubes from their magnets, they were all end-of-life with rather low power so there would always be questions as to whether whatever was found would also apply to a healthy tube. I had a 5517C that would lock with decent output power (265 µW) but required over 4 mA for the discharge to remain stable during warmup. Such a tube should behave reasonably normally, laser-wise, but since the future life of high dropout current tubes is unpredictable, it would not likely be in demand and could be sacrificed in the name of science (and curiosity). (In principle, the tube could be remounted and used but I doubt that will ever happen.)
First, a plot of the mode sweep of this laser was made as a reference. Its appearance was similar to that of those shown above. Then the major surgery was performed to remove the glass tube. The initial results were quite strange. It appeared as though there were always 2 modes that were nearly identical except for a burst of randomness where the mode sweep would normally do its mode hop thing. And these were present at both polarizations! It was as though the output was totally non-polarized - rotating a polarizer had almost no effect! The removal process was rather violent at times (but I won't go into all the gory details!), so I put the tube back in its magnet to confirm that it had not been damaged. It hadn't. After ruminating on this totally peculiar mode sweep during my afternoon walk, I began to suspect something in the environment like a stray magnetic field resulting in a transverse Zeeman effect producing a split mode with a very small difference frequency. That sort of strange mode behavior is a characteristic of the mode sweep of a transverse Zeeman laser at some range of relatively low magnetic field strength. (See the section: Transverse Zeeman Stabilized HeNe Lasers.) And there was another HP laser sitting less than 1 foot away! Sure enough, removing that laser produced a mode sweep more along the lines of what would be expected with a HeNe laser tube having a cavity length of 127 mm (longitudinal mode spacing of 1.2 GHz). See: HP-5517C HeNe Laser Tube Mode Sweep Behavior. The top plots are of the normal 5517C laser with the lock point being where the red and blue (F1/F2) polarized modes (lined up with the horizontal and vertical axes) cross, located at minimum output power. The angular shape is due to the distortion of the split neon gain curves resulting from the magnetic field, tube geometry, and other factors. It corresponds fairly accurately to the diagrams shown above. The bottom plots are of the same tube without the magnet and waveplates so that the polarized modes are the (non-Zeeman-split) longitudinal modes. The time axes of the two sets of plots are similar and the plots are approximately aligned one above the other more or less where the mode is centered on the Zeeman split neon gain curve (top) and the normal neon gain curve (bottom). For this tube, the normal polarized modes also line up with the horizontal and vertical axes - probably not entirely a coincidence. (This is not required since the waveplates can correct for any mode angle but it would simplify the alignment process.) However, the appearance is still not quite typical, as it's somewhat polarized with bumps. The red mode in the bottom plot doesn't quite go to 0 as it would in a linearly polarized laser. And where the bumps are, the mode orientation should reverse, but it does not. And although the appearance would suggest a neon gain curve with a relatively flat top (just a small depression in the middle) and steep sides and a lasing width of only about 1.3 GHz, not the 1.5 or 1.6 GHz normally used, this could also be at least in part a result of mode competition. However, with the lower mirror reflectance and thus increased lasing threshold of the higher REF frequency 5517 lasers, there would be a narrower effective gain bandwidth. But some stray magnetic field must still be influencing its behavior to cause the polarized mode behavior. In fact, this tube is extremely sensitive to magnetic fields - much more so than the average run-of-the-mill HeNe laser. But I'm not sure that even if stray magnetic fields were totally eliminated, the tube would revert to totally normal behavior. I removed all sources of stray magnetic fields I could identify including a loudspeaker a couple feet away, and degaussed the tube and housing (though there don't appear to be any ferrous materials there), but there was no change. Perhaps a set of Helmholtz coils to eliminate the Earth's magnetic field would be able to create a sufficiently field-free region of space. This specific tube is not unique though - a 5517B tube produced a generally similar set of plots and has the same sensitivity to magnetic fields and a 5501A also behaved in a non-typical way with no external magnetic field. What this probably indicates is a polarization anisotropy in the laser tube very close to zero, required for Zeeman splitting to predictable and consistent or be present at all. Or the opposite. :) There is one thing that is very asymmetric: the anode-end discharge enters the bore from one side and the cathode-end discharge exits the bore from the opposite side. In most modern HeNe laser tube, these are generally fairly symmetric in both cases. And on the SFPI, there appears to be something very strange going on when looking at the output through a polarizer. It may be a large frequency modulation of the optical frequency as the appearance is of a full amplitude high frequency oscillation in the mode display, but only with a polarizer. The might be due to the HeNe laser power supply ripple, or low level plasma oscillations, or aliens attempting to communicate with Earth. :-) Without a polarizer, the appearance is normal.
Also, note the depression in the blue mode (and total power). Although that may indicate the presence of a Lamb dip - and many aspects of the physical design of the HP/Agilent tubes are consistent with the requirements of a Lamb dip laser - it could simply be an artifact of the way the neon gain curve lines up with the longitudinal mode spacing.
Since all HP/Agilent thermally-tuned lasers employ a tube with similar construction, I would expect their mode sweep behavior to be similar as well. And I doubt this behavior has anything to do with usage - it is simply a characteristic resulting from the design. But it would be nice to be able test a new 5517 tube sans magnet. However, the probability of this happening is somewhat below that of pigs flying.
In fact, plotting the horizontal and vertical components of the polarized modes of a healthy Agilent 5517C tube with no magnetic field as it is rotated through approximately 120 degrees in 15 to 20 degree increments shows some even stranger mode behavior. See Mode Sweep of Agilent 5517C HeNe Lasaer Tube with No Magnetic Field. The plots of the other 5517C laser, above, are very well behaved in comparison. This tube came from a laser removed from service due to random dropouts, likely a result of a bad cathode connection. But it still had output power well above 300 uW and REF below 3.0 MHz. It's not even really possible to define a set of polarization axes as would be the case with a "normal" HeNe laser tube. The best that can be concluded is that the mode variation is largest in the first plot and smallest in the 5th plot. For a normal tube, there would be an orientation with zero amplitude. However, as expected, the total power plot is well behaved.
To reiterate, I believe this sort of crazyness to be normal for HP/Agilent laser tubes including all 5517s, the 5501B, and probably the 5500A/B/C and 5501A as well. Without a magnetic field, they are all highly random polarized (and I mean RANDOM). Healthier tubes may actually actually be more random. In these plots, there is only a weak tendency toward a set of polarization axes and any disturbance will screw up the polarization. In fact, the slightly vibration will result in wild variations in relative mode power during part of mode sweep. Similar vibration will have no effect on "normal" tubes. I do not know if this behavior is due to the tube structure or the mirrors, or some magic sauce. ;-)
However, applying a modest axial magnetic field wipes away fingerprints. :) Once the field strength reaches a value of around 100 G, the randomness disappears. This threshold tends to be slightly lower for HP/Agilent tubes compared to common internal tubes like those from Melles Griot or JDS Uniphase. It is likely related to the randomness with no field which may promote Zeeman splitting of the lasing mode. But at the normal operating field strength of 200 G or above, it's essentially impossible to tell the tubes apart from mode behavior.
The older 5501A also behaves much the same as shown in Modes and Beat Frequency of HP-5501A HeNe Laser Tube 1 With Normal Axial Magnetic Field.
The Agilent 5517E/F/G lasers use a tube with a slightly shorter cavity, but behavior is generally similar with the magnetic field present. During mode sweep, there is a modest variation in total power and this is a somewhat larger percentage of the total power than with the longer 5517A/B/C/D tubes even when new. However, what may not be obvious is that with no magnetic field, the output power from these short tubes may decline dramatically - or even disappear entirely - during a part of mode sweep. Their longitudinal mode spacing is around 1.5 GHz, and two modes just barely fit under the neon gain curve. At that point, the modes are on the tails of the gain curve, near or below lasing threshold. With one sample, the power variation with no magnetic field was from 0 to 450 micro;W. But with the magnet, it was only 300 to 360 µW. The normal lock point would be near the minimum of 300 µW, but that's a lot more than 0 µW! :)
Performing the same tests with one of these short tubes produces even stranger results as shown in HP/Aglient 5517E HeNe Laser Tube Mode Sweep Behavior. The conditions are essentially the same as for the 5517C plots, above. Again, the behavior is smooth and predictable with the magnetic field, but the polarization is totally chaotic at times without it. I don't really know what's happening most of the time. Changes may be taking place on a time scale faster than the sampling rate of the data acquisition system, which is only 60 samples/second. But note that the total power (green) curve is still smooth as expected. It's only when observed through a polarizer the it turns to randomness. Even when the polarization isn't changing randomly, it isn't what one would see with most "normal" HeNe lasers. And here is a closeup of the randomness: HP/Agilent 5517E HeNe Laser Tube Mode Behavior with No Magnet - Expanded. Rather wild, no? :) Someone asked how this tube got through Agilent Quality Control. The answer is that I believe it is this way by design, or at least this behavior when not inside a magnetic field is a byproduct of the design which minimizes asymmetries in the mirrors, bore, and other aspects of the laser tube construction.
Out of further curiosity, I did the same experiment with my custom SG-5517 laser which uses a Spectra-Physics 007 HeNe laser tube. These plots are shown in SP-007 HeNe Laser Tube Zeeman Split Mode Sweep Behavior. Again, the lock point is where the blue and red modes cross close to minimum total output power. While the general character of the plots is similar to those for the genuine HP-5517C, the details differ dramatically. And the SP-007 has none of the hyper-sensitivity to stray magnetic fields that is present with the 5517 tubes so the normal longitudinal modes (no magnet) look like those of a short well behaved random polarized tube. (However, in the interests of full disclosure, these plots were not of the same physical tube, only the same model tube since I didn't want to disassemble the SG-5517.)
Out of further further curiosity, I tried the same experiment with the totally screwed up (as far as mode behavior is concerned) Far East tube. (See the section: A Far East HeNe Laser Tube.) It's about 6 inches long but unlike most typical short tubes, it has a very long radius hemispherical cavity, with a curved mirror of around 1 meter RoC - 3 to 4 times what is common, so the mode volume should only taper a small amount within the active discharge. The gain curves displayed on the SFPI did not show the dramatic asymmetry present with the HP-5517C or even the SP-007. While far from conclusive given the overall pecularity of this tube as well as other basic differences compared to the HP-5517C and SP-007, it is, well, interesting. No plots, sorry. :-)
Excel metrology lasers, which are performance clones of those from HP, use tubes of conventional design. Their mode behavior is definitely not the normal shape and is somewhat asymmetric, but also not nearly as skewed as that of the HP lasers. This might make some sense if they have the 25 to 30 cm OC found in typical short HeNe laser tubes and thus a somewhat long radius hemispherical cavity. Then again, that may be totally bogus. :)
So what about the asymmetric shape of the Zeeman-split gain curves? At his point, it is almost certainly a result of mode competition between the two Zeeman-split modes, and with and between any normal modes that may also be present. The magnetic field splits the gain curve and shifts the two copies apart but doesn't do much more.
However, when I first became obcessed with the strange shape, many mechanisms were considered and ruled out. For example, that the magnetic field takes the original symmetric gain curves and smears them out non-uniformly, or affects the Doppler broadening non-uniformly. But doing so requires that the population for the right-circularly polarized mode be the mirror image of the population for the left-circularly polarized mode, not simply that the populations are smeared and shifted. This severely limits the explanation. Or, even more off the top, that Zeeman splitting produces more than one pair of shifted neon gain curves (i.e., "hyperfine structure") and the weighted sum then represents the net gain curves. Here are some other possibilities that have been pretty much eliminated:
Reversing the magnet end-for-end does swap the shape of the F1 and F2 polarizations. When this test was first performed, I thought that it might reveal some deep hidden revelation. But the change is caused by the handed-ness of the circular polarization swapping due to the direction of the field, not any change in field strength. As further confirmation, it doesn't make any difference in the split frequency, which would seem to be unavoidable if the shape of the gain curves were changing due to the different amounts of gain along the length of the active discharge. However, that in itself is interesting as it means (not unexpectedly) that the direction of the magnet would affect how the waveplates need to be set up to achieve proper F1/F2 orientation and locking. So that arrow in Magic Marker on all HP/Agilent laser magnets pointing to the front of the laser is there for more than simply making sure that lasers sitting side-by-side are anti-social and always repel each-other. :)
The closest to an explanation for the peculiar shape has come from Harold Metcalf, Distinguished Teaching Professor at Stony Brook University, Stony Brook, NY, and that was simple mode competition. In retrospect, mode competition is a natural fit for the observed shapes of the lasing mode power curves. This agrees with the behavior in the very straight slopes and nearly flat region of output power within the region where there is a split frequency present. Basically, although the gain curves are visualized as being separate, there is still only one population of excited atoms, so the two components of the split lasing mode are competing for a limited resource. My interpretation is that at either end of the split frequency region, only a single mode is oscillating while in the exact center, there are two modes with equal amplitudes. Being resource-limited, the sum of the two modes in the split mode must be similar to the single mode at either extreme, and the simplest equation that will then join them is a straight line, which also explains the nearly constant output power in this region. Something similar is happening in the region where there are two separate longitudinal modes oscillating since again, even though the neon gain curves are split, it's still a single population of excited atoms. At the very least, it's easy to argue that the total power as a result of the sum of the two separate longitudinal modes isn't going to be double or more of the power in the split mode when it alone is present (though it is slightly greater). However, it's even harder to visualize the behavior in those regions. It may be time for Matlab.....
There's probably a research paper from the 1960s or 1970s that will make everything perfectly clear. But all those I've found so far have been less than entirely useful. Translation: I couldn't make heads or tails out of the hairy math and there were no pictures or cartoons. ;-)
The slide show may be started by going to HP/Agilent 5517 Zeeman-Split HeNe Laser Mode Sweep Animation and runs in a separate window. ESC to exit. This is known to be compatible with PowerPoint 2007. Once it's stable, if ever :), a version compatible with PowerPoint 1997-2003 may be available. (while the present one is supposed to work with these, the animations fail to load in PP2003. I don't know if there's aproblem with what PP 2007 created, or my version of PP 2003, which has probably never been updated with bug fixes.)
In addition to a very few explanatory slides, there are animations for a normal (non-Zeeman-split) HeNe laser tube such as that from a barcode scanner, a like-new 5517B laser tube, and a somewhat high mileage 5517C laser tube with no magnet. These all have the same cavity length. Unfortunately, testing the same 5517B tube with no magnet isn't likely to happen. :-)
(If you're wondering what happened to the neon gain curves, lasing threshold, and other junk that used to be in the plots and shows, I decided (1) they were hard to draw, (2) didn't add anything useful, and (3) I couldn't figure out what to do with the gain curves beyond the lasing region anyhow. So they are history!)
While viewing the animation of the normal HeNe laser, observe the following:
The total power is the sum of the individual lasing modes with an approximate estimate shown on each frame of the animation. The cavity length is 127 mm (~5 inches), which is similar to that of the 5517A/B/C/D and 5501B. There is only a single lasing mode for more than 50 percent of mode sweep. (The cavity of the 5517E/F/G is about 20 percent shorter, but the plots for these lasers should be very similar.)
While viewing the animation of the 5517 mode sweep, observe the following:
While viewing the animation of the naked 5517 laser tube (no magnet), observe the following:
All the frames of the normal 5517 mode sweep are also available in a separate PP show as HP/Agilent 5517 Zeeman-Split HeNe Laser Mode Sweep Sequence. Note how the output power - which is the sum of all the red and blue modes present as denoted as Total Power changes by less than 2 percent within the region where there is the single split longitudinal mode (frame IDs 925 to 075). (If drawn to scale, the separation between the two lasing lines would not even be visible.) Where there are two normal modes (red and blue separated by the FSR of the tube cavity), the output power is somewhat greater, possibly because the lasing modes are far away from each-other most of the time and thus have different sets of excited atoms to stimulate. End-of-life lasers will have much lower power within the Zeeman-split region but a more dramatic increase in power away from it since the gain relative to the lasing threshold is smaller.
All the frames of the mode sweep of the normal tube (no magnet) are also available as HeNe Laser Mode Sweep: 127 mm (~5 inch) Cavity Length Showing Effect of Mode Competition.
Enjoy! ;-)
These lasers consist of the laser tube assembly, potted (brick) HeNe laser power supply, beam sampler, connector PCB, and control PCB. mounted on a an metal chassis. Any of the parts can be replaced in under 5 minutes using common tools, with only minimal or no adjustment or alignment.
Laser tube assembly:
All of these consist of the actual glass HeNe laser tube potted with some sort of rubbery material inside its Zeeman magnet, beam expander, and adjustable waveplates. The heater/cathode is attached via a 2 pin plug while the anode has its own single pin high voltage connector. The HeNe laser tube ballast resistance of about 100K ohms is conformal molded into the silicone insulated HV cable. The bifilar-wound heater inside the laser tube has a typical resistance (cold) of 8 ohms on most tubes. (The one exception I know of is an early 5517E which may have even been a prototype, where it was around 4 ohms.) When at operating temperature, the resistance is spec'd to be higher by a constant factor since the actual temperature can be determined based on the known thermal coefficient of resistance of the heater wire.
Only 4 screws hold the tube assembly to the chassis for lasers in the small cases (all the 5517s except the 5517A, as well as the 5501B). One or two will be flat head screws which provide either a fixed axis for horizontal (pan) alignment, or self alignment (no adjustment permitted). All of these tube assemblies appear physically identical, except for the 5517E (and probably 5517F/G) which are slightly shorter. (They, of course, differ with respect to the REF/split frequency.) The larger tube assemblies found in the 5517A, 5518A, and 5519A/B mount with 3 bolts and have machined alignment pins so no adjustments are needed or possible. They, too, are physically identical except for one small area of the casting that needs to be cut away if installing a 5517A tube into a 5519A/B laser to clear the internal +/-15 VDC power supply.
I first made some measurements of the fringe field strength by adapting a clamp-on DC ammeter for this purpose. (It was handy and I didn't have a gauss meter.) Since it uses a hall-effect device somehow, I figured it would respond to magnetic fields and sure enough, the sensitivity on the 200 amp range is about perfect for these magnets. The only modification made to the meter was to put a non-magnetic 0.015" shim in the clamp to reduce variability due to jaw contact movement. The clamp was positioned flat against the side of the magnet centered front to back and oriented to maximize the reading. To verify that the meter wasn't drifting or being magnetized, it was checked periodically against one of the 5517B magnets which served as a reference.
Later after attempting to reconcile conflicting theory and observations, I built a simple gauss meter. See the section: Simple Gauss Meter for Measuring Zeeman Magnet Strength
Here are the data for a variety of HP/Agilent laser magnets. Most are complete tube/magnet/optics assemblies, but since there is little ferrous metal in the tube, whether the bare magnet is tested or the entire assembly shouldn't make much difference. The fringe field is significantly lower than the interior field. While there is some variability, this ratio can be used as a guide in predicting to be a relatively constant ratio for magnets from the same model laser, and even among all the 5501B and 5517A/B/C/D/E lasers since the magnets appear to be made of the same material and have the same dimensions. Thus these values can be used for comparison. However, the 5501A magnet differs enough that the measured values may not correlate with those from the other lasers though the one I checked appears at least as accurate:
REF/Split
Laser Tube Frequency <- External Field -> <-Interior Field ->
Model Part Number Range Relative Absolute Predicted Measured
-------------------------------------------------------------------------------
N1211 24703-60207 ???-??? MHz 33.3 270.0 G 366.7 G
5501A 05501-60006 1.5-2.0 MHz 29.5 239.0 G 322.7 G 345 G
" " " " " " 32.7 265.0 G 357.7 G
" " " " " " 34.6 280.4 G 378.5 G
" " " " " " 35.2 285.2 G 385.0 G
" " " " " " 37.5 303.9 G 410.2 G
Mean 33.900 274.7 G 370.8 G
5501B 05501-60102 1.5-2.0 MHz 24.9 201.8 G 273.4 G
" " " " " " 29.2 236.6 G 319.4 G
" " " " " " 24.0 194.5 G 263.5 G 256 G
* " " " " " " 42.3 342.7 G 285.5 G
" 05501-69202 " " " 26.1 211.4 G 285.5 G
Mean 26.050 211.0 G 285.0 G
5517A 05517-60301 1.5-2.0 MHz 25.9 209.9 G 283.3 G 259 G
5517B 05517-60201 1.9-2.4 MHz 33.6 272.3 G 367.5 G
" " " " " " 33.7 273.1 G 368.6 G
" 05517-68201 " " " 29.5 239.0 G 311.7 G
R " " " " " " 33.4 270.6 G 365.4 G 362 G
" " " " " " 36.2 293.3 G 399.0 G
Mean 33.280 269.7 G 364.0 G
5517C 05517-68217 2.4-3.0 MHz 34.3 277.9 G 375.2 G
" " " " " " 41.1 333.0 G 449.6 G
" 05517-68218 " " " 38.5 312.0 G 421.1 G 350 G
" 05517-68249 " " " 35.0 283.6 G 382.9 G
Mean 37.225 301.6 G 407.2 G
5517D 05517-68224 3.4-4.0 MHz 37.2 301.4 G 416.9 G 380 G
" " " " " " 38.1 308.7 G 416.8 G
" " " " " " 38.8 314.4 G 424.4 G
" " " " " " 39.3 318.4 G 427.7 G
Mean 38.300 310.3 G 419.0 G
5517E 05517-6???? 5.5-6.5 MHz 44.8 363.0 G 490.0 G
5517FL 05517-68253 >7.0 MHz 47.1 381.6 G 515.2 G
5517G 05517-6???? >7.2 MHz ???? ???.? G ???.? G
The entries for "Interior - Predicted" use the value 1.35 (from the reference 5517B) to generate the value for the interior of other magnets which still have their tube in place from the exterior measurement. The entries for "Interior - Measured" apply to magnets for which the tubes had been removed so that the interior was accessible. Note how some can be quite far off so take this with a grain of Alnico magnet material. :)
The 5501B marked with a "*" was not included in its average because the value was such an outlier. My suspicion is that this laser was the only one I ever found with a factory (or service center) installed magnetic shunt to reduce the REF frequency and it was on the opposite side of the magnet from where I measured the field to be over 50 percent higher than any of the others.
The magnet from the 5517B marked with an "R" is what I use as a reference field to assure consistency if measuring other lasers later.
There is only a single data-point for the 5517A because in order to make a measurment that would be consistent with the others, the bare magnet had to be removed from it's cast metal casing, and only one naked 5517A magnet was available.
In general, while the magnet strength on average did increase along with the laser model's REF frequency range, there were quite spectacular exceptions, like the 5501B with the highest reading of any of the common HP/Agilent lasers, only exceeded by the hyper-REF frequency Agilent-only 5517E/FL! (It is believed that this magnet may have had a field reudcing shunt installed at the factory to bring it down to earth, but that is long gone.) And 5501A/Bs generally have stronger magnets than 5517s with higher REF frequencies. What any of this may mean, assuming that my measurements have any basis in reality, is that there must be variations in the tubes like the OC mirror reflectivity and that it's likely that tubes and magnets are hand-selected to achieve the desired REF frequency before being mated permanently. And/or, the field strength may even be fine tuned after assembly to optimize REF.
But there's another issue that tends to go unnoticed: A long cylindrical permanent magnet such as these has a field that is fairly uniform near the center inside but tapers to 0 at the ends and then reverses outside. (This is unlike an electromagnetic solenoid has a uniform field inside extending beyond the ends and tapering off.) But Older HP/Agilent lasers are designed so that the magnet is exactly the length of the bore discharge and thus include the tapering region. This short-changes the Zeeman effect since a portion of the bore discharge sees a lower field strength. The newer high-REF Agilent lasers using a shorter tube seem to have dealt with this at least partially since the bore discharge is slightly shorter and the magnet length is the same. Whether this was deliberate or simply a fortuitous accident (with Agilent not wanting to redesign the magnet) is not known. At some point, I need to do an experiment with an extended magnet to see how much higher that alone would push the difference frequency.
HeNe laser power supply:
Very old (perhaps roughly pre-1990) lasers used Laser Drive model 111-Adj-1 HeHe laser power supplies which had adjustable current via a pot on the laser connector PCB. At least most of them did. I did find one that had the same part number but no third wire. All later versions use VMI power supplies with a fixed current of 3.5 mA. However, the pot is still present on the connector PCB even on lasers made after 2006 (and probably to the present day), but does nothing.
There are four versions of the VMI power supplies used in these lasers. The two oldest ones (VMI PS 148 and VMI PS 217) have the same HP part number of 0950-0470. Then around the year 2000, the power supply changed to the VMI PS 373 with Agilent part number of 0950-4459, which is found in most recent lasers. I know that switching from the 148 to the 217 reduced the residual current ripple from over 3 percent to less than 1 percent because I measured it. I do not know what changes were made in the 373 but it also has very low ripple, nor what other differences there may be between these models. VMI claims all of this is proprietary information. Can you believe that? :) However, I have de-potted a dead model 373 and have reverse engineered its schematic. See the section: VMI 373 HeNe Laser Power Supply. The final version I've seen is the VMI PS 253 (Agilent part number 0950-4073), only found so far in the 5517FL and other late model lasers with the shorter tube. It is potted inside a metal case so the appearance is unmistakable. Preliminary dissection of a PS 253 indicates that it is similar to a PS 373 but with a modified PCB layout, an additional filter capacitor and inductor, and some minor changes to the ripple reducer and high voltage resistors. The PCB is actually slightly larger than that of a PS 373 with additional cut-away sections, presumably to prevent arc-over. It is screwed to the metal case in three places providing direct grounding, and thus eliminating the need for the aluminum shield plate often found bolted to the older bricks. I'm not planning on fully reverse engineering the PS 253 but will probably do some spot checking to identify any major differences.
Until recently, the only defective power supplies I've ever found in HP lasers were nearly all Laser Drive 111-Adj-1s. And one type of failure may result in the adjustment pot having no effect with the power supply pumping way excessive current (like 6 or 8 mA) through the tube. With luck, the ballast resistor catches fire and explodes before the tube is damaged. :( :) I've also see a few VMI PS 148s that had excessive ripple, so something in its output filter had blown. But if I hadn't been checking ripple on a bunch of these power supplies, it probably would have gone unnoticed, since as long as the tube stays lit, performance of a healthy laser probably wouldn't be affected in any significant way, though the MEAS signal might have a bit more fuzz on a scope display due to the current ripple producing amplitude ripple in the laser output power. However, the additional ripple would make the effective dropout current go up, so a marginal tube might start sputtering on that supply. But now having tested several dozen late model lasers (2004 to 2006), several defective VMI PS 373s have turned up. They either draw no current, have lost regulation, or draw excessive current. Some also seem to be very sensitive with respect to input - they will start and run fine in a laser but may refuse to start reliably or at all if powered from a bench supply. Go figure. And on a few, while the output current remains well regulated, it has dropped to around 3.41 or 3.42 mA - or was set that way at the factory due to bad quality control! Reduced current would normally not matter very much, but could result in a premature failure should the tube's dropout current increase, as it often does after long hours of runtime. Now Agilent wouldn't want their lasers to fail early, would they? :-)
Beam sampler:
These consist of a first non-polarizing angled plate to sample a portion of the output beam and a second non-polarizing angled plate to take this and split it between the LCD switch, and reference photodiode. The LCD switch attaches to the Control PCB via a 4 pin connector - 2 pins for the LCD drive and 2 pins for the photodiode behind it. The reference photodiode is actually mounted on the Control PCB and simply pokes its head into the beam sampler assembly. Beam samplers for all model lasers appear to be identical and interchangeable.
Connector PCB:
Aside from the Mil-style connector to the outside world and the 24 pin connector to the Control PCB, this has some filter capacitors; fuses for +15 VDC and -15 VDC; and the Power, Laser On (really same as Power), and READY LEDs. The one pot does nothing except for really old lasers with the Laser Drive 111-Adj-1 HeNe laser power supply brick.
Very old lasers had a case interlock switch to disable the laser tube from being powered if the covers were removed, and a "Service" switch to override this. :) Both of these switches have been eliminated, though the PCB pads and wiring for them are still present, but bypassed. It's worth removing both switches on lasers that have them and adding the required jumper diagonally between the center pads that are closest together on both switches. (Do NOT just add the jumper - the switches must be removed!)
Control PCB:
After removing the cover on one of these lasers, the most obvious assembly aside from the tube is the PCB which controls the heater inside the laser tube based on inputs from a pair of photodiodes (mode and REF). Over the years, there have been three versions:
Since the introduction of the 5517A laser through the early 2000s, all the these lasers used the same Type I Control PCB based on SSI TTL logic for timing and an analog feedback loop. The one in the 5517A is physically larger and not interchangeable with those in the small lasers, but is nearly identical electrically. The main functional difference is the addition of a second line driver and a connector so it can provide the MEAS signal used in 5518A and 5519A/B lasers. There had been virtually no change in the design over 15 years or more, except that a modification to the internal REF receiver makes newer lasers require somewhat higher optical power to lock than older ones. This is probably to avoid false locking when there may be substantial ripple from the HeNe laser power supply or due to plasma oscillations in the laser beam. There are no useful indicators on the Type I Control PCB, only one wimpy LED that duplicates the function of READY on the backpanel. There are no LEDs on the control PCB in the 5517A.
But since sometime after Agilent was created, at least two new versions of the Control PCB have appeared. FPGAs/FPLDs and/or a reasonably high performance microprocessor or DSP have replaced discrete TTL logic, though there are still some analog parts. If one wants to count transistors, I bet the newest one - the Type III Control PCB - has over 1,000 times the number of transistors as the Type I Control PCB! These are also almost entirely surface mount (SMT), with major parts on both sides of the PCB in the latest version.
The first of these, the Type II Control PCB, is electrically and physically interchangeable with the older Type I Control PCB and first appeared in lasers manufactured around 2004. It is based on a Xilinx XC2S50 Spartan-II FPGA which replaces the discrete TTL state machine and most of the logic, and everything else is in more modern (and available) SMT parts. While very reliable, a failure for any reason other than an obvious problem like a blown fuse or bad DC regulator with no underlying cause would likely render it non-repairable except by Agilent or an authorized service center since it's then just a black box with no real way to easily troubleshoot. A service manual may exist but I've never seen one. And even if it did, sophisticated test equipment and a surface mount rework station would be required to have any chance at repair. However, this version has all the same jumpers and temperature set pot, so normal testing and adjustment is similar to that of the Type I Control PCB. The easiest solution would then be to simply swap in a known good board (either version), of which there should be plenty available in lasers with bad tubes.
It's not clear what, if anything, the Xilinx-based controller adds to the laser, other than to make it more proprietary and difficult to service. After all, features are not being constantly changed or added, nor will there be security issues due to computer viruses - it doesn't run Windows! :) So, periodic firmware upgrades and bug fixes really aren't required, which is a primary benefit of using Field Programmable Gate Array (FPGA) technology like that of the Xilinx parts. The Xilinx XC18V01 PROM can be reloaded but I kind of doubt this would ever be necessary! However, there is no doubt that the discrete through-hole TTL and analog parts dating from the 1970s are becoming more difficult to find and expensive. So the Type II Control PCB is probably now used for all Agilent 5517B/C/D lasers, not only for those with "military calibration" or OEM or special requests. Swapping in an Analog Control PCB results in no obvious differences in performance, and this would be my suggested method of repair unless there were special requirements. And now (2009) 5517B lasers with these Control PCBs have been turning up from 2004, 2005, and 2006, probably removed from service in wafer fabs after degrading to the edge of Agilent specs for REF frequency, or a specific number or hours of service. An identical laser from the same source and a manufacturing date of 2003 had a Type I Control PCB. So, that may mark the transition to the newer technology.
All the jumpers and their approximate locations are identical and the time spent in the major states and time to lock (READY on solid) are about the same as with the Type I Control PCB. The behavior while warming up and after locking is indistinguishable from that of the Type I Control PCB so it really isn't even possible to determine which one is inside the laser without removing the cover. The only obviously similar electronic component common to the two is the large white film capacitor for the feedback integrator, and perhaps the heater driver transistor. So, it's likely that the objective of this redesign was simply to eiliminate all the older SSI/MSI TTL logic and other obsolete through-hole parts, but that it is functionally identical to the Type I Control PCB with essentially the same logic inside the Xilinx FPGAs and linear circuitry in SMT ICs.
More recently, another type has appeared, called the "Type III Control PCB". I first found one in a 5517E (this model laser isn't documented anywhere except here!) and thought it might be unique to the -E version. But I have since also seen one in a 5517D-C29, manufactured in 2004. But the Type III Control PCB seems to be a total redesign, with no effort made to be at all similar to the Type I Control PCB with its jumpers and test points, or even in the specific algorithm it uses during warmup and locking.
Here are the three types of Control PCBs used in the 5517B/C/D/E lasers.
(As noted, the one in the 5517A differs very slightly in form factor and has a small amount of extra circuitry for use in the 5518A and 5519A/B. As far as I know, it only comes in the original Type I version.)
The Type I and Type II Control PCBs function in a virtually identical manner, requiring about 2 minutes for the READY LED to start flashing, and another 2 minutes to come on solid. And as noted, they also have more or less the same jumpers and testpoints, as well as the temperature set-point pot. The Type II Control PCB may in fact simply be essentially an emulation of the Type I Control PCB using an FPGA and more modern surface mount parts. That same large integrator capacitor is present, though the smaller sample-and-hold caps are not, or at least they aren't the same type.
The Type III Control PCB is not at all similar to the others. It has none of the same jumpers and several different testpoints, an unused connector and a large unpopulated header (functions unknown), and no pots at all. It does have an RS232 port no doubt for setup and testing and almost certainly for access to a digitally-maintained run-time meter. There are also a pair of micro-DIP switches - and a pushbutton, which I fianally dared to push, and as expected, seems to be master reset. :) It is based on a SHARC processor with many digital and analog SMT parts as well as large Lattice FPGAs on both the front and back of the PCB.
For more, on the Control PCBs and their operation, see the sections: HP/Agilent 5517 Laser Control PCBs and Locking Sequence, HP-5517E/F, and Agilent 5517 Laser RS232 Communications.
The entire purpose of redesigning the controller more than once is somewhat perplexing. Doing it the second time with such complex digital hardware seems totally nuts, unless it was actually done for some custom high performance application, and it was simply convenient to use in late model 5517 lasers. (Given the likely relatively high cost of components including the SHARC processor and Lattice FPLDs, I find this rather hard to believe though. And it's a multilayer PCB with components on both sides, additional connectors, and other stuff.) The same very limited inputs (a pair of photodiodes sensing the modes through the relatively slow speed LCD switch and another photodiode behind a polarizer generating the REF signal) and outputs (tube heater current) are used in all 5517 lasers so it wouldn't seem to be possible to implement a significantly higher level of frequency accuracy or stabilization no matter what sort of control scheme is used. About the only thing that might be done is to actually compute the REF frequency from the REF photodiode signal and fine tune the lock position to maximize it once the basic stabilization using mode balance has been achieved. The peak of the REF frequency function may be a more accurate means of locating the Zeeman-split gain curve center. But except for NIST-level precision, the analog method is really just fine, so even if this scheme is implement, it's not clear what customers would require it. And from my observations of the REF frequency while locking, it doesn't seem to make any effort to maximize it, but stabilizes at a point much lower than the peak, with the same sort of slow variation once locked as the Type I Control PCB.
So, a combination of several explanations make the most sense:
As of 2012, I've still only seen exactly two lasers with the Type III PCB. These both had manufacturing dates well before 2005, yet several lasers from as late as 2007 still use the Type II PCB.
If anyone has more information on these digital control PCBs, please contact me via the Sci.Electronics.Repair FAQ Email Links Page.
Also see the section: Common Problems with HP/Agilent 5517 Lasers.
Locking sequence with Type I Control PCB:
This also applies to the 5518A and 5519A/B since they all use the 5517A Type I Control PCB. It is also generally applicable to the 5501B though some details differ slightly.
From power-on to READY takes around 4 minutes for most 5517 lasers - all those NOT using the Type III Control PCB. Even on the original Type I Control PCB, a state machine based on counters, flip-flops, and gates determines the timing. This may be true of the Type II Control PCB as well, except that the state machine would be inside a Xilinx FPGA. Who knows how the Type III Control PCB with its SHARC CPU implements this algorithm (which tends to take much longer than 4 minutes, reason unknown)! The following is paraphrased from the 5517A manual, which assumes the Type I Control PCB implementation. (All timing is approximate as the main clock is a 555 timer on the Type I Control PCB!):
Thus, under normal conditions, the laser will be locked and ready to make a measurement (approximately) 150 seconds after the READY LED starts flashing. Note that the only check to make sure the laser is locked is that the REF signal is present. Since this only occurs for a small percentage of the entire longitudinal mode sweep cycle, REF will not remain on for long without active feedback, so this is a reliable test. The laser will in fact continue to repeat the above sequence forever if REFON is not detected. Typically, this will occur when the output power from the laser tube has declined to below the REF detection threshold of the internal optical receiver after years of hard work. However, some marginal lasers will go through the sequence several times when powered up as the output power from the laser tube gradually increases with warmup until the amplitude of the difference frequency signal exceeds the REF detection threshold.
Locking sequence with Type II Digital Control PCB:
The Type II Control PCB has 3 yellow state LEDs near the top right corner of the large Xilinx chip. These provide some information about where the controller is in the warmup process and they have a 1:1 correspondance with the major modes of the Type I Control PCB. I'm not sure the times in each state are identical for the two but they are close. Here is a rough chart of their behavior for a normal 5517C laser:
Time READY State Comments
--------------------------------------------------------------------
0:00 000 Power on (WARMUP Mode)
0:01 001
0:02 000
0:07 00X 001-000-001-000 in three seconds.
0:10 000 Remain here 3 seconds.
The previous two entries repeat approximately 14 times,
dependant on time to reach set-point temperature.
1:26 Blinking 010 (HEATER QUALIFIED Mode)
1:32 Blinking 01X 011-010-011-010 in three seconds.
1:35 Blinking 010 Remain here 3 seconds,
The previous two entries repeat approximately 16 times.
3:03 Blinking 110 (OPTICAL Mode)
3:09 Blinking 11X 111-110-111-110 in three seconds.
3:12 Blinking 110 Remain here 3 seconds.
The previous two entries repeat approximately 9 times.
3:58 ON 000 (LASER READY)
The blink rate for READY is about 1.5 Hz.
Locking sequence with Type III Control PCB:
The Type III Digital Control PCB seems to go through many more gyrations during warmup than either of the others, including several times where READY flashes multiple times separated by a period of inactivity, and then finally flashing READY continuously for two minutes until it locks - the latter being similar to what the Type I Control PCB does. The entire process consistently takes much longer than the 4 minutes typical of a laser with the Type I Control PCB, up to 10 minutes or more. The behavior is not obviously different whether a weak (but functional) laser tube, or one that greatly exceeds minimun output power specs is used, though it may take slightly longer with a below-spec tube. After all this, the end result seems to be exactly the same.
Here is a chart of the typical startup behavior for a very healthy 5517B tube installed in a (previously) 5517D laser with this Type III Control PCB:
Time READY State Comments -------------------------------------------------------------------------- 0:00 1111 Power on 0:01 0000 0:02.0 0001 0001-0010-0100 sequence in less than 0.5 sec. 0:02.1 0010 0:02.2 0100 0:03 Flash 1100 MSB LED and READY LED flash briefly. 0:04.0 0001 0001-0010-0100 sequence in less than 0.5 s. 0:04.1 0010 0:04.2 0100 0:05 Flash 1100 MSB LED and READY LED flash briefly. 0:06.0 0001 0001-0010-0100 sequence in less than 0.5 s. 0:06.1 0010 0:06.2 0100 0:07 Flash 1100 MSB LED and READY LED flash briefly. 0:08 0100 3x(Step 0, Step 3, Step 5) sent on RS323 port. 0:15 0101 "LASER" sent out RS232 port. 1:00 0110 1:35 Flash 5 X110 1110,0000,5x(1110,0110). 1:40 0110 2:14 0000 2:15 Flash 5 X110 1110,0000,5x(1110,0110). 2:20 0110 2:55 Flash 6 X110 1110,0000,6x(1110,0110). 3:01 0110 3:35 Flash 7 X110 1110,0000,7x(1110,0110). 3:42 0110 4:12 Flash 8 X110 1110,0000,8x(1110,0110). 4:20 0110 5:02 Flash 8 X110 1110,0000,8x(1110,0110). 5:12 0110 5:40 Flash 10 X110 1110,0000,10x(1110,0110). 5:50 0110 6:45 Flash 32 X110 1110,0000,32x(1110,0110). 7:17 Flash 96 X111 96x(1111,0111). 8:53 ON X000 1000,0000,1000,0000,...
The State refers to the 4 SMT LEDs above the upper left corner of the Lattice chip near the center of the PCB. The MSB is green while the three LSBs are red. All times are approximate. "Flash" is just the briefest pulse of light. "Flash n" denotes "n" flashes at a 1 Hz rate with a 50 percent duty cycle The duration for the 1110,0000 state changes in each "Flash n" sequence is relatively short (perhaps 100 ms for each of the two states). The minimum value for "n" seems to be 5, but it tends to increase as the laser warms up. (I'm not positive it's monotonically increasing though.) Once the REF frequency can be sustained by the feedback loop, it continues flashing for 32 seconds, and then switches to state 0111 for 96 seconds prior to becoming READY. Until that time, the READY LED and the MSB state bit track each other almost perfectly. But then, the MSB state bit (1000) continues to flash (but now at about a 1.2 Hz rate) while the READY LED remains on solid, And there is just a hint of the 0100 state bit flickering dimly, possibly the actual feedback loop in operation. :)
Multiple runs from a cold start may differ slightly in the number of "Flash n" sequences and the values of n, as well as other details, but always take much more than 4 minutes (typical of the Type I Control PCB). The very healthy tube will lock in 7 to 9 minutes while a weak but usable one might take 11 minutes or more. In all cases where the laser successfully locks, the last two minutes will be identical in behavior to that of the other two Control PCBs with READY flashing continuously until it stays on solid. A tube that is very weak or with no detectable beat (REF) frequency will result in only occasional very short abortive flashes, and no conclusion (at least not in 15 or 20 minutes).
One annoying difference between this Control PCB and the others is that the signal level for REF and ~REF seems to be much lower - about 2 V p-p open circuit and less than 1 V (maybe as low as 0.5 V p-p) terminated, instead of more than 5 V p-p, and the 5508A display apparently doesn't accept this as a valid signal. So even if the laser comes READY within 10 minutes (the maximum allowed by the 5508A), it still comes up as "Laser Fail", which isn't recoverable without power cycling the 5508A (which means the laser as well if it receives DC power from the 5508A). However, my home-built SG-MD1 display has no problemss. :) I wouldn't be at all surprised to learn that the signal levels are programmable - somehow.
Here is a chart of the typical startup behavior for the 5517E with its similar Type III Control PCB. The tube is probably below the Agilent spec for minimum power, but locks without problems so the sequence of event is probably not affected singificantly:
Time READY State Comments ----------------------------------------------------------------------- 0:00 1111 Power on. 0:01 0000 0:02.0 0001 0001-0010-0100 sequence in less than 0.5 s. 0:02.1 0010 0:02.2 0100 0:03 Flash 1100 MSB LED and READY LED flash briefly. (Repeat the previous 4 events 27 times.) 1:15.0 0001 0001-0010-0100 sequence in less than 0.5 s. 1:15.1 0010 1:15.2 0100 1:16 Flash 1100 MSB LED and READY LED flash briefly. 1:17 0100 3x(Step 0, Step 3, Step 5) sent on RS323 port. 1:27 0101 LASER sent on RS232 port. 1:52 0110 2:25 0000 2:26 Mode 16 XXX0 16x(1110,0000,....,0000). 3:35 Mode 12 XXX0 12x(1110,0000,1110,0000,....,0000). 5:10 Flash 15 X110 1110,0000,15x(1110,0110). 5:40 0110 5:48 Flash 32 X110 1110,0000,32x(1110,0110). 6:10 Flash 96 X111 96x(1111,0111). 7:46 ON X000 1000,0000,1000,0000,...
The last part of the sequence is essentially identical to that of the other laser, but the initial behavior differs significantly. This one appears to keep track of the mode cycles, or at least flash the State LEDs in response to them! "Mode n" denotes "n" times where the Zeeman beat is on, at least momentarily. Also note that the READY LED only tracks the MSB State bit near the end. I assume that the AM29F040B (4 Mbit flash memory) is the firmware NVRAM but there is no version number so I don't know that they differ, Bbut they must as everything else on the two Type III Control PCBs appears identical including the DIP-switch settings.
For more on the 5517 laser Control PCBs, see the section: HP/Agilent 5517 Laser Construction.
Note: This default procedure is what would be found in the 5517 operation and service manual. It's possible that to "fudge" the specifications, slight modifications may have been made during factory testing. I've seen several lasers where the set-point is significantly higher than would be accounted for by this procedure. I doubt any of these lasers had been tampered with, or that the setting drifted that far on its own. A higher temperature set-point may have been used to slightly increase output power and/or to slightly reduce the REF/split frequency.
When installing a replacement tube in any 5517/18/19 or 5501B laser, the temperature set-point should be adjusted. This is the only user adjustment for these lasers. While the laser may appear to work fine without performing this adjustment, doing so will assure that lock will be maintained over the spec'd temperature range for the laser. A DMM (preferably with a clip lead on the negative probe) set to measure 200 to 400 mV, a medium flat blade screwdriver to remove the cover on the small (rectangular) lasers or Philips or Torx for the large lasers, and small flat blade screwdriver to turn the trim-pot will be required.
The laser should be unpowered for at least 2 hours prior to performing the temperature set-point adjustment:
This specific value of 1.285 applies to the 5501B and all 5517s using a "long" tube. This includes the 5517A/B/C and older 5517Ds. Those using a "short" tube like some newer versions of the 5517D, as well as the 5517E/F/G may require a different value. On a 5517DL manufactured in 2008, it was found to be around 1.37. Since that laser locks in the normal time and is fully functional, the higher value is assumed to be close to optimal. This value is independent of the controller. So, if installing a different type tube, the controller should be adjusted using the appropriate value for the tube type.
While HP/Agilent lasers are very good for their intended metrology applications, they can't compare to the best stabilized HeNe lasers like those from Laboratory For Science, Spectra-Physics, and others. There are issues with both short term variation in optical frequency as well as long term frequency drift. The 3 most significant are probably:
Replace the HeNe laser power supply with a low noise/low ripple type or add an external ripple reducing circuit to its output. The older VMI 148 had particularly high ripple, but even the VMI 217 can stand improvement. I have not tested the older Laser Drive 111-Adj-1 or the newest VMI 373 for ripple. But the VMI-373 already has a ripple reducer built in. See the section: Reducing Ripple and Noise in a HeNe Laser Power Supply with a Switchmode Regulator.
(I'm assuming modifications to the common Type I Control PCB. I do not know if it's possible to do this easily to either of the digital control PCBs. At the very least, cuts and jumpers would be much more difficult on the dense surface mount PCB. And, since there are so many of the older Type I Control PCBs available, why would you want to!)
Remove the LCD panel and its photodiode. Drill a hole in the beam sampler PCB and mount a polarizing beam sampler (e.g., polarizing beam splitter cube and a pair of silicon photodiodes) on top of the PCB. (It might even be possible to build this into the plastic housing instead.)
The schematic for one possible modification is shown in Upgraded Electronics for HP-5517 Lasers 1. This references the part numbers found on the 5517A/B/C/D Type I Control PCB, and probably the 5518A and 5519A/B as well.
The dual trans-impedance preamp for the photodiodes generates separate S and P mode signals. These feed the "Subtracting-Sample-and-Hold" circuit modified so that when in "Optical Mode" under feedback control, it passes both straight through - no holds allowed! During "Warmup Mode", it must pass the normal heater drive signal. The added preamp can be made from any stable dual op-amp mounted on a little circuit board perhaps stuck on top of U12, the LF13331D quad JFET analog switch, and attached to the photodiodes via twisted pairs. A 1M ohm pot in parallel with a 1 nF capacitor should suffice for the op-amp feedback, providing enough gain for all but the weakest laser tubes. The op-amp, U101, isn't critical - something like an LT082 would suffice. A few cuts and jumpers will be required, but on the wide open through-hole layout of the Type I Control PCBs, that shouldn't be too difficult. An alternative would be to remove the LF13331D and install an IC socket in its place. Then, build a little PCB that plugs into that with the LF13331D and preamp circuitry on it. Add an offset pot and it will then be possible to fine tune the optical frequency. It may not end up pretty, but should work great! It may be easiest to do the modifications in two stages: First replace the LCD and its PD with the polarizing beam sampler and preamp, and confirm that the correct polarizations are selected - the system should lock normally. Then disable the LCD selection logic so that both signals are passed at all times.
I later implemented a simpler set of modifications as shown in Upgraded Electronics for HP-5517 Lasers 2. This should produce similar results but with a wee bit less flexibility:
Wiring of the lower "POWER AMP DRIVE" switch (U12B) was unchanged (enabled by "DISABLE").
See Modified Beam Sampler and Offset Adjust Circuit for HP/Agilent 5517 Laser for a photo of these modifications.
The first two sets of changes were implemented first. These worked fine with the locking characteristic after warmup, total time-to-lock, uncertainty in REF frequency, and slow oscillation in REF frequency appearing very similar to the behavior of an unmodified laser. This is actually a rather surprising and unexpected result, so more study will be required. :) A discrete time system has been converted into a continuous time system without doing anything to the loop parameters and there were no dramatic changes the system response. Interesting.... However, later I did confirm that actual locking to the modes occurred almost immediately after the laser entered "Optical" mode (about 100 seconds after READY starts flashing). I also confirmed that if the photodiode polarity was incorrect, it would repetitively pass through the lock point at a rapid rate but never stabilize there. I had expected it to lock to the opposite crossover point of the two modes, but apparently the slope there is so much different that it never latched on, so to speak. Or, possibly it would have locked there eventually but I did not wait long enough.
For the record, the laser first tested with these modifications was a somewhat high mileage 5517C with a power output of around 240 µW and a REF frequency of around 3.3 MHz, the latter being outside the spec'd range for the 5517C (2.4 to 3.0 MHz). The uncertainty in REF frequency may be 200 Hz or more. The variation starts out with a period of around 16 seconds and deviation of around 0.003 MHz. Over a few hours, it slows down and finally stops (or becomes so long as to not be obvious).
Some tests:
Then a 220K ohm resistor placed in parallel with the 2.15M ohm resistor. This also had no detectable effect once locked. But, while the time-to-lock didn't change that much, the locking behavior was more rapid after the initial warmup.
Later, I installed the modified control PCB and beam sampler in a certifiably healthy low mileage 5517B (510 µW, 2.32 MHz). Locking was fine and both the randomness and periodic variation in REF were still present, though subtly different. The amplitude of the randomness was slightly lower - perhaps averaging 50 Hz compared to 200 Hz. The period started at about 10 seconds and the deviation was about 0.0045 MHz. However, running this laser with its original Newer Digital Control PCB resulted in essentially identical behavior. However, the deviation as well as the amplitude of the randomness also appear to be affected by exactly which longitudinal mode pair (i.e., exact temperature) at which the laser locks. A later power-on cycle resulted in a deviation of almost 0.01 MHz. Turning my offset control too far (accidentally!) resulted in the laser losing lock and then reaquiring it after the offset was turned back toward center. But the behavior had changed! The deviation in particular had dropped from 0.01 MHz to 0.004 MHz or less. Nothing else was different other than (presumably) where it locked!
The detailed character of these artifacts remains a mystery. The randomness may in fact be a faster but lower amplitude oscillation in REF frequency superimposed on the larger slower one but it's hard to tell without actually recording it, which I'm not sure I am eager to do. :) Since other evidence suggests that there isn't a corresponding variation in optical frequency to go along with the variation in REF frequency, this peculiarity may be a fundamental characteristic of the Zeeman laser and have nothing to do with the stabilization at all. Or, they may be the result of some sort of etalon effect. The time constant of the slow down in the periodic variation in REF frequency is too long to be anything but thermal in origin. HP/Agilent laser tubes have at least 4 planar uncoated glass surfaces outside the laser cavity and these are not wedged or set at an obvious angle to minimize back-reflections. In addition, the optics of the beam expander telescope and beam sampler have several optical surfaces. Since the structures these are mounted on are all mostly temperature independent of the controlled thermal environment of the mirror spacing rod, it's possible that one or more are forming some sort of external resonator with its longitudinal modes interfering with the normal lasing process very slightly. Maybe.
And eventually, I will have to set up the dual laser setup to check the optical frequency stability.
A second order effect is external magnetic fields, but this really shouldn't be significant unless other Zeeman lasers are living nearby, or you want to run this inside an MRI machine. :) And for the purist, air pressure and seismic activity also affect optical frequency, but the three modifications described above should reduce the short and medium term (up to days, probably not years) variation by more than an order of magnitude. Long term drift of optical frequency will be dominated by changes in the laser tube gas pressure and fill ratio from use, and this can't be easily controlled. But periodic diddling with the offset pot can compensate for those. :)
And, no, there is nothing labeled "RS232 Port", even on the Type III Control PCB. But, there was a header a with a suspiciously appropriate number of pins (10) near the SHARC chip, so I started looking at voltages, and sure enough, pin 3 on the header had -9 VDC on it, and was occasionally pulsing to 0 V. So, I made up a cable to my ancient Kiwi laptop, and sure enough, there was ASCII being spit out at 9600 baud! :)
Header Pin DB9 Pin Signal
----------------------------------------------------
3 2 Data from 5517 (transmit)
5 3 Data to 5517 (receive)
9 5 Ground
The "DB9 Pin" is the result of using an IDC cable wired directly to a DB9 connector. These may be salvaged from old PCs as they are often used to attach the mainboard to the rear panel. The pin numbers will be the same on the PC (not swapped). It's 9600 baud and full duplex (the laser echos characters typed). I have no idea about start and stop bits and parity, but suspect they don't much matter.
The few interesting things I've discovered so far are:
The complete sequence of what's sent from the RS232 port with a working tube (for either controller) from start to finish is:
Step 0 Step 3 Step 5 Step 0 Step 3 Step 5 Step 0 Step 3 Step 5 LASER READY
That's it! See, I told you it wasn't very exciting. :)
A below spec 5517FL that required over 15 minutes to lock produced a very slightly lengthier output (also beginning a couple minutes after being powered on):
Step 0 Step 3 Step 5 Step 0 Step 3 Step 4 Step 5 Step 0 Step 3 Step 5 LASER F8 Step 0 Step 3 Step 5 Step 0 Step 3 Step 5 Step 0 Step 3 Step 5 LASER READY
It's possible that this laser aborted due to the low power and generated the "F8" (whatever that means) and then tried again. Its output when READY came on solid was only 52 µW. (The minimum spec for the 5517FL is 65 µW.) But notice the "Step 4" thrown in near the beginning! Another mystery.
Perhaps flipping one of those DIP-switches will put it into Verbose mode, but picking the wrong one might erase the Universe, so I'm not willing to risk that - just yet. :)
At the very least, the runing time is probably maintained in NVRAM and it would be nice to know how to access that!
If anyone has more information on these digital control PCBs and their RS232 or other diagnostic port, please contact me via the Sci.Electronics.Repair FAQ Email Links Page.
More to follow, maybe, but these "First Contacts" are encouraging. :)
HP-5517 power/reference connector J2
Pin Function
------------------------------------------------------------------
A No Connection on 5517 (MEAS signal level on 5508A) (1)
B ~MEAS (Not used on 5517) (2)
C MEAS " "
D Signal Return (MEAS)
E ~REF (Zeeman beat signal from internal optical
F REF receiver's differential line driver.)
G,H Ground
J +15 VDC Sense
K +15 VDC
L -15 VDC
M +15 VDC
N,P Cable Shield
R Signal Return (REF)
S Ground
T +15 VDC
U Cable Shield
Notes:
The 5517 laser head connector looks like a standard MIL-style bayonet lock type but apparently is special built for HP by Amphenol. The part number may be PT06A-14-18PZ but this has not been confirmed. It's possible the mate to these is available direct from Amphenol or a direct distributor, but probably not a standard item at an electronics distributor. It's actually the same connector as used on the 5500C laser head (PT06A-14-18P), but the keying is rotated 270 degrees, not that that helps much. I did check a Mouser catalog and indeed, the standard connector has the keying rotated by 270 degrees like the 5500C. One mating connector from an original HP cable is labeled: 97 USA/CTI 26SOU 851-06P14-18PX50-44. Searching for any of these part numbers only seems to result in non-stock items with no listed prices - "Ask for Quote". You know you're in trouble when this is involved! :) Used cables for the 5517 are available for around $100 from various surplus dealers and often on eBay. But the standard cables may be 10 or 20 feet long and much more than is needed if the other parts of the interferometry system aren't being used. For power and the reference signal, the mating connector and a few wires should suffice. It would be a pity to chop up an expensive high quality cable simply for the connector. However, it is possible to modify a standard PT06A-14-18P See the section: Making HP Interferometer Cables.
The pinout for HP 10881A/B/C cables with a 5 pin DIN connector (looks like an old PC AT keyboard connector) for power is as follows:
Pin Function 5 Pin DIN Female
------------------------ 2 o
1 Ground 4 o 5 o
2 Ground
3 No Connect? 1 o 3 o
4 -15 VDC ___
5 +15 VDC [KEY]
The color coding of the power wires for the 10881D/E/F cable (which has only spade lugs) is strange:
A working HP/Agilent HeNe laser power supply brick, 15 VDC power supply rated at 2 amps or more, and a laser power meter will be required. The power supply doesn't need to be well regulated for these tests, but it should be well filtered and not go much above 15 V even unloaded. A simple wiring harness should be constructed with mating headers for the 2 and 3 pin connectors on the HeNe laser tube and HeNe laser power supply, respectively:
Both wires to the 15 VDC power supply should be #18 AWG or larger.
Note: If the connector on the HeNe laser power supply brick has a white wire in the center position, it requires a control input and cannot be used for testing when not attached to the (rear) Connector PCB using this simple procedure. Most do not, so find a different one. :-)
Now for the testing:
Both these connectors are keyed and only go one way, don't force. :)
WARNING: The high voltage on the HeNe laser tube and power supply takes some time to discharge, up to a few minutes or even more depending on the version of the HeNe laser power supply and other factors. Take care when disconnecting the tube. It's not dangerous but could be a slightly shocking experience.
The first one (closest to the tube) is a 1/4 waveplate which converts the left and right circular polarized Zeeman modes to F1/F2 linear polarization. The second one (at the output) is a 1/2 waveplate that rotates them so that they are aligned with the X and Y axes of the laser. If the output of the laser tube was truely pure circularly polarized, then the 1/2 waveplate would be unnecessary as the orientation of the 1/4 waveplate could be set to do this. So, the raw output of the tube must be asymmetric in some way as indeed, there is usually a preferred orientation for the 1/4 waveplate, though the effect is not usually that strong. (Excel metrology lasers which also use Zeeman splitting and are functionally similar to the HP/Agilent lasers only have a 1/4 waveplate. It's not known whether the tubes Excel used are not subject to the asymmetry, or whether they simply concluded that the improvement resulting from the use of a 1/2 waveplate was not worth the effort and cost.)
My first comment on doing anything to the waveplates in HP/Agilent is: "if it ain't broke, don't fix it.". :) I have *never* seen an HP/Agilent laser requiring waveplate adjustments. The mode purity and orthogonality are always very good or excellent even for 20 year old lasers. These simply do not change on their own. However, adjustments may be required if moving a waveplate assembly from one tube to another, or between a 5517 and 5501B or vice-versa. Or, if someone before you (we won't name names!) attempted to "repair" the laser by turning everything that could be turned and moving everything that could be moved.
I don't know how the waveplates are set up at HP/Agilent but what I do is to maximize MEAS signal quality while monitoring in an interferometer and minimizing the MEAS signal strength with a polarizer oriented along the X or Y axis. With the stage in motion, the signal waveform should remain clean with minimal jitter. With the stage stationary and the polarizer aligned with X or Y, the signal should approach zero and a 10780 optical receiver set at maximum sensitivity will drop out (signal detect LED goes off and there is no output) except possibly with a very high power laser. The primary adjustments are the orientation of the inner barrels of the waveplate assembly - the one closer to the tube being the 1/4 waveplate. Only if these do not succeed in achieving acceptable signal characteristics, should adjusting the tilt (outer barrels) be contemplated. And before this is done, make sure you're not simply being obsessive compulsive - it's possible to easily make things much worse. Marking the original orientation of both sets of adjustments is strongly recommended - the HP/Agilent approved blue paint is not adequate for this.
While it's possible to adjust the waveplates at the output of all HP/Agilent lasers while attached to the tube assembly, locking only occurs over a small range of orientations of the 1/4 and 1/2 waveplates. If lock is lost, the laser takes a minute or more to reacquire lock, and only if the setting is within that narrow range. If one forgets where a "good" setting is, this can end up being a total unknown and frustrating task. However, for the 5501B and all 5517s, setting the REF jumper to "LO" (or the position next to NORM, which is the same thing) before turning the laser on will force the controller to think there is a valid reference signal once it goes into the locked state, regardless of whether it's locked properly or not. While adjusting the waveplates will still cause lock to be lost, it will be reacquired almost immediately when they are within a fairly wide range. Then it isn't necessary to wait for the state machine to do its thing, taking minutes, since the locking remains under analog control regardless of what's happening with the reference signal. Note: Do NOT move the REF jumper away from NORM while the laser is warming up as it may get confused and switch to analog feedback before reaching the set-point temperature.
However, even this isn't ideal since lock can still be easily lost. What's really needed is a way to maintain the lock point while still allowing arbitrary adjustment of the waveplates. So, I modified a spare 5517B laser chassis so that the waveplate assembly could be removed from the tube and mounted out front where any adjustments only affect the MEAS beam. Then it's easier to set both waveplates to optimize signal strength and quality without worrying about the laser losing lock.
At first, I figured the easiest thing to do would be to add a beam-splitter and direct a portion of the main beam through a 1/4 waveplate to the beam sampler mounted on the side, extending the wires for the LCD switch and photodiode, and adding a mount on the front for the waveplate assembly being adjusted. Using only a 1/4 waveplate should be adequate for locking, even if the F1/F2 orthogonality isn't optimal (normally fine tuned using the 1/2 waveplate present in all HP/Agilent lasers). In fact, I had half completed that kludge with a wart hanging off the laser when I realized that it would be much easier to simply add the 1/4 waveplate *inside* the beam sampler just beyond the first 45 degree beam-splitter mirror. So, that's the way it was done in the end: The 1/4 waveplate "pellicle" was popped out of the waveplate assembly from a 5501A laser tube and stuck in a slot cut in the plastic beam sampler housing. The waveplate axis (which is marked on the pellicle) must be at a 45 degree orientation. It isn't worth attempting to determine ahead of time whether it should be +45 degrees or -45 degrees, so flip a coin. :) If it tries to lock at the wrong location, reverse it. One caution about doing this is that the pellicle - made of optical grade mica or something similar - is extremely fragile. Even a slight bend will delaminate it and introduce a permanent unsightly blemish. However, even when damaged in this way, it works well enough and the laser locks normally. For locking, neither the 1/4 waveplate orientation or quality is particularly critical. As long as it's close enough locking will be reliable but F1 and F2 may not be exactly equal.
The result can be seen in: HP/Agilent 5517B/C/D and 5501B Waveplate Adjustment Adapter. The 3 hex standoffs are mounted so the waveplate assembly is a snug fit and only a single screw at most is needed to secure it.
This same approach could also be used with the 5517A (or 5518A or 5519A/B) except that removing the waveplate assembly from one of these laser tubes requires a combination of a jack hammer and TNT as they are attached with 5 ton adhesive. :) However, the beam sampler assembly is identical so the 1/4 waveplate version could be installed in any of the "modern" lasers (but not the 5500A/B/C or 5501A).
Once modified, the laser can't be used with a normal (complete) laser tube assembly (with its waveplate assembly installed) since there is already a 1/4 waveplate inside the special beam sampler. But the modified beam sampler assembly with waveplate adjustment adapter attached can easily be swapped with a normal one when testing complete tube assemblies.
Note that with the 1/4 waveplate inside the beam sampler set up so it will lock with a 5517 controller, no change is required to adjust waveplates for a 5501B tube as the lock point is the same. Its waveplate assembly will be adjusted with the F1/F2 axes swapped. Aside from the REF frequency difference, the waveplate orientations are what distinguish a 5517B/C/D from a 5501B or vice-versa. With no waveplates on the tube assembly, they are identical (except for the REF frequency).
With this scheme, it became almost trivial to adjust a bunch of waveplate assemblies for optimal performance. :)
A missing or low -15 VDC supply will not prevent the tube from operating but the laser will never lock.
The tube and HeNe laser power supply in the 5517, 5518, 5519, (and 5501B) can easily be tested using a 15 VDC power supply without powering the rest of the laser. The power supply must be regulated and capable of a current of at least 1 amp:
There still can be problems once power is applied to the heater inside the tube but at least this test proves that the tube isn't dead. Also note that if output power is measured in this unlocked state, it may be 50 percent or more greater than once locked.
There is no practical way to boost output power from a weak tube, at least not significantly and without side effects. But note that the power may increase 10 percent or so with an hour warmup compared to what it is just after locking. So, the situation may not be quite as bad as it appears initially.
However, a laser with power below the 180 µW HP/Agilent specifications may still be very usable, especially with simpler interferometer setups or fewer interferometer axes. The optical receivers are quite sensitive and only a few µW is sufficient for a stable beat frequency signal. With my crude setup using a 10780A optical receiver at its default threshold setting, 12 µW from the laser resulting in 8 µW at the receiver is sufficient power. Adjusting the 10780A's threshold setting or using a 10780C optical receiver which is more sensitive would require even less power.
But a low power laser may have other problems. See below.
All Type I Control PCBs made after somewhere around 1990 are wired this way yet the PCB artwork was never updated. I don't know whether this modification was done to prevent locking where the REF signal is so low that it might be corrupted by amplitude ripple resulting from HeNe laser power supply current ripple, or simply to sell more lasers since they will fail to lock sooner. :)
For late model 5517s, the Type II Control PCBs will lock below 25 µW if the "REF" jumper is set to "LO" and they will generate a REF signal, though it's not known what threshold is needed to operate normally. The corresponding parameters for the Newest Digital Control PCBs are not known at all. See the section: HP/Agilent 5517 Laser Construction.
There are two other work-arounds for an inability to lock due to low power where you don't want to modify the Control PCB:
The first of these is probably preferred as it doesn't require permanent modifications to the laser. An external 5 or 10 percent beamsplitter and optical receiver are required but the beamsplitter can probably just be a microscope slide at 45 degrees and any of the optical receivers would be suitable for generating the reference since it operates at a fixed frequency not greater than 4 MHz regardless of the laser model. The obsolete 10780A or 10780B can be obtained very inexpensively.
It may also be possible to change a component in the reference detection circuit but I have so far been unable to obtain a datasheet for the actual IC that is used there - HP part number 1826-0775 or the manufacturer's part number 1DA7Q (assuming this isn't simply some random collection of characters that was never updated since Google has no clue about it!).
There is most likely nothing really wrong with the laser but the tube is simply high mileage or one that has difficulty starting. Some tubes become like this. Even the Ph.D. types at a major laser company really don't know why. Aside from time wasted twiddling one's thumbs, a tube that takes a long time to start is hard on the HeNe laser power supply, but the ones in these lasers seem tough. Do check the DC voltages, particularly for the +15 VDC supply, which is what powers the HeNe laser power supply. While the rest of the laser may run on 12 VDC or below, lower DC voltage means proportionally lower starting voltage for the tube. And lower voltages may be more stressful both on the HeNe laser power supply and other parts of the laser, as well as being more likely to take a long time to start with an uncooperative tube.
In several instances, I found that shining a light on the *back-end* of the tube would promote starting in an otherwise uncooperative laser. Electrical discharge initiation is known to be sensitive to light and radioactivity, so this effect isn't entirely surprising. A radioactive source would work but putting a radiation warning sticker on the laser might invite a visit from Homeland Security. So, I opted for LEDs instead. :) For some tubes, a high brightness red LED shining on the glass extension at the back of the tube is sufficient to reduce the starting time from a minute or more to a couple of seconds. A blue LED seems to be even more effective especially for particularly uncooperative tubes. The LED can be conveniently wired to the HeNe laser power supply on the back of the connector PCB. I now routinely install an LED if there is any hint of slow start. It can't hurt and so far, has seemed to help significantly. For a higher level of sophistication, add a circuit to turn off the LED once the tube starts!
It's also possible that electrical leakage is reducing the effective starting voltage. If there is a smell of ozone while the tube is trying to start, then corona is present from the anode terminal.
CAUTION: Take care to avoid stressing or bending the wire connections to the tube terminals when removing the old foam and cleaning and insulating the anode, as well as reinstalling the tube assembly in the laser. Avoid applying force (especially side-ways) to the cathode/heater terminals and their glass-to-metal seals, and avoid repeated bending of the anode wire since re-attaching that should it break off could be challenging.
Attach a multimeter on DC Volts across the 1K ohm resistor. The reading will be 1 V/mA. Power up and start at the default current setting for the internal HeNe laser power supply of 3.5 mA. If increasing the current results in a stable output, then the problem is almost certainly the dropout current as noted above. The current will need to be slightly beyond where the laser is stable. 3.75 or 4 mA shouldn't hurt it or significantly reduce life expectancy. There's no choice anyhow as this may be the only practical way to get these tubes to stay lit! If the tube is unstable even at 4 or 4.5 mA, then the problem may be the power supply, or the ballast resistor attached to the tube (quite unusual).
Limited anecdotal evidence suggests that a laser repaired in this manner will run continuously with useful power for several months. And, of course, if only turned on when needed, for much longer.
Adding an anode ballast resistance without increasing the laser tube current may work in marginal cases. But in my tests, even as much as 35K ohms only reduced the dropout current by 0.1 or 0.2 mA. So, it alone is probably not a reliable solution for a tube that doesn't stay lit. But adding some modest anode ballast resistnace (10K to 20K) is worth doing to reduce the chance of amplitude ripple as discussed below.
CAUTION: DO NOT allow a laser to continue sputtering for a long time. This may damage the laser tube and destroy the power supply. I've had 5517 lasers where the HeNe laser power supply had been blown due to unattended sputtering, though it's not clear if there was any damage to the tube.
But with one somewhat high mileage 5517B, sputtering for 10 or 20 seconds seems to have done something bad, from which it may or may not recover. This laser produced around ~180 µW at READY, 200+ µW fully warmed up and had been consistent over several months of occasional power cycles. I was intending to sell this as a cheap emergency spare since it seemed to be reliable and had a like-new REF of 2.12 MHz. Although less than 1/2 the output power when it was new, it would meet specifications. Or that's how it was. After completing tests before shipping and shutting down, I realized I wanted to check something else and switched the 5508A (which provides its power) back on. At that point, the laser did not start and started sputtering. Power cycling a couple times didn't help, but leaving it off for a minute or so allowed it to start up and come READY normally. However, the output power had dropped to 119 µW with a REF 2.4 MHz. After running for 48 hours, and power cycling multiple times approximately 2 hours on/2 hours off, the output power climbs to just over 200 µW with REF back down around 2.12 MHz, thus near original condition. However, the behavior seemed to have changed as 2 or 3 hour-long power cycles were now required to rachet the power back up after being off for several hours. Possibly, such power cycling would have gotten even more power originally, but there had been no reason to try. Then the last time I tried it after being off for several days, it came back nearly to its original power and REF without multiple power cycles. Go figure. :) I have never seen a sputtering condition cause damage so quickly. In fact, I've had difficulty getting sputtering done deliberately even over an extended period of time to have any effect, and when it did, the power increased! So this is another mystery. The cause here is unknown, although some sort of release of contaminants is suspected.
On two 5501A tubes I tested, rapid sputtering as a result of a defective HeNe laser power supply caused the output mirror inside the tube to literally have a hole blown in the exact center of its coating, rendering the tube useful only as a magnetic paper clip holder/desk ornament or paperweight! One tube had a hole just about the size of where the beam would have been (or more likely, the bore) as can be seen in Hewlett Packard 5501A HeNe Laser Tube with Missing Coating in Center of Output Mirror. But the other had a clear hole in the coating over 2 mm in diameter!
Note that in a 5501A or 5500C, sputtering may either be due to a defective HeNe laser power supply (probably the potted module), the laser current being set too low, or the tube itself being unable to stay lit at any current setting. With the current setting being under user control, it's critical to set it so that the tube will stay lit. The current should be set according to the recommended value on the label (if any), but subject to the constraint that it be at least 0.2 to 0.3 mA higher than the dropout current after a 1 hour warmup to assure reliable operation in the long term. If there is no value listed, then assume 3.0 mA or adjust for maximum output power when locked between 3.0 to 3.3 mA, but subject to the same constraint. (The nominal operating current may range from 2.6 to 5.1 mA, according to the 5501A manual. But most are between 3.0 and 3.3 mA, so if there is no value listed, it's safer to keep it within this range if possible.) The current may either be measured by installing a mA meter between the tube cathode post (on the side of the large glass bulb of the tube) and its connecting wire, or by measuring the voltage on the laser current testpoint, which is series with a 390 ohm resistor to ground. So, the current will be V/390. The testpoint is accessible on the left side of the rear connector PCB, just above the laser current adjust pot, R11. Leave the right side cover in place to activate the interlock switch that enables the laser to turn on.
Aside from problems in the electronics, a very unusual cause might be an intermittent connection *inside* the HeNe laser tube. Broken welds are possible, but for HP/Agilent tubes, what's more likely is just bad contact with respect to the cathode terminal, a pressed-on slide fit in the 5501B and later lasers with a spring contact to the terminal post as shown in Closeup of Spring Cathode Contact Inside HP/Agilent 5517 Laser Tube. (The overall cathode connection spring can be see in the back of the top photo in Tube Used in HP-5517B Two-Frequency HeNe Laser.) At first I thought that the discoloration was due to overheating, but then I checked other tubes - even ones that were new rejects (don't ask) - and it was there as well. So, perhaps it is due to heat treatment to make the steel springy. :) But the post in all cases isn't shiny but the dull black, brown, or gray. Some early versions like 5501Bs may have the post against a single edge. And it's simply a metal tab pressing on the through-glass terminal in the 5501A. As the parts expand, the result is momentary loss of contact. I've never actually confirmed this in an HP/Agilent tube though I do have a 5501B that is suspect. While not that common, it does happen with conventional tubes. And the tubes used in Zygo 7701/2 lasers are notorious for a similar phenomenon, usually appearing after 10,000 or 20,000 hours of use. The cause is thought to be a bad contact between the evaporated metallic coating used as a cathode and the end-cap.
Needless to say, there is no truly guaranteed practical fix other than installing a replacement tube. However, such events are much more likely before the system reaches thermal equilibrium. So, simply running the laser for awhile before use may be sufficient to reduce the frequency of occurrence of these glitches to zero. And thus lasers run 24/7 may never experience them after the first few hours.
One time, I was 100 percent sure that a bad internal connection was the problem with a 5501A. But it turned out to be much simpler. The tube would drop out at random times anywhere from a few seconds to hours apart (generally less frequent after warming up). The tube was absolutely healthy in all other respects - great power, instant start, and stable over the full range of laser current adjustment. But the symptoms always remained with the tube when it was installed in two known good laser chassis. The tube was even connected to a stand-alone HeNe laser power supply and then, tapping on the tube would sometimes induce a dropout. However, I was suspicious of the anode contact as jiggling the HV wire would also tend to cause dropouts. And, indeed, with the front optics assembly removed, the anode terminal was found to be only a short stump (probably original) flush with the glass. And there were also bits of RTV Silicone stuck to it (origin unknown). I had tried to clean that terminal early on in this saga with no change in behavior, but RTV Silicone bits don't come off easily, especially if they were visible! So, they were preventing the spring contact from seating firmly against the terminal. Or something. :)
Another possible cause of dropouts is a bad ballast resistor. This is also extremely rare because the ballast resistor is conservatively rated and only dissipates less than 1.25 W at normal current (3.5 mA). But an HeNe laser power failure resulting in excessive current for an extended period of time could damage the ballast.
Viewing the beat frequency from an optical receiver on an oscilloscope will show all edges except the one used for triggering the scope to be fuzzy as the bogus signal modulates the position of the zero crossings, rather than the clean waveform that is expected. (But check and touch optical alignment as poor alignment can also result in a fuzzy signal.) The oscillation itself will show up when only one polarized Zeeman mode is presented to an optical receiver (e.g., by blocking the return beam from the interferometer) or by using a photodetector and oscilloscope. In the latter case, it will be seen as a sinusoidal waveform that is present *without* a polarizer. With a polarizer, the beat frequency signal will be riding on top of the ripple. A typical ripple amplitude is 10 µW but this can vary greatly. It will also appear by itself at the output of an optical receiver while the laser is warming up between those times when the normal beat frequency signal is present.
The exact cause of the bogus signal is not known but it probably has to do with the tube's negative resistance. The ballast resistor for these tubes is located 5 to 6 inches from the anode (which is much longer than the 2 to 3 inches usually recommended for HeNe lasers) and the wire between the resistor and tube may run close to the grounded chassis, adding capacitance. So this certainly makes such problems more likely.
For tubes that meet HP specifications for output power (180 µW for most models), it is probably not necessary to do anything about this oscillation unless specific measurement issues can be directly tied to it. In fact, I've seen it in lasers that appeared to be virtually new in all other respects, so it may simply be considered normal!
There are two ways of eliminating the amplitude ripple:
Adding a cathode ballast resistor would probably eliminate the oscillation as well but this is not an option with 5517/18/19 or 5501B lasers since the cathode is attached to the heater used for thermal tuning inside the tube and it must be near ground potential. A cathode ballast resistor should be acceptable on the 5501A.
I don't know for how long these cures will be effective or whether they work in all cases. And sometimes, both will be required. If the increased current is needed to fix a tube that won't stay lit, try that method first.
With some lasers, there is amplitude ripple at a lower frequency, typically 50 to 100 kHz and adding ballast has little or no effect. This is due to residual current ripple in the switchmode HeNe laser power supply. The older VMI 148 has 1 to 2 percent current ripple which is enough to produce easily detectable amplitude ripple in the laser output unless the 3.5 mA (default) current is optimal (slope of output power versus tube current is 0). New supplies like the VMI 217 and VMI 373 have a built-in ripple reducer (active filter) which virtually eliminates this phenomenon unless it's broken. :( :) I haven't seen a laser where the low level ripple was of any consequence once locked though - the beat waveform is clean, especially with respect to the full cycle. If the scope is triggered on the rising edge, then there may be some fuzz on the falling edge, but not subsequent rising edges. And, it's generally very small.
From measurements of the field strength of the magnets in a variety of HP/Agilent lasers, it's clear that on average at least, higher REF frequency lasers have stronger magnets, though there is a lot of variability even for the same model (e.g., 5517B). If HP/Agilent can play with field strength, so can we! :)
So, the approach to decreasing the REF frequency is to slightly reduce the Zeeman magnetic field. Don't panic, there's no need to take the laser to an electromagnetic can crusher or ultra-high field Government magnet lab to zap it! All that's required is some duct tape and bailing wire. Well almost. :)
CAUTION: Not all of these approaches are fully reversible. So, be sure that this is what's really desired before converting your almost working laser into a paperweight!
There are several low tech ways of modifying the magnetic field strength:
This scheme worked much better than I had originally expected. See: Zeeman Frequency Reduction Using "Tin" Can Stock. It was trivial to decrease REF for a 5517C from 3.3 MHz (way out of the spec'd range of 2.4 to 3.0 MHz) to 2.8 MHz using steel from a two-seam can. And the output power climbed from 245 µW to 275 µW! Similar size strips from a one-seam can only brought REF down to 2.9 MHz. With both sets, REF dropped to 2.50 MHz and the output power climbed to 290 µW. Using the original thicker strips on an out-of-spec 5517B brought REF down from 2.49 MHz to 2.15 MHz (spec'd range of 1.9 to 2.4 MHz), which is probably lower than the value when the laser was new. The field on the outside of the magnet decreased by about 18 percent with one set of strips and 24 percent with two sets according to my crude measurement, but it's not clear how this translates to the field strength inside the magnet, and I'm not real confident of its accuracy anyhow. And several sets of strips on another 5517B brought REF down from a way out of spec 2.8+ MHz to 2.1 MHz, with power climbing from 275 µW to over 300 µW. In all tests, the REF and MEAS signals remained clean and free of artifacts at all times.
One benefit of thin steel strips is that their effect appears to be largely reversible if they are removed, so REF returns to nearly its original value. However, this may not always be true with shunts and is definitely NOT the case with the techniques described below.
Apparently, HP may have used the shunt technique at least when they had no other choice. Among my pile of dead HP/Agilent tubes, I found an ancient 5501B (1987) that had a 1/4x1/4x3 inch steel bar RTV'd in an inconspicuous location under the tube. At first I thought it was a bar magnet, but after removal, no residual magnetism could be detected, nor would it retain any magnetization when swiped on a stack of powerful ceramic magnets. The 5501B tube was quite dead with exactly 0 µW of output power so the bar's effect on it could not be determined. But when stuck to a 5517B, it decreased REF by about 0.35 MHz, which would be just about optimal. Perhaps the factory was in a pinch and needed a 5501B when none were available, so they decided to down-size a 5517B tube. Or perhaps it wasn't HP at all but some service company trying to squeeze more life out of a high-mileage laser and no Campbell's soup cans were handy! Would anyone really do that? ;-)
In summary, use thin steel strips where it is desired that the reduction be reversible. Otherwise, rods or bars - or in extreme cases, magnets, would also be suitable.
And to reiterate: None of these REF reduction techniques represent a fountain of youth for HP/Agilent lasers. The tube does not become any healthier but simply operates in such a way that a bit more life - probably measured in months, not years in 24/7 service - can be squeezed out of it while still being within specifications. There is no risk of creating rogue modes as that's only possible with a stronger magnetic field (see below), so this is a low risk procedure as far as the potential for interferometer errors is concerned.
While an unscrupulous seller might try one of these stunts to enable a nearly dead laser to be sold as "like new condition, meets all HP specifications", when used with full disclosure, modifications such as these could squeeze some additional life out of a laser otherwise useful only as a doorstop. And it's possible that a visual examination may reveal the remains of multiple Campbell's soup cans or 10 penny nails stuck to the magnet. ;-)
Placing the magnet assembly from a 5517B in contact with the magnet assembly of the 5517C having the 3.3 MHz REF resulted in a 4.1 MHz REF - higher than the upper limit of a 5517D (3.4 to 4.0 MHz). The output power did decline but remained over 200 µW. For all intents and purposes, the laser behaved normally with no obvious rogue mode problems. A more modest boost - if the magnets were separated by perhaps an inch - would have resulted in a very nice 5517D. Or, applying a similar magnetic field to a healthy 5517C would have also converted it to a 5517D. Since an entire 5517 magnet assembly duct-taped in place might not be very attractive, to be practical, bar magnets of some kind or pieces cut from an HP magnet would need to be secured to the magnet assembly. But, they won't stay on their own because the like poles will be repelling. Using a pair of Alnico bar magnets converted a 445 µW/2.40 MHz 5517B into a 435 µW/2.86 MHz 5517C. However, there is the real risk of slightly demagnetizing the Zeeman magnet with strong bar magnets. In this case, when the bar magnets were removed, the laser locked at 470 µW/2.32 MHz. Further manipulation resulted in a reduction to below 1.90 MHz. :( :) On a high mileage 5517B that started at 2.8 MHz, REF went below 2.0 MHz. Using somewhat weaker ferrite magnets may be lower risk in this regard, but more of them or larger ones will be required to achieve the same REF increase as their strength is lower.
However, in some cases, adding magnets will have little or no effect, or the opposite of what's intended even with the correct field orientation. For example, it was possible to convert a 350 µW, 2.70 MHz 5517C into a 300 uW, 3.4 MHz 5517D with a pair of 5/8x3" AlNiCo magnets. Without peaking inside, it would be impossible to detect that this was not a genuine 5517D. But using the same magnets on a 450 µW, 2.69 MHz 5517C had virtually no effect on the REF frequency, and with additional magnets, REF actually declined. The explanation for this behavior is simple: Rogue modes are appearing that are reducing the mode pulling effect even though the separation of the gain curves continues to increase. In fact, eventually, they will be so far apart that the output power of the Zeeman modes will go to zero and there will be no beat at all. This has been confirmed by monitoring the output of the 5517C on an SFPI as magnets were added. With 1 magnet added, it continued to lock, but a single rogue mode appeared with an amplitude of about 60 percent of the Zeeman modes and the lock point was actually significantly offset from the normal balanced position. While the normal balanced position would be an equilibrium point with the Zeeman-split modes and rogue modes on either side present, it apparently is not stable. Once locked, there was still a beat but it was actually at a lower frequency than the original 5517C since the existence of the rogue modes distorts the effective gain curves and reduces the mode pulling effect. With 2 magnets, only rogue modes were present, there was no beat, and the laser never locked. So such contra-expected behavior of REF with added magnets may actually be a reliable indication of the onset of rogue modes, though this is not known beyond a shadow of a doubt. ;-) But it is almost certain that the two 5517C lasers had different magnetic field strengths (probably around 250 and 350 G, respectively) so there was much more headroom with the first one. Performing these tests did result in a small lingering reduction in REF frequency when the added magnets were removed, of between 0.05 and 0.1 MHz. This is likely to be permament.
CAUTION: I do NOT recommend using high-strength rare earth magnets for any of this. Aside from the tendency to squash flesh and other vital body parts, there's no telling what excessive and irreversible effects they will have on the Zeeman magnet. For example, extreme demagnetization or a highly non-uniform field after they are removed may be all too likely. However, it's possible that through the careful use of super powerful magnets, the strength of the Zeeman magnet and thus REF could in increased slightly without requiring additional magnets to be glued in place. But there may not be any way to do this significantly or consistently. Having said that, by using a pair of rare earth magnets and appropriate manipulations, it was possible to boost the REF frequency of the healthy laser above by around 0.2 MHz, back up to about 2.10 MHz, which is safely in the acceptable range for a 5517B. This had not been possible using Alnico magnets. And on the high mileage 5517B that had its REF reduced from 2.8 MHz to 2.0 MHz using Alnico magnets, rare earth magnets were able to bring it back up to 2.30 MHz. However, successfully increasing REF appears to be a hit or miss proposition. On another 5517B, no matter what was done with the rare earth magnets, REF continues to decline. :( So, don't count on this as a savior. :) Add a few Alnico magnets instead.
The following are causes that may produce a variety of symptoms:
Variations in the beam sampler behavior even among units considered to be good can result in a shift in the optical frequency of 5 MHz or more. In fact, from my admittedly limited tests of 3 supposedly good beam sampler assemblies, this may be the dominant factor affecting optical frequency in lasers with a similar number of hours on the tube. However, I do not know if it is due to variations in the LCD panels, or the other optics of the beam sampler. While the laser will still easily meet specifications (5 MHz is only about 0.01 ppm), this is an annoyance in the elegance of these systems department. :) But, without comparing the laser's output to a reference, the only symptom may be a larger than expected mode imbalance, though a visual inspection and electrical testing of the LCD panel as described below may identify marginal units.
Some LCDs also seem to cause a continuous or intermittent hunting behavior of the optical frequency with a larger deviation than is normal. All HP/Agilent lasers exhibit a slow variation in optical frequency with a period of around 2.56 seconds and a deviation of 100 kHz or so. The frequency deviation in some lasers may be up to +/-0.5 MHz or even more. Since this amount of wobble in the optical frequency is well below HP/Agilent specifications, this should probably be filed under the "well that's interesting department" rather than considered a serious issue. The only way to detect it would be to beat (heterodyne) two lasers together and look at the difference frequency. The cause is a direct result of the stabilization loop implementation using an LCD to alternately select each of the two polarized modes, rather than using the more conventional polarization beam sampler and a pair of photodiodes.
To test the LCD panel, remove the two screws on top holding the beam sampler assembly in place and the two mounting screws for that end of the control PCB to allow it to be pushed away from the laser body slightly. Then, it should be possible to pop the beam sampler off of the laser. Working over a soft pad should the LCD panel fall out, remove the small PCB with the photodiode, and use fine tweezers to pull out the elastomer ("Zebra Stripe") connector pads. Then, the LCD panel should slide out of the plastic housing. Simply rotating a polarizer behind it while looking through from the front may reveal an obvious problem like two or more sections that have different polarization orientations. The entire panel should be the same. Or, as in one I tested, half the cover glass may have broken off! However, the more common scenario is that a leak develops in the edge-seal and part of the LCD "delaminates" and becomes inoperative. This will also be visible by inspection even without the polarizer. If the appearance is normal with or without a polarizer, apply 5 V across the electrodes on either side of the LCD panel (where the elastomer pads were pressing) while looking through it and the polarizer. (The LCD can be placed on top of a polarizing filter on a white piece of paper.) When applying the voltage, the density should change dramatically and uniformally. CAUTION: If doing this for more than a few seconds, use a 50 Hz 5 V p-p (AC) signal with no DC component - prolonged DC current through the LCD can damage it. However, momentary DC won't hurt.
Since the elastomer strips may tend to stick to the LCD making it difficult to remove, even easier is to test the LCD in place by removing the sampler PCB and covering the reference PD port (side) and output window (front) with pieces of black tape. Put a polarizer over the input port (back) and look through the sampler port (top) with the input port facing a brightly illuminated white surface. Then, one orientation of the polarizer should show a uniform bright field while the orthogonal orientation should show a uniform dark field. Apply the voltage and these should flip states.
Or, just swap in an LCD panel from another laser and see if it now locks! :) The entire beam sampler assembly is identical for the 5501B and all 5517 lasers in the small case, but the plastic housing and photodiode PCB differ for those like the 5517A in the large case.
The easiest way to check for ripple is to make an adapter that has the return current going through a 100 ohm resistor with a shielded cable at least a foot long. Adding a pair of 1N4148 diodes of opposing polarity in parallel across the resistor is a good idea to protect the scope should the resistor decide to fail open. The scope probe can then be attached to the end of the shielded cable. I don't recommend connecting the scope directly near the laser because it may pick up RFI from the HeNe laser power supply or control PCB.
Any of these may be present immediately at power-on, or happen after components warm up. The simplest way to confirm is to substitute a known working HeNe laser power supply, either an identical unit from another HP laser, or something that will provide 3.5 mA at 1,500 to 1,800 V (typical of a 1 to 2 mW laser). See the information above on measuring current for hookup details.
Where an adjustable power supply brick is found to be bad, it can be replaced with a fixed current (HP/Agilent) supply as long as the tube was being run at the default current of 3.5 mA - which would be in the vast majority of cases unless the tube was high mileage and would only stay lit on higher current. I had to do this for a 5517A. The original supply would start and run long enough for the laser to lock, but would then sputter a few times and blow the internal fuse. Swapping in a supply from a 5517C repaired that.
The PWM scheme of the 5501B appears to be a bit more prone to electronic problems, usually a bad Q6 (D45H281 or D45H5/8/11, PNP power transistor) for the positive heater drive. If it becomes weak or dead (but not shorted), warmup will be much slower and the laser may never reach the temperature set-point, or may not lock. If it shorts, one or both internal fuses may blow and Q7 (D44H5/8/11, NPN power transistor) for the negative heater drive may be damaged. I've also seen a bad U20 (SG3524) PWM chip whose positive heater drive output was dead, which may be related to these same failures.
For detailed service information, see the section: Additional HP/Agilent Resources. While there is nothing on the 5517 laser specifically, the electronics of the 5518A (part of the 5528A Measurement System) and 5517A is identical except that the 5518A has an additional PCB (the internal optical receiver). And the electronics of 5517B/C/D lasers using the Type I Control PCB is close enough to that of the 5518A to be useful for troubleshooting and repair.
5518As with a serial number of below 2532A02139 have the same REF frequency specifications as the 5517A (1.5 to 2.0 MHz) and can be used exactly like a 5517A laser with the turret/shutter wheel set to OTHER and ignoring the optical receiver. 5518As with a serial number of 2532A02139 and above have a REF frequency specification of 1.7 to 2.4 MHz but should work fine as 5517As as well.
The chassis, laser tube, and Connector and Control PCBs are identical to those of the 5517A. An additional optical receiver PCB which plugs into the control PCB is added inside the front of the laser, and the front bezel and shutter assembly differ for the 5518A. See the section: Notes on the HP/Agilent 5517 Two Frequency HeNe Laser for more information.
There are two apertures at the output-end of the laser. The top one is the normal laser output, with the usual control wheel for a large opening (normal), small opening (alignment), and closed. It is also the return port for straightness measurements only. A second aperture below it is for the optical receiver. This aperture is used for all measurements except straightness. It has a control wheel for large (normal) and closed (which then has an alignment target printed on the exposed surface). A large Turret Ring behind the apertures has two positions: Straight and Other. For straightness measurements, it inserts optics in the normal laser output aperture to direct a return beam there to the optical receiver, and a microswitch is activated to change the gain of the optical receiver. (The laser output power is also reduced somewhat in this position, so the optical receiver needs to be more sensitive.) There are also "Laser ON" and "Signal" LEDs on the front bezel. Laser On is the same as the LED on the back panel. Signal is lit when there is enough of a return beam to the optical receiver to be useful.
A 5517A can be converted to a 5518A by installing the optical receiver PCB and adding a small polarizer oriented at 45 degrees inside the turret assembly to generate the beat signal to the photodiode. The result will be identical to a 5518A except that it won't be able to do straightness measurements since the additional optics (and microswitch to control the optical receiver sensitivity) are not present.
Testing the 5518A is straightforward. Once READY comes on solid (4 to 5 minutes), with the turret/shutter wheel in the STRAIGHTNESS position, a common flat mirror can be used to reflect the beam back into the output aperture. Or, in the OTHER (Normal) position, a retro-reflector like a cube corner or roof prism can be used to direct the beam into the bottom aperture. Even without monitoring the electrical signals, if the SIGNAL LED comes on, the laser is probably fully functional.
Make sure the shutter/turret wheel is in the OTHER (Normal) position when measuring output power. Otherwise, you'll get a very disappointing reading - in the STRAIGHTNESS position, it's cut by about 75 percent.
Several photos of a 5518A laser head can be found in the Laser Equipment Gallery (Version 2.42 or higher) under "Hewlett Packard/Agilent HeNe Lasers".
Some 5518As include one additional component, not present in any other HP/Agilent laser I've seen, and that is a shield or cover surrounding the area of the beam expander, purpose unknown, but possibly to prevent stray scattered light from reaching the optical receiver photodetector.
I have reverse engineered the schematic for the Optical Receiver PCB shown in Photos of HP-5518A Optical Receiver PCB. See: HP-5518A Optical Receiver Schematic. Most of the component designations are arbitrary since very few had anything on the artwork. Although it performs a function similar to that of external optical receivers like the 10780C, the circuit is considerably simpler and nearly identical to that of the reference receiver on the Control PCB. The built-in photodiode can be seen below the hole through which the output beam passes. The two pin header attaches to the microswitch in the current assembly that selects gain based on whether it is set for "Straight(ness)" or "Other". The gain is increased in Straightness mode since the outgoing beam passes through a non-polarizing beam-splitter and the return beam reflects off of it
The one trim-pot on the PCB is for sensitivity. The adjustment procedure assumes the laser is producing an output power of at least 100 micro;W when locked (READY on solid). It sets the sensitivity so that the optical receiver will work reliably over at least a 10:1 range of return optical power. A Retro-Reflector (RR, cube-corner) and an OD0.5 or OD1 Neutral Density (ND) filter are required. It is assumed that the 5508A is being used with the laser:
I would also like to find the non-HP equivalent of the receiver IC U1, HP part number 1826-0775, listed as 1DA7Q on the HP schematic of the 5517B laser, which (among others) uses the same IC. If anyone has a standard part number and/or datasheet, please contact me via the Sci.Electronics.Repair FAQ Email Links Page. Of course, maybe 1DA7Q was just a random text string intended to be replaced by the actual part number and that never happened! :) A different revision of the schematic shows the manufacturer part number as 1826-0075 which could be another typo.
The case style of the 5519A/B is similar to that of the 5517A (and 5518A) and the three mounting holes on the feet are tapped M8x1.25. (You were no doubt unable to sleep not knowing this vital information!) The tube assembly is very nearly physically interchangeable among all these large-case lasers. The "very nearly" means that a small piece of the casting of an older 5517A or 5518A tube assembly may need to be cut away to provide clearance for the internal DC power supply, not present on those lasers. (Newer 5517A and 5518A tube assemblies have already incorporated this change.) So, where the higher REF is not needed, a 5517A or 5518A tube assembly can be installed relatively easily. See the section: Notes on the HP/Agilent 5517 Two Frequency HeNe Laser for more information on the tube itself.
The Control PCB of the 5519A/B laser heads is functionally and physically similar to those in the 5517A and 5518A lasers. However, it is not known with certainty whether it is also a direct replacement. The PCB layout has changed somewhat but the only obvious electrical difference is that the REF and MEAS differential outputs are transformer coupled on the 5519, rather than being directly coupled from the 75114 drivers as they are on all other older 5517 and 5518 lasers (as well as the 5501B). (Later 5517s with the Type II Control PCB also use a transformer.) This was probably done to improve isolation and immunity to ESD damage and doesn't affect compatibility with measurement electronics. Like all the other lasers, the Control PCB requires +/-15 VDC. The internal switchmode power supply provides only +15 VDC while a miniature DC-DC converter on the Connector PCB generates -15 VDC.
The 5519A/B is particularly easy to test since it plugs into the AC outlet! Once READY comes on solid (4 to 5 minutes), with the turret/shutter wheel in the STRAIGHTNESS position, a common flat mirror can be used to reflect the beam back into the output aperture. Or, in the OTHER (Normal) position, a retro-reflector like a cube corner or roof prism can be used to direct the beam into the bottom aperture. Even without monitoring the electrical signals, if the SIGNAL LED comes on, the laser is probably fully functional.
As with the 5518A, make sure the shutter/turret wheel is in the OTHER (Normal) position when measuring output power. Otherwise, you'll get a very disappointing reading - in the STRAIGHTNESS position, it's cut by about 75 percent.
Many photos of a 5519A laser head can be found in the Laser Equipment Gallery (Version 2.31 or higher) under "Hewlett Packard/Agilent HeNe Lasers".
The N1211A implementation uses a tube assembly similar to that of a 5517 designed to have a high output power (probably close to 1 mW), with a low split frequency (1.6 MHz typical). Since the AOM RF drive frequencies can be selected to generate an almost arbitrary difference frequency, the split frequency from the laesr tube can be almost any value and a lower split frequency results in higher output power (down to a point). Based on photos I've seen and a sample tube I have, the N1211A tube assembly is physically similar to that of the small 5517s (e.g., 5517B) but has a telescope that produces a 1 mm beam optimal for the AOMs (rather than the 3 to 9 mm beam common to all other HP/Agilent lasers), and the entire assembly is slightly shorter. For reasons only known to the designers, the hole spacing of the mounting feet is just incompatible enough to be annoying. A diagram is shown in Internal Structure of Agilent N1211 Laser Tube Assembly and a photo in Tube Assembly used in Agilent N1211A Laser. Upon casual examination, the difference might go unnoticed. But it is a bit shorter since the output goes through a simple collimator rather than a beam expaander. And the feet are machined rather than cast. Otherwise, it's really a 5517 tube that runs on a 5517 HeNe laser power supply and will lock using a 5517 controller. One sample I have has an output power over 600 uW and it may have been removed from service due to low power! See Agilent N1211A Laser Tube Assembly Installed in 5517B Body.
The output from the tube assembly goes to a Polarizing Beam-Splitter (PBS) which separates the F1 and F2 components. These each pass through an AOM with its own RF drive frequency. If, for example, these are 80 MHz and 94 MHz for F1 and F2 (where F2-F1=1.6 MHz), then the resulting output frequencies will be offset by 80 MHz and 95.6 MHz (F2 + 96.6 MHz) - (F1 + 80 MHz) resulting in a difference frequency of 16.6 MHz. These two new components (call them F1' and F2') are coupled into individual polarization-maintaining optical fibers which terminate at a Remote Optical Combiner (ROC). Another PBS then merges the two components into a free-space beam with a diameter of 6 mm or 9 mm for the N1212A or N1212B ROCs, respectively.
Physically, a system using an N1211A is much larger than any of the 5517 lasers. While the laser tube and controller are similar to those in the small 5517 lasers, there is significant added complexity in the optics and AOMs, and their drive.
The 10780A is used in interferometry systems using the 5501A, 5501B, 5517A, or 5517B laser heads. It contains a silicon photodiode behind a focusing lens and polarizing filter oriented at 45 degrees, a preamp, a comparator to generate a digital signal from the heterodyne beat of the two polarized modes of the Zeeman-split lasers, and a differential line driver. The primary output is called called "MEAS" and its complement "~MEAS". There is also a Beam Indicator LED which will be lit when there is enough power to produce a reliable beat frequency signal. (This threshold is adjustable.) The 10780B appears substantially similar to the 10780A except that the threshold pot is accessible without removing the receiver cover.
The pinout of the main connector (J1) is:
BNC Pin PCB Pin Function
-----------------------------------------------------------------
1 (LL,F) 1 ~MEAS (Zeeman beat signal pair from
2 (UL,F) 2 MEAS differential line driver.)
3 (LR,M) 3 Return (also BNC shell and receiver case.)
4 (UR,M) 4 +15 VDC
_____
| |
| |
| TP |
| |
| |
MEAS | x o | +15 VDC
~MEAS | x o | Return
|_____|
The PCB pins are counted from the edge of the board. I don't know the official designations of the pins on the funny bi-sex 4 pin BNC connector. LL (Lower Left, etc., F for female and M for male) reference the connector with the receiver oriented vertially - with the optical input and Beam Indicator LED at the top. (Rather than buying the way overpriced mating cable, I fashioned a 2 pin female header for power that fits only one way into the male pins, and a separate 2 pin male header for the MEAS signal. These were then glued into a BNC shell. It's not as pretty as the original but it works. The official mating connector is a Souriau 21P106-1, which Arrow Electronics has on backorder and the cost is not listed so might be more than an arm and two legs for each one. :( :) Other distributors also carry Souriau products but I haven't yet found an actual way to buy one of these connectors.
The small 4 pin (male) LEMO on the 10880A/B/C (Optical Receiver Cable), 10881A/B/C (Laser Head Cable), and possibly others has the following pinout:
Red Dot
|
v
+15 VDC o o ~MEAS/~REF
Return o o MEAS/REF
+15 VDC is required to power the optical receiver when a 10880 Optical Receiver Cable is used with a measurement or servo axis card for MEAS. But testing a 10881A Laser Head Cable shows continuity to +15 VDC in the laser head and this doesn't make sense for REF - both the laser head and card would be voltage sources.
Note that the 10791, one of the types of cables that is used to connect the 5517 laser heads to DC power and the measurement electronics, has a 4 pin BNC plug like the one that mates with the optical receivers. The REF outputs of the laser are on the male pins with +15 VDC and GND on the female pins. This connector should normally NOT be attached to the optical receiver!
There is also an external test-point called "Beam Monitor" on a feed-through pin sticking out above J1. This is the intermediate rectified and filtered signal used for the threshold detection and is useful for peaking the alignment. However, it is NOT a linear function of signal strength being highly compressed at the upper end.
The case should not be connected to the optical metal chassis or Earth ground (I assume for single point grounding noise considerations). Use Nylon screws through the plastic insulated mounting holes at each end. If any plastic pieces are missing (as is often the case with used receivers) add insulating washers if necessary.
The 10780A and 10780B are now considered obsolete as they are not guaranteed to work with 5517C/D and later interferometer lasers over the full specified velocity range since the spec'd upper cutoff frequency is too low (5 MHz). However, HP/Agilent specs are often very conservative. A 10780A I tested using a function generator and LED operated from below 40 kHz to over 8 MHz. It actually would probably be usable down to around 10 kHz but the waveform was somewhat distorted below 40 kHz. The sensitivity as determined by the voltage on the Beam Monitor test-point was down to about 50 percent of what it was at 5 MHz, but some of that fall-off might have been due to my LED/driver, and it is non-linear. The replacements are the 10780C (free space optical input) and 10780F (fiber optic input, though some of these may actually have the 10780C model number and/or be designated 10780U). The 10780C and 10780F have a guaranteed frequency range from 100 kHz to 7.2 MHz. But for experimental use, when using a single interferometer, or when not requiring high velocity in one direction, the 10780A or 10780B should be fine and typically much less expensive on eBay. :-)
Any of these HP receivers make good general detectors for optical heterodyne beat signals within their frequency bandwidth since they will operate over a wide range of input optical power from a few µW to 1 mW or more without adjustment. They will also operate with similar optical pulsed signals and work fine to detect the chopped drive of some of my LED flashlights! :) However, note that although the 10780F/U can be used with free-space input, to do so will require a polarizer at 45 degrees to be added. And since there is no lens to focus the light onto the small area photodiode, the maximum sensitivity is much lower than for the other optical receivers.
Here is a high resolution scan of a 10887A card:
I've figured out the base address set by the DIP-switch (SW1) and IRQ settings on the 2x8 pin jumper block (J14) to the right of SW1:
J14 Position
IRQ ISA Bus Pin (Left to Right)
------------------------------------
2/9 B4 1
3 B25 2
4 B24 3
5 B23 4
6 B22 -
7 B21 5
The jumper for IRQ5 was found to be installed. The functions of J14 positions 1 to 3 are not known. They do not connect to any edge pins but position 1 also has a jumper installed.
After about a year of waiting, I finally acquired Windows software for the 10887A. Although it's supposed to be for a dual-axis system, once I figured out the settings (above), it stopped complaining about "10887A Not Found" and seems to be happy enough even though there is no second axis. I didn't have a 5519A/B laser available, but I did have a 10883B cable which adapts the 10887A to a 5518A laser. But I also didn't have a real 5518A laser, so I put one together using a 5517A and the optical receiver PCB from a defunct 5518A. The control PCB in 5517As is identical to the one in 5518As with an extra row of pins to connect to the optical receiver PCB, unused in the 5517A. After drilling a second hole in the output aperture disk, the laser is indistinguishable from a genuine 5518A except that it cannot be used for straightness measurements. (The turret doesn't have the periscope optics to direct a return beam into the laser aperture down to the photodiode.) However, until I realized that there was no polarizer in front of the photodiode and added one, although the software was happy with the signal level and reset properly, the position refused to change. The signal detect LED did behave rather strangely, tending to be on when the alignment was sub-optimal and going out when perfect. Apparently, with that marginal alignment there's enough REF in the return beam for a signal to be detected, but no actual MEAS from the interferometer, so the phase of REF to REF was constant and the position remained stuck at 0.00000. Once that little detail was resolved, the display began to behave normally. And this software seems to be rather capable and cool. ;-)
There are several status LEDs on the 10887A PCB. While I don't really know what they indicate, my observations are as follows:
I'm still looking for more information on the 10887A, 5528A or 5529A manuals, etc. If you have anything like this you're will to share, please contact me via the Sci.Electronics.Repair FAQ Email Links Page.
5517/5508A adapter pin-out
Mil DB25
Pin Pin Function
-----------------------------------------------------------------------------
A 1 MTR (MEAS signal level to meter on 5508A)
B 2 ~MEAS
C 3 MEAS
D 15 Signal Return (MEAS)
E 5 ~REF
F 6 REF
G,H 7,10 Ground
J 11 +15 VDC Sense
K 12 +15 VDC
L 8 -15 VDC
9 -15 VDC Sense
20 -15 VDC
21 -15 VDC
M 23 +15 VDC
N,P 13,16 Cable Shield
R 18 Signal Return (REF)
S 19 Ground
17 Ground
22 Ground
T 24 +15 VDC
U 25 Cable Shield
4 NC
14 NC
DB25 male:
MTR ~MEAS MEAS NC ~REF REF GND -15 -15S GND +15S +15 CSHLD
1 2 3 4 5 6 7 8 9 10 11 12 13
14 15 16 17 18 19 20 21 22 23 24 25
NC MGND CSHLD GND RGND GND -15 -15 GND +15 +15 CSHLD
The Power, REF, and MEAS signals are also brought out to terminal blocks so they can be monitored or easily attached to test equipment like a frequency counter or oscilloscope, as well as the 5501A reference connector.
For testing 5517s (all versions) and 5518As, the 5508A is used directly with the DB25 adapter. For testing 5501As, a separate DC power supply is used but with the 5508A powered and fed with REF and MEAS via the terminal blocks.
5517/5508A connectors
Since the circular MIL-Spec connector for the 5517/5508A is non-standard, the connectors available from electronics distributors need to be modified. The Amphenol part number is PT06A-14-18P-SR and includes a strain relief. Various electronics distributors and even Amazon.com (!!) have them for under $30 in single quanity, which is actually quite good as these things go. (But the prices on Amazon vary quite wildly, so some searching is worthwhile.) The modification turned out to be easier than I had anticipated. The pin block is made of rubber and can be pushed out with a piece of 1/2" copper pipe in a drill press. (Though a very slightly larger cylinder would be a bit better.) First, go around the periphery from both ends with a thin blade which will free most of the rubber from the adhesive used to secure it in place. The pipe fits around the pins without mashing them and only contacts the rubber. Push in increments, making sure the rubber doesn't get too misshapened or skewed in the process. The screw-on strain relief (if present) or some other suitable spacer with a hole in it will be needed under the connector to allow the rubber block to be pressed clear of the shell. Then reinstall in a similar way after aligning with respect to the 5517 or 5508A connector. There will be some damage to the rubber, but it should not affect anything unless you're a purist. Even without any adhesive, the fit is really snug enough, but won't be a Mil-Spec connector that's waterproof. :) It would also be straightforward to fabricate a "punch" that matches the pin pattern. That may reduce collateral damage, but doesn't seem to be worth the effort unless 1,000 of these connectors need to be modified.
And a note about trying to salvage HP cables if all the required connections aren't already present: Forget it. The cover on the laser-end connector consists of a thick rubber boot on top of a hard plastic conformal molded inner core. While the boot can be slit from end-to-end and peeled off, I doubt it is realistic to remove the core without damage to the connector and pins. I gave up after seeing what would be involved since I didn't have any TNT handy. :) So, for example, an ET-319283 adapter cable which has the 5517 connector at one end and a 7 pin LEMO at the other, possibly intended to connect a 5519A/B to a 5508A Measurement Display isn't useful to power a laser since the DC power connections are not present. (The 5519A/B has a built-in switchmode power supply that runs off the AC line.)
5501A/B connectors
The 5501A and 5501B use a pair of 4 pin circular connectors. The power connector is standard with a suitable mate being Amphenol PT06-8-4P-SR. (This is also available from Mouser, but is more expensive than the 18 pin connector!) The reference connector has the keying rotated 45 degrees but a similar push out and reinsert approach works, though more care is needed to assure that the rubber doesn't get destroyed. The diagnostic connector (present only on the 5501A) mates with the standard PT06-14-18P-SR. Unless you're into automated monitoring, building a cable for that is probably not worth it. See the sections on the 5501A/B, above, for pinouts.
Scans of original product brochures for the Model 200, 220, and 260 lasers, and html versions, as well as general desciptions and a price list can be found at Vintage Lasers and Accessories Brochures and Manuals under "Laboratory for Science". The brochures include a nice description of the principles of operation and applications considerations in addition to the specifications.
The following brief descriptions include extensive contributions from David Woolsey (http://www.davidwoolsey.com/).)
There were three Laboratory for Science stabilized HeNe lasers known to have been produced and sold:
All three models had the same size power supply/control box but the laser head for the Model 260 was longer than those for the models 200 and 220. The user controls and general operating procedures are also basically the same for all models.
A number of features and attention to detail set these lasers apart from most other commercial stabilized HeNe lasers that are or have been available. These are described with respect to each model in the following sections. Unfortunately, clever ideas and implementation are often not the most important factors in determining the success of a product or business.
Even with the superb technology, not many of any of these lasers were ever sold. The total production run for all the years of the product line from the early 1980s to sometime in 1995 was soemthing like: 300 for the Model 200, 60 for the Model 220, and only 10 for the Model 260. There are references to other models ranging up to 280 in the product literature, but someone who actually worked at Laser for Science throughout the years of ultra stable laser production never heard of them going beyond the discussion stage.
Ironically, the extensive discussion of retro-reflections in the product brochures may have scared off potential buyers. Nearly half the text in the brochures for the LFS-200, LFS-220, and LFS-260 is related to the effects and mitigation of retro-reflections which some people might interpret as a deficiency with these lasers. Retro-reflections are a problem with all lasers, but especially with lasers designed to have the best stability performance. Other manufacturers tend to simply mention retro-reflections in the operation manual - not the product brochures! - as something to be avoided, but even there, they don't dwell on it.
Even experienced laser jocks find it hard to understand how reflected light with a power level 1/100,000,000 or less compared to the intra-cavity power can have an effect on the behavior but it definitely can with these type of lasers.
If anyone has schematics, a service manual, or other detailed documentation for any of the Laboratory for Science lasers (or an actual Laboratory for Science laser!) stached away they no longer need, please contact me via the Sci.Electronics.Repair FAQ Email Links Page.
The LFS-200 and LFS-220 are described in the order in which I acquired samples.
Although LFS is now out of business, other companies do offer transverse Zeeman stabilized HeNe lasers. One example is NEOARK (Japan).
Among the features and attention to detail that sets the Model 220 laser apart are:
The well known commercially available HeNe lasers I'm aware of implement very few, if any of these. And note that many duel frequency Zeeman like the 5501A/B, 5517, and others, use simple dual polarization mode stabilization techniques despite their being Zeeman lasers with fancy price tags. :)
Scans of an original product brochure for the LFS-220 can be found at Vintage Lasers and Accessories Brochures and Manuals under "Laboratory for Science". A much more compact html version is at Model 220 Ultra Stable Laser Brochure. The brochure includes a nice description of the principles of operation and of course, the specifications.
The first Model 220 I (Sam) acquired on eBay - S/N 51 - has IC date codes and PCB fab dates between 1981 and 1986. But if the serial numbers started at 1 (or even 10 as has been suggested) rather than 50 and only 60 lasers were ever built, it may be much newer than 1986, possibly between 1988 and 1992. So what if the chips are a bit moldy, they haven't changed in any way other than dropping in price by 1 or 2 orders of magnitude since 1981. :) Maybe LFS bought their chips from PolyPacks (a popular surplus outfit for cheap chips that also no longer exists). ;-)
While the tube in this laser is weak - around 0.8 mW on a good day which is about half the minimum power spec - this is more than adequate to provide a stable beat frequency signal. Originally, the laser was going through what appeared to be normal warmup, but would not lock after the warmup period and the Lock Level indicator came on. The Model 220 has a headphone jack to permit listening to the PLL error signal (as do the other models as well) and a knob to adjust the PLL gain. And while the knob affected the sound in the headphones, there was little correlation with anything else. It was like a bad SciFi movie sound track! I was thinking there must be electronic problems preventing a stable lock from being achieved. Fortunately, all ICs are standard 4000-series CMOS and common analog parts. Unfortunately, it's not likely that a schematic will be available given how few of these were probably built. Google gas been totally incapable of finding much of any useful information beyond the brochure for the Model 200 on David Woolsey's Web site and a few journel references citing the use of Laboratory for Science lasers for such-an-such research.
However, a miracle happened. Someone sent me the user manual for the Model 220 and lo and behold, that empty socket under the controller I had been pondering since acquiring the laser needed a jumper plug to complete the internal signal paths! It provides access to all the critical input and outputs of the internal architecture of the Model 220 controller with the intent to permit the use of an external frequency reference, remote control and monitoring, and other advanced functions. The jumper block must have either fallen out in shipping, or the previous owner had been using the remote hookup and kept the cable. No wonder it didn't work. Nothing was connected together! So there were electronic problems of sorts. :)
With the default jumper plug constructed and installed, *everything* started working in a manner that actually made sense. The Mode LED went on and off as the modes cycled in the HeNe laser tube during warmup and the headphones produced a satisfying chirp a couple times during each mode cycle. When the 30 minutes or so warmup time was completed, the laser locked instantly!
The sound from the headphones is nearly pure white noise and the beat frequency appears rock stable on an oscilloscope and around the 425.8317 kHz it should be based on the PLL synthesizer BCD switch settings of 511. (The frequency is: 3.4 MHz*m/M where M is the switch setting and m is 32, 64, or 128, preselected based on the laser tube to provide the maximum number of possible discrete Zeeman frequencies.) I intend to check it on a frequency counter but have little doubt that will also show the correct frequency with crystal accuracy. Unfortunately, I don't have a spectrum analyzer or an iodine stabilized laser to check it more precisely. The stability should increase is allowed to warm up for longer - 90 minutes is the time to reach spec'd performance. Originally, I thought it might not be working quite correctly due to the sound from the headphone jack having rumbling and other non-white noise components, but I now believe that may have been due to acoustic feedback since I was actually listening using a stereo amp.
Here are some photos:
The HeNe laser tube construction is nothing special, at least on the outside. Like the two mode stabilized HeNe lasers, a Spectra-Physics 088-2 or similar tube would work. But the actual tube used by LFS was apparently custom built though. Some, if not all, were filled with isotopically pure Ne20 or 22Ne to provide the narrowest linewidth and/or to select the precise line center, and possibly 3He as well. Later ones were made with a special bore support spider that eliminated the "slip-stick" behavior during warmup of some other designs.
The waste beam from the HR-end of the tube is used for the reference beat tone. It has a polarizing filter between the tube and the photodiode and a glued-on wedge to make sure the waste beam can't reflect back into the bore. There is an AGC circuit of sorts for the photodiode so that a usable signal can be obtained as the tube ages regardless of a (reasonable) decline in tube power.
The tube is rather elaborately suspended as can be seen in the photos. The suspension provides some degree of vibration isolation and there is even a fine thread screw (visible on the top of the laser head) to rotate the tube by a few degrees. The complex suspension was designed to minimize stress in the glass envelope and eliminate stick-slip noise due to length changes of the overall tube. It also allowed the tube to be rotated by a 100 pitch screw adjustment without twisting the tube at all. This was desirable to align the tube's birefringence axis (mode orientation) precisely with the magnetic field.
The entire laser head is thermally regulated by a temperature controller which is the circuitry on the lone PCB inside the head. The temperature set-point can be adjusted via a pot accessible from underneath the laser head. Power resistors attached to the baseplate on which the tube and magnet assembly is mounted provide the heating and an LED on the rear panel of the laser head shows the amount of power to the heater by its intensity. The baseplate bolts to the outer aluminum case with close-fitting end-plates. Although perhaps not obvious from the photos, the wall thickness is much greater than that of most other HeNe lasers.
There is also a rather elaborate transducer attached to the tube. While serving a similar function to the heaters on many mode stabilized lasers, the design was optimized for fast response. Power to this heater is what is controlled by the PLL responsible for locking the laser.
The transducer consists of a dense "zig-zag" run of copper wire about 3.5 inches long Epoxied directly to the outside of the glass tube envelope. The wire is oriented (back and forth) along the long axis of the tube, *not* as a helix or coil (it is not an inductor). When a current is run through the winding the wire heats up and immediately pulls (stretches) the glass with it. The response bandwidth is something like 10 kHz since the length change between the mirrors did not have to wait for the glass to heat up. With the wire arranged along the tube axis all of its change in length was in the intended direction - unlike with a the the more common coil arrangement.
With a simple coil, the initial change in dimension when current is applied is an increase in winding *diameter* which pulls the glass with it (expands the tube diameter) and causes an initial *shortening* of the tube. The shortening is followed by a lengthening as the heat from the transducer diffuses into the glass. This is not a good way to make a fast feedback loop. Also unlike other heater schemes generally used, with the wire directly attached to the tube glass, there is nothing in between to limit the response as with taped on thin-film heaters.
On the anode-end of the HeNe laser tube (the front of the laser head and output) is a collar with two LEDs on it and a trim-pot. Only the anode wire connects to this collar. One LED is lit when the tube is first turned on. Inside the collar is a temperature regulator for the output mirror. There is a small amount of internal reflection in the mirror that gets back into the laser cavity and this is the way it was tamed. There is a thermistor regulated heater in there that uses the laser discharge current for power. The voltage drop across the heater box will vary, but the current through it is held constant. So, the mirror temperature is regulated so that the etalon formed by its front and rear surfaces has a peak covering the neon gain curve resulting in a constant transmission without retro-reflections. For the approiximately 5 mm thick mirror - 7.5 mm optical length - the FSR is 40 GHz, compared to 1.5 GHz for the Doppler-broadened neon gain curve. So, the peak is rather broad in comparison, but keeping it centered helps long term stability.
The rear mirror had a simple prism made of cover glass that was Epoxied onto it so that the internal reflection was removed by putting it off axis. The Epoxy was made to be thicker at one side than on the other by supporting one side of the cover glass with little tabs of tape. This method couldn't be used on the output mirror.
When the output window is under proper thermal regulation both of the LEDs on the thermal regulator enclosure should be half lit. The upper one lit means heating and lower one lit means cooling. The pot adjusts the temperature set-point.
And note that neither anode or cathode is at ground potential! Don't ask how I (Sam) found this out. :( :) This was apparently for noise suppression. Grounding one end of the tube will risk inserting some 60 Hz hum onto the tube current through ground loops and such. Talk about paying attention to every last detail!
The HeNe laser tube is driven by a linear power supply with totally exposed components once the controller cover is removed. Not even a plastic shield! It is the typical voltage doubler with parasitic voltage multipler for starting. Four power transistors provide current regulation in the cathode return. While at first glance it looks similar to many other linear power supplies of the early 1980s, it was designed to put out 5 mA at 1,200 Volts with a supply ripple of about 1 mV! That gives it a SNR of around 127 dB. This was necessary in order to reduce the very small fluctuations in laser power output due to supply ripple, and their corresponding phase noise, to a minimum. This was somewhat tricky to do back then. Specifically, the current regulation control circuit has better components and additional filtering compared to common commercial HeNe laser power supplies. The PCB traces were also apparently arranged to minimize pickup of hum and noise from the nearby power transformers. A partial schematic I traced of the Model 200 HeNe laser power supply can be found in the section: Laboratory for Science Model 220 Laser Power Supply (LS-220). I still need to determine the details of the current regulation circuit (lower right in the schematic) but it's diffiult to make out because the PCB can't easily be removed from the controller case.
And speaking of details. There are some zener diodes in the power supply. If they are clear glass, room light getting in via the ventilation slots will end up modulating the power supply current, so they should be painted or replaced! Mine has the silver painted variety so I guess it's OK.
The controller has two PLLs. One is used as a frequency synthesizer to produce a highly stable reference derived from a 3.4 MHz crystal. The reference frequency may be set via 3 rotary BCD switches accessible through holes in the case. The other PLL then locks the Zeeman beat to the reference once the laser has reached operating temperature (about 1/2 hour). Thus, the reference determines the exact place on the neon gain curve where the laser will operate. (A little typewritten note on the unit I have states that the center of the 20Ne lasing line corresponds to a setting of 511.) So, maybe my laser tube is filled with isotopically pure gases.
There are 3 indicators on the front panel. The "Lock Signal" lamp on the right shows by its intensity, the approximate power to the heater transducer attached to the tube. The indicator on the left is called "Reference" and is on all the time at relatively low intensity. It is a power indicator but at a reference brightness that should be similar to the "Lock Signal" indicator when the laser is optimally stabilized. The LED at the top is called "Mode" and goes on and off during warmup as the modes cycle. When locked, it will be on at partial brightness.
A switch on the rear panel can be used to override the PLL output and select heater at max or off, to adjust the lock temperature, either because the tube is at too high or too low a temperature for stable locking, or should it lock onto a "bad" point of the Zeeman frequency response function.
The headphone jack is used not only to check on the laser during warmup and to confirm that stabilization has occurred, but also is a sensitive detector of back-reflections, which may be a destabilizing influence. Effects of optics resulting in back-reflections will be heard as transient tones in the headphones. (The headphone output may also be connected to the "Line", "CD", or "Tape" input of an audio amplifier.) Waving anything in front of the laser is audibly detectable, as are any sort of vibrations including gently touching the laser or even the table it's on, or walking across the floor. If the output is piped through loud speakers, having the volume above a very low level will result in acoustic coupling into the laser tube and a very noticeable increase in audio level as well as a change more toward non-white noise.
There is also a calibration jack which provides a beat frequency signal and DC power source for the Model 225 Zeeman Beat Frequency Range Register, whatever that is. :)
For an overview of the operating principles, which seem to track the actual implementation quite closely, see the following patents. (For the model 220, the main patent of interest will be #4,468,773.)
And a non-LFS patent for a green (543.5 nm) transverse Zeeman laser (though I don't know if it actually uses that term):
The patents also include a number of relevant references.
About two months after snagging the first LFS-220, I obtained another one, also on eBay - S/N 36. Its tube is a bit hard starting but has slightly higher power than the first - about 1.1 mW. After replacing 2 transistors and a diode which may have been bad or may have been killed when I accidentally shorted the high voltage to the Mode light bulb socket (don't ask!), it also works quite well. Internal construction appears virtually identical to S/N 51.
At some point in the future, I plan to combine the beams of the two LFS-220s and record and plot the frequency of the beat signal to determine the actual stability. I'll have to complain to the LFS QC department if they don't meet published specifications!
However, I do wonder about the frequency stability that can actually be achieved under real-World conditions. For one thing, beat frequency is a sensitive function of the magnetic field. Although somewhat shielded, anything magnetic in the vicinity is likely to have some effect on the magnetic field of the tube and thus lock point.
I have also built an experimental setup using a normal barcode scanner tube in a transverse magnetic field. While turning this into a stabilized transverse Zeeman laser is unlikely to occur, I have captured some plots of it's behavior. See the section: Two Frequency HeNe Lasers Based on Zeeman Splitting.
I have acquired a scan of the operation manual for the Model 220 laser but have not gotten permission to make it public as yet. However, much of the same technical information with respect to theory of operation can be found in the brochures at Vintage Lasers and Accessories Brochures and Manuals and in the patents. In fact, the block diagram in the operation manual is taken directly from Fig. 1 of Patent #4,468,773.
Among the features and attention to detail that sets the Model 200 laser apart are:
The well known commercially available HeNe lasers I'm aware of implement very few, if any of these except for tube testing, which would be essential.
At the same time, the electronic implementation (see the schematics) is a bit too simple and could benefit from a few things like an integrator in the feedback loop and bypass capacitors!
Scans of an original product brochure for the LFS-200 can be found at Vintage Lasers and Accessories Brochures and Manuals under "Laboratory for Science".
Both the laser head and controller for the LFS-200 are superficially identical to those of the LFS-220 except for the lack of a tube rotation knob on the laser head. Operation is generally similar as well, including the use of the audio headphones for locating back-reflections. However, the tube lacks the heated OC mirror and of course, the additional rotation hardware. The shutter lever on the laser head selects among NP (Non Polarized), off, and CP or LP (Circularly Polarized or Linear Polarized, apparently depending on serial number). The manual says the latter is CP (Circular Polarized).
The interior of the laser head also differs in a number of ways. The HeNe laser tube appears to be a bit shorter than the one in the LFS-220 and the anode is at the HR-end. The mode pickoff optics and photodetectors are in a little box behind the HR mirror with their premap mounted on the side. There is an offset trim-pot for the mode position accessible from under the laser head. The heaters and temperature controller are mounted on the baseplate as with the LFS-220.
The controller box is arranged roughly the same way as for the LFS-220 but the locking circuitry is substantially simpler having a total of three 8 pin DIPs: LF412 and LM358 op-amps, and an LM2905 timer, presumably for the warmup delay. But there are 6 pots for adjustment (in addition to the user accessible "volume control" servo gain knob). The HeNe laser power supply is similar to the one in the LFS-220 but several additional high voltage filter capacitors have been added on the Control PCB to zap the unsuspecting. Some versions also have one or two additional pots, as well as an unidentified object in the vicinity of its control circuit, purpose unknown.
Here are some photos:
There were several versions of this power supply, at least two without the additional trim-pot, and another with the 4 extra capacitors mounted on the same PCB. The latest I've come across (from laser serial number 98) has two additional trim-pots, circuitry in a shielded enclosure at the bottom right of the PSU PCB, and an active filter to go with the auxiliary cap bank, which is part of the Control PCB. This is shown in LFS Model 200: HeNe Laser Power Supply 2. and LFS Model 200: Auxiliary Capacitor Bank with Active Ripple Filter.
And those large electrolytic capacitors are part of the HeNe laser power supply waiting to zap the unsuspecting! :) Later versions include an active filter in addition to the filter caps. Someone must have noticed 0.1 pA of 120 Hz ripple in the tube current and couldn't live with that. :-)
There was also a version of the LFS-200 made specifically for Teletrac which included an optical receiver module bolted onto the front of a relatively standard LFS-200, but the laser was painted black. :)
The designers at Laboratory for Science appear to have been more obsessed with retro-reflection or back-reflection (same thing) than at any other stabilized laser company. This is understandable considering the higher level of performance that is being achieved with the higher bandwidth servo system more sensitive to cavity perturbation. For example, while other stabilized HeNe lasers will simply use a polarizing beam splitter or two to separate the modes making sure to angle all reflective surfaces to prevent back-reflection, the LFS-200 has added the QWPs after the polarizers. The optics stack sandwich for each mode visible in the photo of the HR-end of the LFS-200, above, is something like:
Plexiglas back-plate | Amber filter | Polarizer | QWP | Plexiglas front-plate -> PD
Two passes through the QWP (out and back) result in a 90 degree rotation of the polarization axis so any reflected light is blocked by the polarizer.
There is also a significant amount of electronics in the laser head including the laser head temperature controller and photodiode amplifiers. Reverse engineering those would require ripping apart a laser head - something I'm not planning on doing any time soon.
Note that there were many engineering changes over the course of manufacturing relatively few lasers, with little if any documentation or revision numbering on the PCBs. So, don't be alarmed if there are discrepancies between the schematics and the PCBs in your laser!
Scans of an original product brochure for the LFS-210 can be found at Vintage Lasers and Accessories Brochures and Manuals under "Laboratory for Science".
The controller for the LFS-210 is identical to that of the LFS-200 except for a slot on the left side providing access to the low/high (red/blue) side lock switch. (The switch is present on the LFS-200 controller but not user-accessible.)
The mode position trim-pot found under the LFS-200 laser head has been replaced by a 10-turn pot mounted on the backplate. The frequency/intensity stabilization select switch is behind behind a plastic plug on the right size of the laser head. Up is F and down is I. Hidden inside are a pair of trim-pots to adjust the offset and something else for the lock position. The photodiode closer to the front of the laser is for the vertically polarized longitudinal mode for intensity feedback while the one closer to the back is for horizontally polarized longitudinal mode also used for frequency stabilization.
Everything else about the LFS-210 is identical to the LFS-200. Refer to the previous sections for details and schematics of the circuitry that's in common. (I don't have schematics for the photodiode preamp board, which differs slightly.)
The tube in the 260 was 15 inches long. It lased on three modes, giving it a more complex inter-combinational beat frequency pattern. About 50% of the power was in the central mode and a polarizer could be used to discard the other two modes since they were polarized orthogonally to it. This would get rid of the beats.
Some of the tubes were filled with single isotopic neon. Most were not. The isotopic mix did not depend on the model type though.
The tubes used in some of the later lasers were custom made by Shasta Glass (R.I.P.). These tubes had a specially designed capillary support "spider" that produced no "stick-slip" noise as the tube changed length under regulation. Other than that, there was nothing any different between the tubes used in the Laboratory for Science lasers and the tubes used in supermarket barcode scanners. We did exploit mirror defects that were typical of the type of laser tube though. Some types of sputtering artifacts can make a laser less prone to mode hopping. Also, since the mirrors were imperfect, there was a small amount of birefringence in them that we exploited as well. They were cheap tubes, but with lots of sorting and characterizing. We used about 2/3 of the tubes we bought.
The transducer was one of the fundamental, and patented, ideas that made the Laboratory for Science lasers better than any others. All the lasers used the same transducer system. One of the other patents was related to the phase locked loop electronics on the Model 220. (See the patent list above.)
A Model 220 was used by IBM in the first Atomic Force Microscope (AFM). The Model 220 could be used to measure distance changes on the order of 1/20 of an Angstrom right out of the box. Compare that to what the "competing" HP laser could do ("Position/distance resolution down to better than 10 nm") and then compare the price tags.
NASA bought a 260 for the robot that they made to test the tiles on the Space Shuttle. The robot had a YAG laser to hit the tile with a high power pulse that, due to the resulting thermal shock, would make the tile ring. The 260 was used to detect the ring modes. All this was done without contact or close proximity to the surface.
If you need a tube replacement, the right thing to do is contact Dr. Seaton. He may be able to supply you with one (even though the Lab is nominally out of business). It'll cost a bit over $1,000 installed, I would guess. There are quite a number of subtle things about tube replacement and it is best left up to someone who has done it before (unless you consider your time to be of very little value).
Why aren't there other lasers like these available today?
There are much simpler solutions available now for lasers with a coherence length of a few hundred meters. Distributed FeedBack (DFB) diode Lasers can have coherence lengths of a couple hundred meters, power outputs of many times what the Model 200 put out, cost much less than the Model 200, turn on and stabilize quicker, and don't die as easily when abused. (However, DFB lasers do not provide a self-referenced absolute frequency, as do stabilized HeNe lasers. --- Sam.)
As for the Model 220, I am not quite sure why nobody is making an equivalent system now. I suppose that there is just no significant demand for 1 mW of optical power with 20 km of coherence length. Also, there is only so much that modern manufacturing will get you in this case because there is just too much "hand tweaking" that went into these lasers.
LFS could have charged 2 or 3 times as much as they did and not lost sales. There was no place else to turn, short of much more complex and expensive iodine stabilized lasers and such, for the 220 and 260 levels of performance. The Lab almost got involved in making an iodine stabilized system. I think I recall Dr. Seaton claiming that it would have something like 0.01 Hz stability.
The RB-1 consisted of two pieces. The first RB-1 I saw had laser head S/N 1 and controller S/N 2, so at least two of these systems were built and I had mismatched pieces. However, I have photos (below) of RB-1 S/N 8 with very similar construction, which still looks like someone's science fair project. :) The thing clearly wouldn't be caught dead going out to a paying customer, though it's almost certain that the RB-1 or its successor eventually morphed into the Newport NL-1 (maybe "Newport Laseangle 1"?) as a result of a merger or buy-out. However, I've yet to see an NL-1 (or production RB-1 if there ever was such a thing) in person.
The RB-1 laser head contains the HeNe laser tube, with wrap-around heater, a beam sampler assembly that diverted all of one polarization to a photodiode and part of the orthogonal polarization to another photodiode, and preamps for the photodiodes. The base is a 3/4 inch thick aluminum slab with a 1/8 inch aluminum cover sealed with foam rubber.
The HeNe laser tube was from Uniphase, a garden variety model with a length of about 8 inches, which is somewhat unusual, probably rated around 2 mW. A tube length of 6 or 9.5 inches being more common, at least today. However, there is a modern Uniphase that's similar, the 1018 at 8.5 inches, rated 2.5 mW.
The beam sampler includes a polarizing beamsplitter cube to extract one of the mode signals and prevent it from reaching the output at all, and a separate angled plate to extract a portion of the orthogonal mode. A pair of EG&G SGD-100A photodiodes (may be similar to the Perkin Elmer FFD-100) fed LF356 op-amps. (EG&G is now part of EXCELITAS.)
The controller houses a linear DC power supply, standard Laser Drive HeNe laser power supply brick, feedback circuitry, and heater driver. There were controls on the front clearly not for an end-user, like 8 or 10 gain settings and a fine gain control for one of the op-amps, selection of which mode signal to pass to an output, a current meter for the heater, and so forth. People who typically use these things would have no clue of what to do with the knobs and switches. I've yet to see a user manual for the RB-1.
While the mounting of the HeNe laser tube is somewhat overkill and the beam sampler is a nice solid unit with an adequate number of adjustments, the electronic construction of both the laser head and controller are, to put it politely, a disaster. Everything is on those copper strip prototyping boards, with capacitor upon capacitor added in various places no doubt to tame noise pickup or instability. (Someone must have had stock in a capacitor company!) The designers must have had a goal of using strange and hard to find connectors wherever possible which they did for the separate cables of the photodiode signals (blue multipin) and heater drive (microphone two pin). Power for the HeNe laser tube in S/Ns 1 and 2 came from a standard Alden on the controller but at the laser head had both the medium voltage BNC on top for the positive and the normal BNC on the bottom for the negative. In S/N 8, the high voltage cable is hard-wired into the laser head. Maybe the engineers were getting zapped too often. :)
Here is a composite photo of S/Ns 1 and 2:
Here are some photos of Laseangle RB-1 S/N 8 courtesy of eBay seller rdr-electronics:
The 7900 is a dual mode polarization stabilized laser essentially similar to the Coherent 200, Spectra-Physics 117/A, and others. It consists of a rectangular laser head which contains the controller and HeNe laser power supply, and a separate box with DC power supplies and possibly a status indicator.
The specifications and a photo for the 7900 can be found at Mark-Tech Model 7900 Frequency Stabilized Laser. And the 7910 at Mark-Tech Model 7910 Frequency Stabilized Laser
(It appearas as though the actual Mark-Tech Web site is history. It's not clear now long Google will maintain theses, but I have backup copies of the HTML and photos.)
The HeNe laser tube appears to be a Uniphase 098-2 or similar, 2 to 3 mW. It uses a Laser Drive power supply.
The one interesting difference between the 7900 and most other similar lasers is that the heater to control the length of the HeNe laser tube is painted or coated on the outside of the tube, rather than being a thin film heater or wound with wire. This should potentially have a more predictable response and thus lower frequency/phase noise once locked.
During initial warmup, the controller runs the heater at rather high power until a reference temperature is reached, and then closes the feedback loop. Since it doesn't need to wait for the temperature to reach equilibrium, this greatly reduces the lock time to under 5 minutes. This is similar to that of the HP/Agilent lasers which use custom and expensive HeNe laser tubes which have an internal heater wrapped around the bore. Most other stabilized HeNe lasers using off-the-shelf tubes take 10 to 20 minutes to lock. The tube does run rather hot though, but this is probably normal.
The stabilization feedback is implemented in 2 op-amps with some other stuff to monitor the heater temperature, do the switchover from preheat to feedback mode, and generate status signals.
The output power when locked on the sample I have is about 0.9 mW. (The spec'd minimum locked power is 0.5 mW.) So, this one appears to be basically in like-new condition even though it has a manufacturing date of 1984 making it 24 years old, with a serial number of 112. And I bet they started at S/N 100! :)
In addition to connections for +/-15 VDC and ground, there are a pair of status signals from the laser head. One goes high (around +12 VDC) a few seconds after the laser locks. The other is open collector, and turns on at the same time. However, this signal will start pulsing if the lock is interrupted - for example, if the beam to the photodetectors is momentarily blocked. The pulsing continues even after lock is re-established. There is a third signal, also open collector, that is always on. I have no idea what that does.
Here is the pinout for the circular connector (J5, mating connector is AMP/Tyco part number 206434-1 with possible pin part number 66507-9). The same pin numbering is also used on the internal PCB header:
Pin Function Commecnts ----------------------------------------------------------=------------------- 1 Ground 2 Ground 3 +15 VDC Direct to HeNe laser PSU, +12 V reg elsewhere. 4 -15 VDC 5 Ground 6 Lock/Error (Blink) OC, on when locked, 2 Hz for loss of lock. 7 Lock Low initially, +12 VDC when locked. 8 Unknown OC, on all the time.
(There are two similar circular connectors on this laser but only one of them is wired to anything internally.)
The same case seems to be used for a fancier Mark-Tech laser as there are obvious tapped hole locations for mounting additional optics and other stuff. This may be for the model 7910 which appears to be an interferometer laser used in measurement/calibration systems. However, unlike those for similar applications from HP/Agilent and Zygo, it is probably NOT a two-frequency Zeeman laser but simply a model 7900 with an internal optical receiver using simple quadrature A/B fringe counting in the interferometer. See Model 7910 Interferometer System.
Here are two photos of the interior:
Here are the optical and stabilization specifications for the 05-STP-901 (from Melles Griot):
Optical Specifications ------------------------------------------------- Output Wavelength: 633 nm Minimum Output Power: 1 mW M2: <1.1 Beam Diameter (1/e2): 0.5 mm Far-Field Divergence (1/e2): 1.60 mrad Polarization: Vertical, >1000:1 Spatial Mode: TEM00 Longitudinal Mode: Single Stabilization Characteristics - Frequency Stabilized Mode ------------------------------------------------------------- Frequency Stability (1 min/1 hr/8 hr): +/-0.5/2.0/3.0 MHz Power Stability (1 min/1 hr/8 hr): 1.0% rms Frequency Offset: +/-150 MHz Temperature Dependence: 0.5 MHz/°C Stabilization Characteristics - Intensity Stabilized Mode -------------------------------------------------------------- Frequency Stability (1 min/1 hr/8 hr): +/-3.0/5.0/5.0 MHz Power Stability (1 min/1 hr/8 hr): +/-0.1/0.2/0.2% rms Frequency Offset: +/-50 MHz Stabilization Characteristics - General --------------------------------------------------------------- Noise: 0.05% rms Lock Temperature Range: 10 °C to 30 °C Time to Lock: <30 minutes
There are also an 05-STP-903 and 05-STP-905. These differ only in the default line voltage settings: 230 VAC for the 05-STP-903 (and CE compliant) and 100 VAC for the 05-STP-905. The laser heads are identical.
The specifications for the SP-117A should be similar. Note that the 05-STP-901 and SP-117 do NOT have a precisely specified vacuum wavelength. Thus the specifications simply have "633 nm". :) The same is generally true of most non-metrology stabilized HeNe lasers. If what is desired is a precisely specified constant vacuum wavelength, then a metrology laser such as those from Hewlett Packard (now Agilent) might be a better choice. For one such laser, the vacuum wavelength is spec'd to be 632.991372 nm and typically changes by less than 0.02 picometers over the life of the laser! In terms of optical frequency, that's around 10 MHz of drift. However, note that I do not even believe the HP/Agilent numbers for their vacuum wavelength as they have changed the specifications for no apparent reason! And lasers with the two different specs have identical wavelengths when tested. The only way to know for sure outside of NIST would be to compare them to an iodine-stabilized laser or other laser with a super precisely known wavelength, or a wavemeter calibrated against one.
The first unit I acquired was of relatively recent manufacture (as these things go) - 1996. The only major problem I found with it was a dead HeNe laser power supply brick - a Laser Drive unit rated 4.5 mA at 1,600 V, similar to the one in the SP-117. It appears to be a standard model except for a hand-printed label with "0.03 percent noise". So, it's either built with better filtering or is specially selected for this application from standard units. Using an external HeNe laser power supply temporarily allowed the controller to be tested. However, it appears to be much more finicky than the original SP-117 in Frequency Mode and would only stabilize with one of my SP-117-compatible laser heads. It basically ignored one that had a slightly leaky photodiode and my home-built clone, simply turning on the "Locked" LED but not actually doing anything. All three of these laser heads stabilize reliably on the much older SP-117 controller. I suspect that an adjustment of the gain of the photodiode preamps would take care of this - probably just turning it up all the way. So, perhaps I shouldn't be so hard on it. :)
Switching to Intensity Mode at first resulted in the heater simply turning on. The offset pot had to be adjusted to get the mode signals to be in the required range for the locking circuitry to operate, but then it would lock with perhaps 30 seconds required to settle down.
Switching from Frequency to Intensity Mode or back caused the Locked LED to flash and the Locked relay to chatter for a few seconds. This seems to be normal behavior, as it happens on every one of these lasers I've seen. (The Locked relay provides a set of SPDT contacts that can be used to control auxiliary equipment, though there is no external connector for it. But a cable could be wired to the PCB pads and snaked out through the ventilation slots on the bottom of the case.)
Monitoring the heater drive signal on an oscilloscope shows how sensitive this feedback scheme really is. Even playing music at a moderate level evoked a detectable response. Tapping on the concrete floor resulted in an oscillation that took a few seconds to die out.
The electronic design of the SP-117A and 05-STP-901 clearly has it roots in that of the SP-117 (no A) but in addition to the added circuitry for the intensity mode, the frequency control loop has been upgraded to a pure integrator (with a few additional op-amp stages). The PCB layout is completely new and TL084s have replaced LF347s. But much of it looks like it is unchanged from the SP-117 design. The heater drive is a crude pulse width modulator rather than linear pass transistor. Considering the care with which the PCB is laid out with separate analog and digital grounds and linear everything else, it seems strange to have this source of high level digital noise. And, in fact, the varying heater current results in either thermal or magneticly induced vibration of the tube. This can be detected in the spectrum of the laser output if one looks hard enough. For example, if it is heterodyned with another clean laser. (An upgrade is described later to eliminate this.)The input is +12 VDC from a linear power supply. A source of +9 VDC is provided by a LM317 linear regulator. A 555 timer generates the PWM clock and also the -9 VDC power via a charge pump. These power most of the analog circuitry. Timing delays are implemented using several CMOS monostables. Grrrr. :)
Also see the sections starting with: SP-117 and SP-117A Stabilized Single Frequency HeNe Lasers for more details of the electronics including complete schematics for both the SP-117 and SP-117A/05-STP-901.
For best performance, the controller should be adjusted to match a the specific laser head. (Intensity mode may not work at all if swapping heads without readjustment.) There are only three pots inside, so this isn't that complex a procedure! Remove the cover by taking out the 4 Philips-head screws on the bottom near the feet at the edge of the case.
First, power up the laser and check the 12 VDC power supply at the mainboard PCB connector. There are two pots on the power supply PCB. The one closer to the HeNe laser power supply-side is the voltage adjust. (The other one is unlikely to need adjustment.) You'll need a tiny right angle flat blade screwdriver or bent flat hairpin turn the pot. Set the voltage for 12 V +0.25/-0.0 V. (The supply generally within tolerance unless someone before you touched the pot.)
Adjustment procedure for MG-05-STP-901 and SP-117A:
Pots R9 and R10 (500K ohms) set photodiode preamp gain while R13 (50K ohms) sets output level in INTENSITY mode. All measurements should be made with respect to AGnd (TP7). An oscilloscope is desirable for the INTENSITY mode adjustments but not essential.
It's possible that an earlier PCB revision of the MG-05-STP-901 or SP-117A may have different parts designations. In particular, the SP-117 - no "A" - has other part numbers but it should be obvious which pots and test points to use.
Confirmation of correct PD assignment:
To achieve rated specifications in Intensity Stabilized mode, it is essential that the single polarization mode used for locking correspond to the one that makes it through the polarizing beam-splitter at the output of the laser. This is unlikely to be incorrect in a laser that has never been serviced since coming from the factory, but is easy to get wrong if the laser head has been partially disassembled. So, if the trim strips on the laser head covering the cylinder holding set-screws are intact, it's probably safe to assume this is correct. The quick test is to confirm that the I Mode Level pot (R13) increases output power in Intensity Mode when turned in a clockwise direction. If the output power decreases, either the PDs are swapped or the output polarizer is rotated 90 degrees. Neither of these can occur unless (1) the factory assembly was incorrect or (2) someone went inside the laser head!
Note that the end result will be similar if the intensity signal is sensed by the photodiode at the rear of the laser head with the polarizer in the correct orientation, or by the photodiode on the side with the polarizer rotated 90 degrees. The first one may be preferred because the two polarizers will be operating in identical ways, though the difference may not be noticeable or even detectable.
To fully test, it may be worth doing the following:
The voltage on TP6 should drop to near 0 V with the paper in place. If it does not, power down, cut the cable ties securing the PD cable, unplug the PD cable and plug it back in rotated 180 degrees. Install new cable ties. Confirm that the swap was successful. (If the voltage are now negative, the connector was rotated by 90 or 270 degrees.)
If they are out of phase, it means the polarizing beam-splitter at the front of the laser is oriented incorrectly, probably being rotated by 90 degrees.
FREQUENCY mode:
The LOCKED LED should not be on at this point (even if it was before the adjustments were made). If it refuses to go off, turn the laser off for a couple minutes and power up.
The system is basically working at this point but for the final adjustments, let it remain this way for at least another hour to give everything time to warmup fully.
Use the PBS rhombus or any polarizer to compare the power in the two modes. (There should be a mark on the laser head case to indicate the plane of polarization.) If it isn't adequately equal, adjust R9 or R10 to make it so. Only a series of small adjustments should be needed, giving the system a few seconds to reach the stable position.
INTENSITY mode:
This adjustment can be done without electronic test equipment but it's much quicker with at least a DMM and easier with an oscilloscope:
Here is the procedure without electronic test equipment, but a laser power meter (or calibrated eyeball) and polarizer will be required:
Further adjustment is left as an exercise for the student and your mileage may vary. :)
Adjustments procedure for SP-117:
The SP-117 only has FREQUENCY mode, which is functionally identical to FREQUENCY mode of the SP-117A. There are also 3 pots. R11 and R13 (the two 500K pots) are equivalent to R9 and R10 of the SP-117A with R12 (the 50K pot) adjusting the position on the gain curve. I do not know why a similar function to this last pot isn't included in the SP-117A. Or, perhaps it is and my schematics have errors. Errors? Nah. :)
The LOCKED LED should not be on at this point (even if it was before the adjustments were made).
The system is basically working at this point but for the final adjustments, let it remain this way for at least another hour or so to give everything time to warmup fully.
Adjusting the output polarizer (SP-117/A and 05-STP-901):
Where a true single frequency (single longitudinal mode) is required, the polarizing beam-splitter cube or rhombus at the output of the laser must be oriented precisely. This should only be an issue if either the polarizing beam-splitter has been removed or the laser head has been rebuilt. Of course, someone before you may have fiddled with it without knowing what they were doing, so it may be worth checking! Also, of course, where both modes are desired, the polarizer is removed entirely.
The best way is to use a Scanning Fabry-Perot Interferometer (SFPI) to display the longitudinal modes and then rotate the polarizer until the undesired one disappears entirely even with the gain turned all the way up so it's buried beneath the noise floor. A high speed (>1 GHz) photodiode and RF spectrum analyzer can also be used, adjusting the orientation of the polarizer to eliminate the ~640 MHz beat between the desired and undesired modes.
Another much lower tech way to get close is to display the laser output on a graphing laser power meter or data acquisition system while the laser is warming up and mode cycling. Then, adjust the orientation of the polarizer to maximize the p-p amplitude of the power variation. This will probably reduce the unwanted mode to less than 1 percent of the desired one.
The 05-STP-901, SP-117, and SP-117A are all identical with regard to sources of periodic modulation of the optical frequency.
An external ripple reducer circuit can be added to the output of the HeNe laser power supply. See the section: Reducing the Ripple and Noise in a HeNe Laser Power Supply.
SP-117A Stabilized HeNe Laser Linear Heater Drive Modification shows a simple circuit that should eliminate this source of FM. It consists of an RC filter (including the 15K ohm resistor that that feeds Q1) to convert the PWM to a linear drive signal. The same fancy transistor can be used by removing it from the controller PCB and remounting it on an isolated heat-sink. (Isolation is required because the tab is the emitter, not the collector, and the PCB shorts it to the emitter pin. A heat-sink can be added to the PCB if the trace is cut.) The gain pot is set so that the sensitivity is about the same as with the PWM signal. A good place to start is to set it so that the voltage across the heater is about 10 V during warmup. The 1 uF capacitor shown should be acceptable, but a somewhat larger one may be used to more completely suppress the residual 5 kHz ripple while still maintaining adequate loop bandwidth.
The Newport NL-1 is the "polished" version of the Laseangle RB-1 described above, so Newport didn't actually design it either. The primary changes compared to the RB-1 are cosmetic with much nicer construction of both the laser head and controller. However, the underlying design appears to be virtually identical. It is a dual polarization mode thermally stabilized HeNe laser with very rudimentary electronics (as if the circuitry can be more basic than those in most other similar lasers!). However, while it doesn't even have automatic switch-over from preheat to stabilized operation, there are controls to set "Servo Gain", which could be used to optimize the locking behavior. It also may be locked to either the low ("red") or high ("blue") side of the neon gain curve, along with a trim-pot accessible from the front panel to set the locking point on the neon gain curve. There are front panel outputs to monitor the two polarized mode signals from the photodiodes and something called "Null Test" (which seems to be the amplified difference signal from the photodiodes), as well as a tuning input. And there is a meter to monitor the heater current so it's easy to tell if it is within the stable operating region as thermal equilibrium is reached or the ambient temperature changes.
The laser head uses the same size Uniphase HeNe laser tube as the RB-1 (possibly a 1018), along with the same rather overly elaborate beam sampler assembly. The photodiode preamp with two ICs is now on a real PCB as opposed to a prototyping board, and the power supply filters caps are also on a PCB attached to the rear panel. There are no controls or indicators on the laser head (the interlock plug has been moved to the controller), only the cables (still 3 of them - HeNe laser tube power, heater, and feedback) permanently attached via strain reliefs. The main difference compared to the RB-1 is that the laser head cover is now in three pieces - two thin (but nicely made) side panels and a thick milled top plate! But now 12 to 20 screws need to be removed to get the entire cover off, although only 8 screws (woopie!) for a side panel. :) Of course, there's generally no need to go inside the laser head other than to get zapped on the laser tube anode voltage. :) The one adjustment in there (mode balance) shouldn't really change.
The controller is in a 3.5 inch high 10 inch deep cabinet that could probably be rack-mounted with a suitable adapter kit. It has to win the award for the largest controller for a commercial dual polarized mode stabilized HeNe laser ever made and is about 90 percent empty space! All of the control electronics are on a PCB mounted behind the front panel, with an attached heat-sink for the heater driver transistors. A simple DC power supply is mounted on one side, the Laser Drive HeNe laser power supply brick is mounted on the other, and the power entrance assembly and 2 pin Jones socket interlock connector are on the back. The entire central area is totally empty! :) The laser head cables attach to the front panel probably only because that's where they were on the RB-1. (Although this would make sense if it were rack-mounted.) As with the RB-1, there is both an AC power switch AND keylock switch. Geez, it's not like this is a 10 kW cutting laser!
Operation is very straightforward based on empirical evidence, as I have no user manual: Power up with the Heater Current set at an intermediate value like 0.3 A and Servo Gain controls (Coarse and Fine) set to minimum. Allow it to warmup for 10 or 15 minutes or until a full mode sweep cycle takes more than 20 or 30 seconds. (It would have been nice if they had provided mode indicators - a pair of LEDs monitoring the Passed and Blocked outputs. But a laser power would need to be used for this.) Then, turn up the Servo Gain controls to achieve a stable lock. The Fine control appears to be most important. It seems that almost any setting that isn't minimum will result in a stable lock, but if set very low, there will be an offset. So, the feedback loop probably has no "I" term. Coarse (which is a multiposition switch) only affects loop stability in some obscure way - the meter needle tends to oscillate if set too low. Or something. So, either the labeling isn't quite accurate or Coarse is broken. :) But Fine definitely does have a much more dramatic effect than its name implies! No doubt, the Servo controls could be set more optimally by monitoring the short and long term variation in the output power or optical frequency. But locking is very rapid at almost any settings, and the output then becomes rock stable, with only a very slight drift as the entire system reaches thermal equilibrium. Flipping the Red/Blue Lock switch results in a shift in lock position over a few seconds to the opposite side of the gain curve with nearly identical power. If the tube temperature is too low or too high as evidenced by the heater current being near or at zero or maximum (about 0.6 A), it's a simple matter to move it to a different lock point by turning the Servo Fine control to minimum and allowing the tube to cool or heat by a few mode sweep cycles. So, unlike most boring stabilized HeNe lasers, this one does have a few fun controls to fiddle with. :)
The locked output power on my sample can be set between about 0.6 mW and 0.725 mW using the front panel Tuning trim-pot, though it hits 1 mW during mode sweep. The output power is almost the same for the "Red" and "Blue" lock points. (These probably only differ due to non-ideal components.) There is also a trim-pot in the laser head to adjust the mode balance between the "Passed" and "Blocked" photodiode signals. I do not know what the specs are, but the performance of this laser could be close to the new values. An 8 inch laser tube may not have a rated output power much higher than 1 mW, and the beam sampler inside the laser head diverts 100 percent of the horizontally polarized (blocked) mode and about 10 percent of the vertically polarized (passed) mode to the feedback photodiodes.
Here are photos of the laser head and controller:
Many more photos of an NL-1 laser head and controller can be found in the Laser Equipment Gallery, (Version 3.13 or higher) under "Newport HeNe Lasers".
Nikon apparently developed their own custom HeNe lasers at some point in the past, though it's not known what the intended applications were. More recently, Nikon has manufactured equipment like wafer steppers but they typically use Hewlett Packard/Agilent lasers for positioning, though it's been rumored that early ones may have used Nikon lasers. And other Nikon products such as confocal microscopes and other optical instruments that contain HeNe lasers have used standard models from other suppliers, possibly re-badged Nikon but not made by them or significantly modified for the specific application.
The tube appears to be of generally similar design to the one in the SP-119. However, it is not physically identical, so swapping a Nikon tube into an SP-119 head or vice versa, would not be possible. But it's probably functionally equivalent, though probably with a longer active region and definitely longer cavity. Why else make such a fancy piece of glasswork that's mostly hidden with the cover in place? In fact, it's not easy to even see if it has Brewster windows or where exactly the mirrors are located without extensive disassembly, so little about the details are certain!
Based on the physical characteristics of the NKL-85 tube resembling the SP-119 tube, the presence of a PZT power supply, and the behavior of the output power versus cavity tuning that indicates that there is a Lamb Dip (more below), it's almost certain that the NKL-85 uses Lamb Dip stabilization like the SP-119, the only other commercial stabilized HeNe laser known to do so. What's not likely would be for it to use the equally uncommon gain peak technique, since two peaks are present with a Lamb Dip. But it could use an overly elaborate implementation of conventional really boring single mode stabilization. I only have a laser head at the present time so much remains unknown. Or, it may simply have manual tuning via adjustment of the PZT voltage with no automatic frequency control of any kind! Given what's contained in the laser head, any of these techniques would be possible, but it would be silly to go to the effort and expense to manufacture a custom laser tube with a Lamb Dip and allow it to go to waste.
The HeNe laser power supply and PZT power supply are both low voltage DC-input high frequency inverters using ICs and transistors - no vacuum tubes, can you believe it?! :) But they are still about 10 times the size of modern equivalents.
The circular military-style connector is about 1.5 inches in diameter with 1-1/2 coarse threads (not fine threaded and not bayonet). Here is its pinout determined visually with the aid of an ohmmeter and some labels on the HeNe laser and PZT power supply PCBs. The signal names are mostly mine:
Pin Color Function Description/Comments
------------------------------------------------------------------------------
1 Black Signal Shield Shield of gray coax from Aux Box.
2 NC
3 Red +30 VDC HeNe laser PS +DC power.
Orange HeNe Interlock COM terminal of inner microswitch.
4 Yellow HeNe Interlock X NO terminal of inner microswitch.
5 White Signal Center of gray coax from Aux Box.
6 NC
7 Red PZT INT PZT power supply INT input.
Brown PZT INT NO terminal of outer microswitch.
8 White 0 V/+30 VDC RET Twisted with pin 3/red to HeNe laser PS.
9 NC
10 NC
11 White PD Out Center of green coax from PD Preamp PCB.
12 Black PD Out Shield Shield of green coax from PD Preamp PCB.
13 Black PZT- Shield Shield of black coax from PZT-.
14 Clear PZT- Center of black coax from PZT-.
15 NC
16 NC
17 NC
18 Blue PZT Interlock X COM terminal of outer microswitch.
19 Red +15 VDC PD Preamp PCB +DC power.
Red +15 VDC Aux Box +DC power.
20 White -15 VDC PD Preamp PCB -DC power.
White -15 VDC Aux Box -DC power.
21 Red +5 VDC PZT PS logic (oscillator) power.
22 White HeNe ISense Out Center of Red coax from HeNe laser PS.
(Calibration is: 1 V/mA.)
23 White 0 V/+5 VDC RET Twisted with pin 7/red to PZT PS.
24 Black 0 V PD Preamp PCB power and signal COM.
Black 0 V Aux Box power and signal COM.
25 NC
26 Green Safety Ground Wired to baseplate.
27 Black HeNe ISense RET Shield of red coax from HeNe laser PS.
Standard AMP 0.062 inch diameter pins fit reasonably well though I don't know if they are optimal.
The input to what I'm calling the "Aux Box" mounted inside above the main connector, probably an amplifier, is from a mini-coax connector mounted outside above the main connector. (It may actually be a really tiny coaxial power socket connector. I'm not sure what the mate should be and one I thought might fit was a tad too large.) The only connections between the Aux Box and anything else inside the laser head are for DC power. Its output simply goes back to the circular connector. The function of the Aux Box was originally somewhat of a mystery, though from subsequent testing, it would appear to simply be a preamp to be used with an external photodiode for monitoring the laser output during set up and testing. More below.
The wiring to the controller box is probably set up so the HeNe laser power supply interlock microswitch is in series with its DC power (labeled +30V on the PCB). The microswitch for the PZT power supply would be in series with the INT input (which is actually the power to the chopper in the PZT power supply). However, for testing, these can be bypassed externally so the laser head would work with the cover off (as if it's so difficult to jam something in the switches!).
The HeNe laser power supply doesn't appear to have an internal current regulator. Thus its DC input (despite being labeled +30 V) is actually used to adjust current, with the ISense signal (1 V/mA) providing feedback to the controller. The approximate calibration is:
Input Tube Current
-------------------------
29 VDC 4.5 mA
30 VDC 5.0 mA
32 VDC 6.5 mA
So, since 5 mA occurs at 30 VDC, it may be the nominal HeNe laser tube operating current.
The only other wiring inside the laser head are:
Since there is a 7404 IC (TTL logic) in the PZT power supply, the +5 VDC (which is connected to its Vcc/pin 14) must be present at all times the PZT power supply is running and it must be constant. Thus INT is used to control the PZT output voltage and is the actual power input to the chopper, while the +5 VDC input is used only for the oscillator. The calibration of the output with respect to INT is approximately 100 V/V. So, the range of 0 to 5 V results in an output of about 0 to 500 V, which tunes the cavity over at least two FSRs. I don't know if it is capable of a higher output voltage and have not tested that! Two FSRs is more than sufficient for locking, though a larger range would make the laser more tolerant of thermal changes in cavity length. I had originally thought that there was a resistive sense network for PZT voltage feedback. But resistance readings didn't make any sense so I scraped off its RTV Silicone coating. The "resistor" turned out to simply be a two position ceramic terminal strip like those in old Tektronix test equipment. One position connects the PZT power supply high voltage output and PZT+, the red wire to the PZT. The other position connects the center conductor of the PZT- coax and PZT-, the black wire to the PZT. This means that there is no direct sensing of the DC voltage on the PZT. It's not clear exactly how the PZT feedback was set up in the controller. INT could have been driven to vary the PZT voltage based on Lamb Dip/Mode feedback without regard to the actual voltage, only the relationship of 100 V/V. The actual PZT voltage only really matters when it approaches the upper or lower limit, but the voltage on INT could serve the same purpose. Or it could have used INT to specify a fixed DC voltage on the PZT and had a separate HV power supply in the controller connected to PZT- to control the difference of PZT+ and PZT-. The first is simpler but the clear high voltage insulation on both the PZT HV output AND the PZT- coax center conductor suggest that the latter might be a possibility. Alternatively, the required dither signal to the PZT may be applied via the coax since that would have a higher frequency response than controling the PZT power supply. Then the low frequency offset based on the locking error signal would be via the PZT power supply INT input. The only way to know for sure will be to find a controller for this laser! :)
The laser head I have is in near-mint condition except that someone seems to have removed the metal (probably adjustable) feet - presumably the only things they found useful or figured had any value! Unfortunately, these are also where the cover fastening screws attach. It's a miracle this laser survived shipping. Without the feet, the cover sits slightly lower with the hard metal underside of the top coming within less than 0.5 mm of the fragile glass laser tube and possibly even touching it. The cover was held in place with clear packing tape! However, in all fairness, it was very nicely double boxed. :)
The tube appears to be like new, with a large pristine silvery/dark getter spot with no hint of discoloration even around the edges and I've seen an output power of over 0.84 mW. The tube was labeled 0.8 mW, so that is another indication that it is very good condition. Hopefully it won't require cleaning and/or alignment. Cleaning would be a pain if it's a Brewster tube especially since access appears to be very limited. And I have no idea what, if any, adjustments there are for alignment. There is nothing obvious. The output is vertically polarized, but I do not even know if it's a 1 or 2 Brewster window tube, or an internal mirror tube with an internal Brewster plate to force the vertical polarization and an internal PZT on which the HR is mounted. Or the HR could be mounted on a bellows with an external PZT to move it. Aside from the unknowns, everything else is obvious. :)
One peculiarity was that the first time I powered the laser tube using an external HeNe laser power supply, its output power started at over 0.64 mW and declined to around 0.5 mW after awhile, though it's not clear why. The power came back once the laser was allowed to cool down, not that it gets even detectably warm on any accessible surface! I doubt the decline to be due to be anything wrong with the tube itself like contamination as it starts and runs very well with a perfect discharge color, and the output power always peaks at around 6.5 mA. I've been running it at 5.0 mA to be safe since the actual current rating is not known and the output power is only slightly higher at 6.5 mA. I thought that perhaps the decline could even be due to normal mode sweep since the resonator is a massive casting and might not be going through even one complete cycle over a short warmup. But next time it was powered on, the output power started at about 0.75 mW, and later was above 0.8 mW. (And the range with mode sweep or cavity tuning is only from about 0.71 to 0.84 mW, never as low as 0.5 mW.) I still don't know if even though the getter looks perfect, the tube is still somehow cleaning itself up with multiple power cycles, or there is an alignment issue, possibly with the mirror on the PZT since the highest readings so far have been after exercising the PZT.
After wiring up DC power supplies to run the HeNe laser and PZT power supply, and building a Darlington emitter follower to buffer the output of a function generator to drive the PZT power supply INT input, I was able to watch the laser output with respect to cavity length (mode sweep or tuning) under controlled conditions. The laser output power during a complete cycle varies from around 0.71 mW to a peak of 0.84 mW, decreasing to 0.77 mW, back up to 0.84 mW, then down to 0.71. The Lamb Dip may be the valley at 0.77 mW, though that behavior could be present even without one if the laser were lasing in two longitudinal modes over a portion of the cycle. And the fact that the power doesn't decline further during part of the cycle suggested that the cavity of this laser might be longer than that of the SP-119 and is able to support two longitudinal modes when they are on either side of the neon gain curve. Typical Output Power versus Cavity Length for SP-119 Lamb Dip Stabilized HeNe Laser shows similar behavior for the Spectra-Physics 119 laser (though the cavity length and thus FSR, c/2L, differ for the NKL-85, more below). Depending on the health of the SP-119 tube, the "Mode Hop" points (and their surroundings) may actually result in an output power of exactly 0.0 mW. For the NKL-85 tube with its relatively high output power, the dips there are not nearly as dramatic.
The only way to know for sure if there is indeed a true Lamb Dip is to display the output on a Scanning Fabry-Perot Interferometer (SFPI). The SFPI will show the longitudinal mode structure including whether there are 1 or 2 modes at any given time and their relative amplitudes. To confirm that the valley (or equivalently, a double bump) is present, the laser cavity length much be swept using its PZT over a range where there is only a single longitudinal mode lasing while displaying the mode structure on the SFPI.
And indeed, using a Spectra-Physics 470-03 SFPI, (1) the NKL-85 laser is pure single mode under all conditions and (2) there does appear to be a Lamb Dip centered on the portion of the cavity tuning between the extremes where mode hops occur. So, as the cavity length changes, the single lasing mode moves across the gain curve through the area of the Lamb Dip until it becomes too weak on one side and then mode hops to the opposite side. Since there are never two modes present at the same time, the mode spacing can't be measured directly (either by measuring the distance on the SFPI display or by measuring the beat frequency between them). But mode hops during cavity tuning will be the same distance and were estimated to be 1.125 GHz corresponding to a cavity length of about 133 mm, which is significantly longer than the 10 cm cavity of the SP-119 and explains both the much higher power output and the smaller variation in output power during cavity tuning. However, it doesn't quite explain why it's always single mode as a common HeNe laser tube with a 133 mm cavity length would have two modes over a portion of its mode sweep cycle. I speculate that this may be due to an isotopically pure gas-fill or the gain being fully saturated over a portion of the cavity, both requirements for a Lamb Dip to be present. However, the SP-117 tube has a somewhat similar behavior without being a Lamb Dip laser, though it may still satisfy these requirements. A composite scope photo of the SFPI display is shown in Mode Profile of Nikon NKL-85 HeNe Laser. Compare this with a diagram for the similar (but shorter) SP-119 laser in Longitudinal Modes of Short HeNe Laser with Lamb Dip. The SFPI was scanning at about about 50 Hz while the cavity length was being swept with a triangle waveform from a function generator at a few Hz. The room was darkened and the digital camera was set up to use its normal exposure setting, which kept the "shutter" open for a good fraction of a second. Even so, it took about 40 shots to get even this not very fantastic composite photo with a decent combination of multiple mode peaks and proper focus! The ugly mode peaks are due to the requirement that the reflected beam not go back into the laser aperture since that would destabilize the laser and produce all sorts of noise artifacts. Thus, the alignment of the laser to the SFPI could not be even close to optimal.
It would be quite simple to build a Lamb Dip stabilizer. With modern components, the added circuitry would easily fit in a corner of the laser head. A single IC like the SE5521 LVDT Signal Conditioner could perform most of the required functions. (A Google search will return links to the SE5521 datasheet and app notes.) Among other things, the SE5521 includes a sinewave oscillator and synchronous demodulator. The oscillator output would provide the dither to the PZT and the synchronous demodulator would then generate the DC offset error correction signal to the PZT to lock the cavity length to the minimum of the Lamb Dip. (Other names for this are a lock-in amplifier or phase sensitive detector.) For the Lamb Dip, the objective is to find a location on the neon gain curve where the photodiode's response of the laser output with respect to the PZT dither signal has no first harmonic (inflection point), and the rectified dither signal and detected light output signals are in phase (local minima). However, building a replacement for the controller is not my intent at present. I would much rather find a genuine original one, even if it weighs another 20 pounds! More details on the origin of the Lamb Dip and Lamb Dip stabilization can be found in the sections starting with: SP-119 Laser Principles of Operation.
I haven't found any dates on the laser head or date codes on any parts inside, so its age is not known. The controller is from at least 1981, based on a date code found on a TTL IC. There is a number on the laser head which I originally assumed was a serial number - 24041, but the same number is also present on the controller. So, either I have matching units (which seems highly unlikely given that they came from entirely different sources a year or more apart) or it's really a part number or something else. However, it would be strange not to have any sort of serial number on a laser such as this. If that 24041 is actually a serial number, I'll wager a bushel of stabilized HeNe lasers that this is at most only the 41st NKL-85 to have been manufactured. :)
Based on the minimal number of user controls and indicators and the relative complexity of the circuitry, I assumed that operation would be rather simple. With no laser head connnected, the "INCREASE" and "DECREASE" buttons were able to smoothly change the PZT voltage with the switch set to "BIAS VOLT". And of course, there was nothing on the meter when the switch was set to "POWER MON".
At that point, not much more was known about the controller. None of the trim-pots are labeled and with that many, this rig would either work or it won't when plugged into the laser head. But at first I thought that would have to wait as there was one key piece still missing: The umbilical cable with the strange circular 27 pin connectors at both ends. I wasn't lookiing forward to constructing a set of jumper wires that could be snaked from the controller to the laser head, plugging into the individual pins and sockets! :) However, I reached a compromise. :) By removing the 4 mounting screws and rotating the connector on the controller counterclockwise 90 degrees, the laser head will mate with it when both controller and laser head are on solid ground at the same height. (Leaving the connector detached from the backpanel gives even more flexibility.) Physically and electrically, this works great. But the results at first were, well, slightly strange and at first nothing in the behavior of the system appeared to have changed, except that the laser tube lit up. BIAS VOLT moved smoothly up and down and POWER MON still read zero.
However, using a laser power meter monitoring the output, the laser behaved as expected with respect to the INCREASE and DECREASE buttons changing the voltage on the PZT (BIAS VOLT), with the laser output power varying slightly as the cavity length was changed. The regions where the Lamb Dip and mode hop occur are easily seen. What I finally discovered is that the meter responds to a signal applied to that mysterious mini-coax above the main connector on the laser head, but not to the internal PD behind the HR! Installing a photodiode connected to the mini-coax (cathode to the center, anode to ground) results in a meter reading which is linear with respect to laser power. With the external photodiode, it's possible to use the buttons to locate the Lamb Dip on the meter. Flipping to LOCK then allows the PZT voltage to remain unchanged, but it's not immediately clear if that is presently doing anything other than to disable the buttons. Perhaps the system must warm up for 3 hours before locking!
But that turned out not to be the case. Something was bothering me and I decided to remove the cover on the laser head and check out the wiring of the PD preamp behind the HR. The result was that it is simply an AC-coupled amplifier with a gain of 1 V/µA. In fact, the PD is simply in series with a 0.01 uF capacitor, and the combination is across the inputs of the LM308 op-amp, with a 1M in parallel with 33 pF for the feedback network. After poking around and putting everything back together, a miracle occurred. ;-) At first what was happening didn't make sense. Since the function of the PD behind the HR was now known, the signal at its output should be the AC component of the laser output power as the cavity length is dithered. On an oscilloscope, it showed up at about 2 kHz with a maximum amplitude of about 25 mV. The PZT voltage could still be adjusted and the waveform would change shape depending on where on the gain curve the cavity was centered. But the signal would disappear almost as soon as the buttons were released. Then I realized that not only was this behavior fundamentally different than before, but the PZT voltage liked to "stick" at several discrete values within its acceptable range instead of moving up or down with uniform speed and staying wherever it was when the buttons were released. Now when the SELECT/LOCK switch is in the SELECT position, the buttons still change the PZT voltage as before, but when they are released, the controller immediately searches for and locks at the Lamb Dip. At that point, the CONTROL LED will be on. However, if there are any backreflections from something external (e.g., a photodiode), the LED flickers from light tapping on the laser head or even the table. More robust tapping will make it flicker regardless of any backreflections. If the buttons are released at a particularly "bad" spot, the CONTROL LED may remain off for a few seconds until lock is re-acquired. There is also the sound of one or more relays clicking as the locking process takes place. There are actually 9 or 10 possible lock positions within the acceptable range of PZT voltage. When the SELECT/LOCK switch is then flipped to the LOCK position, all it may indeed do is prevent the buttons from having any effect, so the laser remains locked without the possibility of accidental loss due to over zealous button pressing. :) The mini-coax must indeed be there simply for testing purposes since its signal is not used at all by the locking electronics
I do not know exactly why locking now appears to be working correctly or why it was behaving differently before examining the PD preamp PCB. Possibly there is a intermittent solder joint there and wiggling the wires fixed it, at least temporarily. But a careful examination didn't reveal anything obvious. All the solder joints looked good. Or perhaps one or more of the relays' contacts were dirty and had to clean themselves and the timing was a coincidence. Or perhaps it just got tired of my use of assorted expletives. :)
In fact, it now appears as though operation is nearly fully automatic. The laser will power up into a locked state with the PZT voltage close to mid-range. But if needed after a long warmup or significant change in ambient temperature, the "INCREASE" and "DECREASE" buttons can be used to reposition it. POWER MON is then only required during testing or setup and seems to be calibrated to use a simple photodiode placed directly in the output beam.
If anyone has more information on Nikon lasers, an operation and service manaul :), or a cable or set of connectors, please contact me via the Sci.Electronics.Repair FAQ Email Links Page.
As noted, the L-109 does NOT employ a two frequency technique with orthogonal polarization and external interferometer optics as in the Zygo or HP/Agilent lasers. Rather a Bragg cell (basically an Acousto Optic Modulator or AOM) adds a fixed frequency offset to the return beam from the remote target which adds to the Doppler shift due to target motion. The two beams are then heterodyned by the built in optical receiver shielded with copper foil that feeds the processor. It must get its power via the coax since there are no other connections. Either that, or it's simply a photodiode. The signal out of the optical receiver is equivalent to MEAS in the HP/Agilent lasers. Because the Bragg cell introduces a frequency shift for both the outgoing and return beams that add, REF is double the frequency of the RF signal used to drive the Bragg cell. One interesting twist is that the output mirror of the HeNe laser tube is also one mirror of the Michelson interferometer so that wavefront alignment of the outgoing and return beams is assured as long as the two beams are coincident on the photodiode of the optical receiver. The residual beam entering the laser tube through the mirror does not destabilize the laser because it has been shifted in optical frequency by the Bragg cell. In principle, the resolution and accuracy of this approach should be similar to that of the orthogonally polarized two-frequency lasers with external interferometer optics.
The most detailed explanation of how this all works seems to be in U.S. Patent #5,116,126: Interferometer requiring no critical component alignment. What's on the Optodyne Web site itself is rather sparse. However, some information is available on the "Downloads" page.
The laser head is much smaller than it appears on the Optodyne Web site, only about 2x2x9 inches. The "R" seems to refer to the normal 9 or 10 mm beam diameter. There is also an L-109N with a 0.5 mm beam and normal 1.7 mR beam divergence with no beam expander! And, there is a 20 mm version, probably with an external beam expander (or additional beam expander).
The DC power input is 15 VDC based on measurements of a P-108L controller, though the laser seems to be happy on as low as 10 VDC. Older versions of the L-109 laser head have a funky round LEMO-style connector with 3 mini-coaxes (1 of these appears to be unused) and 2 small female pins for power. The L-109 photos on the Optodyne Web site seem to have separate connectors for the two signals (RF and MEAS) and power.
Here are some photos:
The connector pinout is as follows (view with laser head labels on top):
RF In (Coax)
O
DC- * * DC+
MEAS Out (Coax) O O NC (Coax)
The first L-109R that I acquired behaved rather strangely. I was only powering the laser and its stabilization electronics as I did not have an RF source for the Bragg cell or processor to do anything with the MEAS signal, but I don't think that is the problem. It did lock to a single mode after 10 or 15 minutes. However, after that there is a 2 or 3 minute cycle whereby lock point slowly drifts part way over the gain curve, sits there for 10 or 20 seconds, at which point the green LED on the back of the laser head gradually increases in brightness. Then it moves back the way it came much more quickly and abruptly re-locks at a different location on the gain curve at which point the LED goes out, and the cycle repeats. It's almost certain that it remains locked during the slow drift, just that the lock point is moving for some unknown reason, possibly a fault in the electronics. This can't be normal behavior, though the limited optical frequency/wavelength variation may not materially affect the measurement performance.
But a second L-109R run from a P-108L controller (described below) locked to a stable constant mode position after about 15 minutes, though it did go through several false starts similar in behavior to the first laser head. I retested the first laser on the P-108L with no obvious change, so there would seem to be a real problem.
The output of the first laser when locked (or what passes for being locked!) is about 250 µW, which is probably normal as at least one half the power from the tube is lost in the beamsplitter optics forming the internal part of the Michelson interferometer. For testing without RF drive to the Bragg cell, I adjusted the alignment of the beam expander to output the un-deflected beam. Otherwise, there is very little output power, just a couple of splotches totally perhaps 40 µW! So, depending on the efficiency of the Bragg cell, the locked output could be lower than the 250 µW I measured as it will depend on what portion of the beam is actually deflected. When I tested the first L-109R laser head, I thought it was broken. But after reading the patent, have concluded that this is normal without any RF drive. I do not believe that the altered alignment is the cause of the peculiar behavior as I tested to confirm that back-reflections weren't confusing stabilization by blocking the outgoing beam and monitoring the mode sweep from the waste beam out the other end of the laser tube. There was no change. One interesting tid-bit though: This tube is a flipper for the first 5 minutes or so as it warms up, and then reverts to normal behavior.
There's one other mystery associated with the first L-109R: A skinny blue wire (like wire-wrap wire) is connected to the electronics PCB and runs the length of the laser but is not attached to anything at the other end. All the other wires are fatter. This sort of thin wire is only used elsewhere in the laser head to as cable ties. These can be seen in Optodyne L-109R HeNe Laser Head - Closeup of Interferometer Optics. The unconnected wire runs along the top of the photo and then curves down on the left. Cable ties using similar red and blue wires are also visible in the upper center and surrounding the ballast case at the upper right. It's not a ground wire as there's 1K ohms or more between it and ground, and it doesn't appear to have a counterpart in the second L-109R laser head, though I have not removed the chassis from the case to examine its entire length. At first I thought the wire had ripped out of someplace when I pulled off the rear cover, but I don't see any evidence of that. It looks like it was cut clean. Of course if it was supposed to go somewhere, that could explain the peculiar locking behavior. :)
If anyone has more information on Optodyne lasers, please contact me via the Sci.Electronics.Repair FAQ Email Links Page.
Aside from their very low split/REF frequency of 250 kHz, the Optra two-frequency HeNe lasers are not particularly unusual except for their stabilization technique, which uses both a heater blanket over a portion of the center of the laser tube for overall (slow speed) cavity length control as well as one wound around the restricted portion of one mirror mount stem for fast cavity length control by changing the length of the mirror mount itself. The only technical references I could locate were two Optra patents:
Due to the low split frequency, long conventional tube, wimpy magnet, and PLL locking technique, I originally thought that the Optralite was a transverse Zeeman HeNe laser. Transverse Zeeman HeNe lasers normally have split frequencies in the hundreds of kHz range, not the MHz range as with most other axial Zeeman HeNe lasers. However, after finding the raw output from the HeNe laser tube to be circularly polarized and not linearly polarized, I checked the magnetic field orientation more carefully with my trusty cereal box compass and found it to be pointing along the axis of the tube, and uniform around the perimeter of the ring magnets - the definition of an axial magnetic field! That they are ring magnets and not bar magnets on either side of the tube should have been the tip-off, but I guess I wasn't paying attention! The tube in an axial Zeeman HeNe laser produces a pair of left and right circularly polarized components while the tube in a transverse Zeeman HeNe laser produces a pair of orthogonal linearly polarized components.
However, given the very limited coverage of the magnetic field of perhaps 25 percent of the gain region (the bore discharge) - and probably not being particularly uniform as well - it's not clear how the neon gain curve will be split and what will happen when more than one longitudinal mode is lasing. In addition, the near-hemispherical cavity this tube likely has will result in a tapered intra-cavity mode volume with a corresponding variation in gain. This seems to be leading up to a messy integral better left for the advanced course. :) But ignoring the mode volume issue and simplifying the situation to two parts:
These two pairs of functions could be summed resulting in a combined Zeeman-split neon gain curve. Each portion would look like a lop-sided Gaussian weighted a bit away from the center. (The gain curve diagrams shown in the patent do not represent reality.)
The Optralite uses a 7-1/8 inch (181 mm) random polarized Aerotech HeNe laser tube of conventional design, but with an additional optic glued to the HR mirror, possibly to add wedge and/or an AR-coating to minimize back-reflections and etalon effects. Whether it is specially modified in any other way is not known but all indications are that it is a standard model. And according to the second patent, above, the tube may be an Aerotech LT05R which uses natural Ne but isotopically pure 3He. The tube used in the Aerotech Syncrolase 100 is physically identical - perhaps this is more than a coincidence. A tube of this length would almost certainly result in the production of rogue modes if mated with the type of strong full length magnet found in the HP/Agilent lasers. Normally, when the laser is locked, there should be a single Zeeman-split lasing mode consisting of the two frequency components F1 and F2. Rogue modes are undesirable lasing modes that are present due to additional longitudinal modes fitting under the Zeeman-split neon gain curve. The tubes in HP/Agilent lasers have a cavity length of only around 5 inches (127 mm) to suppress rogue modes, and axial Zeeman HeNe lasers from other manufacturers use tubes with shorter cavities - some are 4 inches (102 mm) or even less. But with the Optralite, this relatively long HeNe laser tube, which has a cavity length of around 6-3/4 inches (172 mm), can be used without the risk of producing rogue modes because the two parts of the Zeeman-split neon gain curve are not spread very far apart by the small low strength magnet.
As with the HP/Agilent lasers, a Quarter WavePlate (QWP) at the output is used to convert the circular polarization to linear polarization, with the same adjustments - rotation and tilt. But the mount is a real pain to deal with as the waveplate itself is glued to the hub of a spherical bearing that can rotate around its axis in an outer shell. This entire affair is simply clamped via a set-screw so it's almost impossible to independently adjust rotation and tilt. However, there are a pair of holes on the front side of the inner hub so it may be possible to make a special tool (like a bent paper clip!) for this purpose. Perhaps that's what they used at Optra. :) There is no Half WavePlate (HWP) like the one present in *all* HP/Agilent lasers. It's possible that since the laser tube can be optimally oriented in its mount during final assembly and alignment, the HWP may be unnecessary. (This is not possible with HP/Agilent lasers since their tubes are embedded in potting compound.)
Feedback is provided by a single photodiode behind a polarizer which monitors the waste beam from the rear of the tube. A CD4046-based PLL locks the Zeeman beat to a reference oscillator using thermal control of the cavity length of the HeNe laser tube. While a PLL can be used with any laser where beat frequencies are present and correlate with lasing mode position, the only other Zeeman HeNe laser I'd seen that used a PLL was the Laboratory For Science model 220, a transverse Zeeman laser which also used a CD4046. (The LFS-260 had a PLL for stabilization as well, but it is not a Zeeman laser.)
There are two heaters for cavity length control. One covers a bit over an inch of the length of the tube near its center to maintain the overall temperature of the tube - DC and low frequency response. The other is wrapped around the OC mirror mount stem to provide high frequency (well relatively speaking!) response. The tube heater is a bit strange, at least compared to those on every other stabilized laser I've seen. It's not a low voltage thin film Kapton heater, but is covered with some sort of high temperature fabric, has a resistance of around 1K ohms, and is powered from 115 VAC via a DC-controlled solid state relay! During warmup, it is run at full voltage but once the system locks, it seems to be driven by a bang-bang-bang control loop with only three states - off, 1/2 voltage, and full voltage, swinging wildly between these even when the laser is locked and stable, and more so if the optical feedback or thermal environment is disturbed. (However, I do not really know if this is how it is supposed to work and the 1/2 voltage state may just be an illusion due to very rapid switching.) At the same time, the mirror mount stem heater has a nearly constant voltage of approximately 7.5 VDC on it when the laser is locked and stable, but which varies in a more continuous manner if the optical feedback or thermal environment is disturbed. The actual drive to the mirror mount stem heater is a filtered version of a PWFM (Pulse Width Frequency Modulation) signal - a pulse train that varies in frequency from about 50 Hz to several kHz and whose duty cycle also changes. Its derived from the PLL somehow but is clearly not directly out of the CD4046. A TIP121 Darlington power transistor configured as an emitter-follower provides up to approximately 10 V to the heater.
Interrupting the beam to the feedback photodiode will result in wild swings of both heater voltages and if done repeatedly, will eventually cause the system to lose lock and return to the warmup state for a few minutes before re-acquiring lock. Touching the tube (which affects its temperature) or blowing on it will produce a somewhat more muted response.
The Zeeman magnetic field is produced by a pair of permanent (ferrite) ring magnets each about 1/4 inch thick and not quite touching, attached to the tube and frame with globs of RTV Silicone. They are approximately centered on the discharge of the tube (the bore), although they only actually overlap perhaps 20 percent of it. The field strength is probably a few hundred gauss.
Several photos of the Optra Optralite laser can be found in the Laser Equipment Gallery (Version 3.02 or higher) under "Optra HeNe Lasers". Four interesting ones are included here:
The CD4046 is the left-most chip in the top row. Only its "Phase Dectector 2" is actually used since the VCR is in effect the laser tuning. The reference clock is an SPG8640BN, a crystal oscillator with programmable divider built in. Other ICs include op-amps (LM324, TL072), a 555 timer, logic (CD4001, CD4013), and voltage comparators (LM339). Can you locate any trmpots? I bet not. :)
The Optralite I have is in good physical condition and seems to operate normally - at times (more below). It locks in about 9 minutes from a cold start, with an output power of about 1.35 mW. Although this may be a bit low for a typical 7 inch HeNe laser tube when new, it is above the CDRH sticker rating of 1 mW!
There are a pair of BNC connectors on the rear panel (see photo, above). The one on the left is for the internally generated REF signal and the one on the right is for the output of the optical receiver which Optra calls SIGNAL (but HP/Agilent calls MEAS). Both outputs are simply amplified from their respective photodiodes with no clipping or conversion to digital signals. So, the amplitude of REF depends on the output power of the laser tube and the amplitude of SIGNAL depends on the strength and alignment of the return beam.
From a cold start at around 65 °F (18.3 °C), the laser goes through around 196 complete mode sweep cycles and then a partial one as the PLL kicks in after about 9 minutes. The first cycle takes about 1 second and the last about 8 seconds. However, each cycle represents a change in cavity length of only 1/2 wavelength rather than the full wavelength of a laser with orthogonal linearly polarized modes, so it's equivalent to 98 of those. But this still seems a bit high.
Interestingly, there is less than a 20 percent variation in the output power of F1 or F2 during mode sweep. This must mean that there are multiple longitudinal modes present throughout almost its entire range. Although the tube is long by axial Zeeman laser standards, it's short enough that under normal conditions with no magnetic field, the mode sweep would be very large, possibly even 100 percent.
At the same time, REF exhibits a wide frequency variation from below 60 kHz to over 310 kHz due to mode sweep. Over most of this range, REF is a nice sinusoid. But near the minimum frequency, the waveform becomes quite distorted, though the beat never disappears entirely as it does over much of the mode sweep in HP/Agilent lasers. The lock point is at 250 kHz - just below the peak. And the PLL maintains REF at precisely 250.000 kHz +/-0.5 Hz averaged over 1 second.
I attempted to replicate this behavior with several HeNe laser tubes of similar length in laser heads including two Melles Griot 05-LHR-911s and an Aerotech OEM1R (which may actually have a tube physically identical to the LT05R). But I only have a single ring magnet with a hole large enough for a laser head to fit and it is quite weak. So, while all tubes produced a beat signal over at least a portion of the mode sweep, none really matched the behavior of the Optralite. Depending on the specific tube and to some extent, the placement of the magnet, the maximum beat frequency ranged from 95 to 165 kHz, the waveform was sinusoidal only over a small portion of mode sweep, and the frequency variation ranged from small to large. For all, the beat totally disappeared at times. Aside from these anomalies, everything was totally consistent. :)
However, there is definitely a problem with this laser. Although it sometimes will remain perfectly stable for hours with the cover removed, it lost lock after a short while when restarted with the cover installed, seemingly unable to increase the tube temperature even when RESET. That 8 second mode sweep cycle just before lock already means that the tube is nearly as hot as it can get since the heater has been running at full power. Assuming that some sort of timing criteria is used to determine when to lock (as there doesn't appear to be any temperature sensor for the tube), perhaps the heater is not supposed to be running at full power during warmup or it should be locking earlier when the mode sweep cycle is much shorter than 8 seconds.
In addition, if disturbed in any way (or perhaps even if not), that remarkable stability is lost as though there is something trying to push it up by a few kHz. REF would go down to 250.000 kHz, then after a few seconds start creeping up until it gets forced back down to 250.000 kHz, and the cycle repeats. It might go up to 255 kHz or higher, or only to 250.1 kHz depending on how it feels. However, given that it is possible under some undefined conditions to maintain 250.000 kHz continuously apparently forever, this suggests some sort of intermittent problem in the electronics. Perhaps an op-amp is latching up. But it could even be noisy power. Or, maybe that spastic drive to the tube heater should really vary more continuously. A schematic would be *really* helpful about now. :-)
The maximum beat frequency during mode sweep would actually correspond to the condition where the optical frequency is centered on the Zeeman-spilt neon gain curve and be the most stable in terms of absolute optical freqency over the long term. So, since the laser locks at 250 kHz which is offset by about 60 kHz suggests that as the tube ages and its power declines, there will be some additional drift in optical frequency, not present with the HP/Agilent lasers that automatically keep the lock point centered.
Since the range of beat frequencies during mode sweep is rather wide, it might be possible to change the strapping of the SPG8640BN to 200 kHz, 166.67 kHz, or perhaps even 100 kHz if a different REF frequency were desired for some unfathomable reason. Hey, another programmable two-frequency laser! :) (The LFS-220 PLL has BCD switches for this purpose, but with enough resolution to actually select the optical frequency to a very high precision.)
As expected, the output from the laser consists of two linearly polarized components, F1 and F2, separated from each-other by the split frequency, and orthogonal to each-other lined up with the laser's X and Y axes. Thus, as with other similar lasers, the beat signal amplitude with an external photodiode is a maximum with a polarizer oriented at 45 degrees. The optical power of F1 and F2 is nearly equal when the laser is locked, but there is relatively little variation even during mode sweep. It is not yet known whether F1 (the lower frequency component) is horizontally or vertically oriented. There is a shutter with positions for "OPEN/NORMAL" (with polarizer at 45 degrees in front of optical receiver photodiode), "CLOSED" (output beam is blocked), and "OPEN/FDL" (with no polarizer in front of optical receiver photodiode - no idea what "FDL" means).
The green "LOCKED" LED comes on once the laser has locked (about 10 minutes from a cold start, possibly once a complete mode sweep cycle exceeds about 8 seconds). It is apparently monitoring the PWFM drive to the mirror mount stem heater as it is a pulse train with a 70 or 80 percent duty cycle and also flickers with back-reflections or when the laser is disturbed in another way such as by blowing on the tube. Pressing the "RESET" button while locked causes the laser to go through a few minutes of behavior similar to the mode sweep during warmup, at which point it then snaps back into lock. I'm not sure under what specific conditions the red "UNLOCKED" LED is supposed to come on. I forced it to light up momentarily by blowing on the tube so lock was lost and then reacquired. It is not on during warmup or RESET. (Originally, UNLOCKED never came on under any conditions, which seemed suspicious. It turned out that the LED was dead. Surely the laser hadn't been unlocked for long enough to wear out an LED!)
Only pins 1 to 9 of the DB15 "REMOTE" connector are wired to anything. Here are their functions. Values were measured with the laser locked:
Pin Name/Description Locked Condition (if applicable)
-------------------------------------------------------------------------
1 REF 250.000 kHz sinewave
2 Ground 0 V
3 PWFM, 50 Hz to several kHz ~65 Hz, 70 to 80 percent duty cycle
4 Similar to pin 3
5 RESET- To RESET button via diode (cathode)
6 Heater transistor drive +8.3 V (filtered version of pin 3)
7 Ground 0 V
8 SIGNAL 250.000 kHz + Doppler shift
9 Ground 0 V
Pins 10 through 15 are not connected to anything.
The laser may be RESET by pulling down on pin 5. While there are three signals relevant to the mirror mount stem heater (pins 3, 4, and 6), the tube heater doesn't seem to be represented at all. Surprisingly, digital status information - like whether the laser is locked - is not present on this connector, at least not in a simple 0 or 1 form. But it can be inferred from the PWFM and heater drive signals.
This laser has no electronic adjustments of any kind as there are exactly *zero* trim-pots on the control PCB. But perhaps some of the component values are hand-selected at the time of manufacture based on the measured output power of the HeNe laser tube and other characteristics once the tube assembly with the Zeeman magnet has been installed. On the other hand as long as the range of beat frequencies during the mode sweep includes 250 kHz regardless what happens as the tube ages and the feedback can handle the normal decline in waste beam output power, there really shouldn't be any need for adjustments!
I've been contacted by someone - who believe it or not - is apparently still using the Optralite lasers and contacted me about possible repair. They told me that (not surprisingly) the Optralite was probably never a real product, but what might be termed a pre-production prototype, and perhaps at most 20 were ever built, with variations in the details of design and construction (cuts and jumpers, specific signals going to the REMOTE connector, etc.). Given the construction quality - which in some ways is superb but in others resembles a science fair project - this makes perfect sense. And some of the quirks I've observed were apparently present in other lasers shipped to customers. No wonder it's been discontinued.
I emailed Optra via the contact form on their Web site asking for anything they might still have available with respect to the Optralite laser but all I received in replay was: "Sorry, everyone who worked on that product has left the company. Good luck.". I followed up suggesting that given the significance of the Optralite in Optra's history, one would think that there would be a company archive with documentation. No reply, not even a grunt. :( :)
If anyone has additional info for the Optralite or other Optra HeNe lasers including brochures, specifications, operation and service manuals, and schematics, please contact me via the Sci.Electronics.Repair FAQ Email Links Page.
Most of the circuitry for the Optralite is contained on the single large PCB on the left side of the laser. For all intents and purposes, the analog circuitry for the REF and SIGNAL photodiode amplifiers (left 1/3rd of the PCB) is entirely separate from the mixed (analog/digital) circuitry for the PLL and thermal control, with series inductors on +/-15 VDC to suppress noise.
Partial specifications
About the only thing known about the 5800 so far beyond these and what it looks like is the pinout for the electrical connector, determined by continuity tests and (hopefully) educated guesses:
Pin Function --------------------------------------------------------- CTR Laser Tube Anode (fat blue, goes to ballast) A PZT for OC (white) B Laser Tube Filament/Cathode 1 (0.4 ohms to pin C) C Laser Tube Filament/Cathode Common D Laser Tube Filament/Cathode 2 (0.4 ohms to pin C) E Photosensor behind HR (red dot) F Photosensor behind HR H Laser Head Heater (yellow, 660 ohms to pin L) J Temperature Sensor (white, 560 ohms to pin K) K Temperature Sensor (white) L Laser Head Heater (yellow) M PZT for OC (blue)
It's not even possible to see the actual laser tube without fairly extensive disassembly (which I'd like to avoid for now) being almost entirely enclosed inside a metal shroud. But it's probably a miniature version of other Perkin Elmer tubes of similar vintage - coaxial construction, with two Brewster windows and dual heated filaments (one spare), but only 3 to 4 inches in length.
Based on this, the patent, the physical size of the laser tube, and the front panel of the 5801 power supply/controller, the 5800 almost certainly uses the Lamb dip for locking. The heater and temperature sensor probably maintain the laser head at a constant temperature and the PZT is then used with the photosensor and a synchronous demodulator to lock to the minimum of the lasing output power curve. The 5801 has a knob to adjust cavity length to locate a Lamb dip, and a switch to enable locking.
Assuming my determination of the connector pinout is correct, the 5800 laser head I have is quite dead. Not that this should be very surprising. It won't start using a modern HeNe laser power supply between the CTR and BCD pins. The test was done without the filaments powered (though they are intact). It would not be good to attempt to run it that way but it should still have been possible to strike a discharge if it were gas intact. I don't believe it's physically broken, only that it has leaked after 40+ years. But perhaps wishful thinking, but having sat for so long, there is a chance that it's simply slow start and would be happier with a filament actually heated. Since I can't even see the tube, testing it with an Oudin/Tesla coil is out of the question. (However, I did use a power supply capable of starting a 5+ mW laser and that did nothing.) In addition, the photosensor tests open so it's not clear if that's bad as well or simply something other than a common photodiode.
More to come.
Case Split Minimum HP/Agilent
Model Size Frequency Power Equivalent Comments
----------------------------------------------------------------------------
LGN-212-1 L 1.5 - 2.2 200 µW 5517A
LGN-212-1M S " " " --- No HP/Agilent equivalent
LGN-212-1M-A S 2.0 - 2.4 " 5517B
LGN-212-1M-B S 2.4 - 3.0 " 5517C
LGN-212-1M-C S 3.0 - 3.4 " --- No HP/Agilent equivalent
LGN-212-1M-D S 3.4 - 4.0 " 5517D
Options for beam size are 6 (default), 3, and 9 mm as with HP/Agilent, though they are spec'd as 5-7, 3-5, and 7-9 mm, respectively. And the spec'd beam divergence is the same for all of them - 0.5 mR. Frequency stability is simply listed as 10 ppb without regard to any time scale.
There are also versions of these lasers with "F" in the model number - e.g., LGN-212-1MF-A. Quoting from the spec sheet: "HeNe gas laser of continuous operation mode, double-frequency, stabilized with inbuilt photodetector of laser emission, reflected from the inner optical interferometer system". I have no idea what that means, but they are in the small case, so it doesn't sound like they are intended to be similar to lasers like the 5518A or 5519A/B. Perhaps more like the 5500A! Or perhaps, PLASMA should hire a Russian to English translator!
But there are two very disappointing specifications associated with the LGN-212 series lasers. One is the "Mean life", which is only 5,000 hours for the LGN-212-1 and 10,000 hours for the others. This suggests that either Russian tube processing is still not up to modern standards or that the tube is very small - probably both. Such a short life expectancy might be acceptable for scientific use where the laser is only turned on for an experiment now and then. But the price would have to be really really low to justify consideration for a metrology application like a wafer stepper that's run continuously 24/7. If that's not enough, an even more troubling specification is the "Mean time to failure" of only 2,000 hours for all of them! So, expect a problem approximately every 3 months! (A typical HP/Agilent laser will operate continuously for several years or longer without requiring attention.) Hopefully a service engineer comes prepackaged with each LGN-212 laser at no extra charge. :)
I have not yet tested any LGN-212 lasers so I cannot say if they do indeed meet HP/Agilent specifications with similar frequency stability and freedom from rogue modes.
Renishaw has systems for machine calibration as well as integration into OEM equipment like wafer steppers. All laser heads include a built-in optical receiver (like the HP/Agilent 5519A/B), so only the interferometer optics are external. Their flagship product for calibration, the XL-80, puts the laser, its controller, optical receiver, and the measurement electronics in a stylish package that is a compact 214x120x70 mm (8.4x4.7x2.8 inches) in size and weighs a mere 1.85 kg (4.1 pounds). So, the only external connections are DC power (from a universal switchmode adapter) and a USB cable to a PC! The transfer rate may be up to 50 thousand updates per second to Renishaw's display and analysis software. The ML-10 was the predecessor to the XL-80 with lower performance, a different interface, and an uglier case. The HS-10, which is similar in performance to the ML-10, is for integration into OEM equipment. It would appear that a separate HS-10 is required for each axis in a multiple axis installation.
I would love to acquire an XL-80. I'd even settle for an ML-10 or HS-10 tossed in the trash after an upgrade. :) But I really can't justify the $34,000 price tag for a new system to tear apart even if it is to enhance the Laser FAQ. And these things simply do not appear on eBay! Raw dinosaur eggs are more common there.
The best summaries are probably provided in the respective brochures for the XL-80, ML-10, and HS-10 by going to Renishaw, "Calibration", "Laser Interferometer Systems", and then "Downloads" (on the right side of the page) where you can browse or use the search box to find these quickly. Also check out "Homodyne and Heterodyne Interferometry" in which Renishaw attempts to justify the benefits of their single-frequency-based laser approach for metrology applications.
This system had a sticker price of just under $5,000 but since around the time of the acquisition of Spectra-Physics by Newport, it is no longer in production. And a search at Spectra-Physics/Newport now comes up empty.
The SP-117 and SP-117A consist of a cylindrical laser head with cables for the high voltage and control signals, and a separate power supply/controller box. The 05-STP-901 from Melles Griot appears to be the same system as the SP-117A except for the front panel decor and color scheme as it has the same specifications, controls, indicators, and connectors - including the strange three-pin Spectra-Physics HV connector not found on any other Melles Griot lasers. In fact, the PCB inside the 05-STP-901 case has the Spectra-Physics logo on it and "Fab" and "Assy" numbers that begin with "117"! As if this is not enough, the 640 MHz mode spacing of the 05-STP-901 listed in the Melles Griot catalog is the same as the Spectra-Physics 088-2 or 088-3 HeNe laser tube used in the SP-117. And, Melles Griot *does* have an 05-LHR-088 tube which matches the 088 physically and has a mode spacing of 641 MHz. Coincidence? I don't think so. :) (However, inquiries to Melles Griot have not indicated any acknowledgement of this, nor the option to actually buy such a tube without the attached laser.) Thus, a new SP-117A would probably have a Melles Griot tube inside since Spectra-Physics has been out of the HeNe laser tube business for some time. See the section: Melles Griot Stabilized HeNe Lasers. However, the Melles Griot tubes have an AR coating on the HR mirror (likely in addition to being wedged) as well as on the OC mirror. The purpose would be to futher minimize backreflections and to get the most power from the waste beam since that's used for the mode sampling photodiode pickup. The minimization of backreflections is by far the most important for two reasons:
So, the existence of an AR coating on the HR is one way of determining that the tube you have is from or for a SP-117A or 05-STP-901 stabilized laser, rather than a barcode scanner. Common non-stabilized HeNe lasers - even very expensive ones - do not have this.
The SP-088 tubes may not have an HR with either an AR coating or wedge. This is of no consequence in an SP-117, but may result in a slow periodic variation in output power if used with an 05-STP-901 or SP-117A controller in Intensity Mode. The reason is that as a result of the etalon formed between the HR mirror coating and the outer surface of the mirror substrate, the waste beam power (used for feedback control) and output power from the laser will not quite track each other due to temperature changes not corrected by the stabilization feedback. The HR of the 088 tube in what I originally thought was an SP-117A head, but must have been an SP-117, had no AR coating or wedge but did have a curved outer surface. Whether this was an attempt to get around this deficiency is not known. But it does not work - when used with an SP-117A controller in Intensity Mode, the output power varied by 10 percent or more over a period of minutes or hours. But since the SP-117 does not have the Intensity Mode, it wouldn't matter there as the amplitude of the two polarized modes used for feedback would track each-other.
A search of the Spectra-Physics Web site used to return a link to a model 117 with some confusing text about the model 117, 117A, and 117B, but that 117 isn't the same as the original SP-117 or SP-117A described below, and the page seems to have disappeared. It even mentioned Brewster windows but none of these lasers ever had Brewster windows! The SP-117C is an OEM version which has the laser tube/heater, electronics, feedback PBS/PDs, and HeNe laser power supply mounted on a baseplate for integration into custom products. There may also have been an SP-117D, similar to the SP-117/A, also for OEM applications. The SP-117B is essentially an SP-117C mounted in a rectangular enclosure with the addition of an output polarizer (which the SP-117C lacks) and mechanical shutter.
The following description applies to both the SP-117 and SP-117A (and MG 05-STP-901) unless otherwise noted. The laser head I dissected was an SP-117, though I expect the newer ones to be very similar. A typical SP-117 is shown in Spectra-Physics Model 117 Stabilized HeNe Laser System. The SP-117A is in an similar package with the Frequency/Intensity mode keylock switch and Locked LED added. The PCB is a completely new layout to accomodate the added circuitry for Intensity mode but everything else is similar. Why change a good thing?
The HeNe laser head is powered from a HeNe laser power supply brick (approximately 1,700 V at 4.5 mA) via the usual strange Spectra-Physics screw-lock HV connector, with a separate cable with a DB9 connector for the photodiode signals and heater power. The only thing non-standard about the brick may be a lower p-p ripple and noise specification but there is no special external regulation of this power supply. However, for it to turn on requires that the interlock plug be present on the back of the controller, that the microswitch inside the HV connector be depressed by a plastic pin in the HV plug, and that pins 2 and 7 on the signal connector be jumpered. (Misadjustmant of that microswitch is one common way these can refuse to start.)
The cylindrical laser head contains the tube, output optics, and beam sampling assembly. A view of the parts after disassembly is shown in Spectra-Physics Model 117 Stabilized HeNe Laser Head Components. Sampling is from the waste beam at the HR-end - simply a polarizing beamsplitter (inside the black cylinder, upper left) feeding a pair of solar cells/photodiodes (glued to the metal bracket attached to it). Since this is at the HR-end of the tube, it doesn't reduce the output power. The entire guts can be pulled out by loosening a bunch of setscrews. Disassembly to the state of affairs in the photo took about 10 minutes, all completely reversible except for cutting small blobs of black RTV silicone holding the laser tube in place in the cut out aluminum cylinder at the top of the photo.
The HeNe laser tube itself used in the original 117 was some version of a Spectra-Physics model 88 - the same type that used to be found in zillions of barcode scanners. However, note that the version used in the SP-117/A has cathode-end output, unlike the anode-end output of the barcode scanner tube. A sample I have produces over 3 mW, so it's probably a higher power version, perhaps an 088-2. As noted, those of more recent manufacture may use the Melles Griot 05-LHR-088. I have no idea if the tube is special in any other way, like having been selected for no more than two longitudinal modes or filled with isotopically pure gases or blessed by the Laser Gods. :) There were no markings of any kind on this one and at this point, I rather think that the only special requirement is that the tube not be a "flipper" - one where the polarization state of the modes switches abruptly rather than remaining fixed. In fact, I rebuilt an SP-117 laser head with a surplus 088-2 tube and it would seem to work fine. See the section: Transplant Surgery for Two Sick Spectra-Physics Model 117 Stabilized HeNe Laser Heads.
The HeNe laser tube has the multiple layer aluminum foil covering Spectra-Physics is so fond of. There may even be more layers than normal, covering a larger portion of the tube than normal. A thin film heater (copper-colored cylinder) is wrapped much of the way around the tube on top of the aluminum foil, and glued in place. Finally, here is an application where the aluminum foil actually might make sense to distribute the heat uniformly. :) The short black cylinder on the right holds a (PBS) Polarizing BeamSplitter (with the reflected output blocked) to select one of the two orthogonally polarized modes of the laser, possibly optional since it was not present on one of the SP-117 laser heads, or an SP-117A laser head that I've seen. So, either, whoever originally had these things salvaged the PBS as the only remaining useful part before selling them, or that is an option not present on all units. On the SP-117, it's a normal PBS cube mounted such that the laser guts have to be slid out of the laser head cylinder to access the set-screw holding it in place. On the SP-117A, it's a PBS that has a rhombus cross-section - a slightly distorted cube - presumably to minimize backreflections. And on the SP-117A, the PBS is secured by two screws accessible from the front once the bezel is removed
Prior to assembly at the factory, the tube must be tested to determine the best orientation for maximum signal change of the two polarized modes since there is no adjustment for this once the tube is mounted with RTV silicone. The specific orientation is determined by slight asymmetries in the tube construction - random factors like mirror coatings and alignment - but should not change with age or use.
After an initial warmup period where the heater is run continuously, the controller enables the feedback loop which monitors the two outputs of the beam sampler and maintains cavity length using the heater typically so that the two orthogonally polarized longitudinal modes are equidistant on opposide sides of the neon gain curve (for best stability when running frequency stabilized) or where one mode is closer to the center of the gain curve (which may be desirable when running amplitude stabilized to get a bit more output power in that mode). Although I haven't measured it, there are probably around 75 complete mode cycles before locking. The switchover occurs when the period of a full mode sweep cycle exceeds about 18 seconds. (If rebuilding one of these lasers, that time may have to be adjusted due to differences in the thermal characteristics of the tube, and the thermal coupling and insulation of the heater.)
The user controls on the SP-117A consist of a switch for power and a switch to select between frequency and amplitude stabilization. There are indicators for AC power and Stabilized. (The SP-117 is physically identical but lacks the mode select switch.) After a warmup period of about 15 to 20 minutes for the laser head to reach operating temperature, the Stabilized indicator will come on and may flash for a few seconds, and after that should remain solidly on. This really indicates only that the stabilization feedback loop is active, NOT that the laser is actually stabilized and meets specs - that may require another minute or so. For the SP-117A and MG 05-STP-901, the behavior is similar in Frequency or Intensity mode. (The SP-117 has no Intensity mode.) In fact, the way they are designed, everything is identical in both modes until the Stabilized indicator comes on, then it switches to the Intensity signal for locking. If power is cycled, the delay to Stabilized is much shorter, so no actual counter delay is involved, just some circuit watching for the mode changes to slow down below that 18 second threshold. Indeed, if the photodiodes are disconnected, Stabilized will come on in under a minute even though the modes are varying wildly. Stupid electronics. :)
The internal circuitry of the controller box is relatively simple and includes some CMOS logic including several monostables (!!) for timing the warmup period, multiple op-amps and comparators, a 555 timer, voltage regulator, and switching transistor for the heater - all standard stuff. A linear power supply feeds the HeNe laser power supply and the control electronics,
Here is the pinout of the DB9 control connector as determined by my measurements. There may be errors.
Pins Function Comments ---------------------------------------------------------------- 1,6 Heater ~19 ohms, 12 V source on pin 6. 2,7 Interlock Shorted. 3 Ground May not be connected on some versions. 4,5 Photodiode 1 Anode is pin 4. 8,9 Photodiode 2 Anode is pin 8.
Although, it's very likely that any reasonably healthy SP-117/A laser head will lock on any SP-117/A controller, adjusting the controller for optimal mode signal swing will result in best stability and permit the location of the two orthogonal modes on the neon gain curve to be fine tuned.
However, since the SP-117 lacks the Intensity Stabilized mode of operation, it's possible that an SP-117 head installed on an SP-117A or 05-STP-901 controller may use the wrong photodiode for intensity feedback and thus not be very stable in that mode. The cable connectors for the PDs inside the laser head are not labeled and it would be possible for them to be swapped with no obvious effects, either at the factory or due to prior service. The proper PD is the one at the back (which correspond to the output beam), not the side. It's easy to test for this and swap the PD connectors inside the laser head if necessary. (Swapping the PDs will also shift the lock position in Frequency Stabilized mode to the opposite side of the neon gain curve - i.e., "red side" instead of "blue side" or vice-versa. But since there is no real specification on optical frequency or wavelength, this would normally not be apparent even with fancy instrumentation.) To test, loosen the outer cylinder and pull it out just far enough so that the PD at the back of the laser tube is visible. Power up the laser and allow it to lock in Intensity Stabilized mode. Once locked, slip a piece of paper between the PD and the laser tube. If the system looses lock, it's using that PD and is wired correctly. If nothing happens, power down, cut the cable ties securing the PD connnector, unplug it and and put back in rotated by 180 degrees.
If anyone has information on the official internal adjustment procedure for the SP-117 or SP-117A controller and/or a service manual or schematics, please contact me via the Sci.Electronics.Repair FAQ Email Links Page.
There are three sheets:
The PWM ramp generator ends up on the analog page only because it is actually in the same chip as the amplifiers.
In the laser head, Photodiode 1 is the one on the side (P-Mode, TP3) and Photodiode 2 is the one in the back (S-Mode, TP6). S-Mode is the signal used for Intensity stabilization and the corresponding polarization that normally exits the front of the laser. (At least that's how the sample I checked was set up!) Other combinations would be equivalent, but it's critical that the same polarization be used for the Intensity reference as the one in the output beam. I assume that the SP-117 is set up the same way (assuming there is consistency in this at all!)
Note that older versions of the SP-117A differ slightly in their design. The most obvious one being that the two relays found in the 05-STP-901 are missing so no clicking when switching modes. :) They appear similar to the SP-117 in other respects like the use of the ICL7660 or ICL7662 rather than 555 for generating the clock and -9VDC. But I haven't examined one in enough detail to identify other differences (and probably never will).
There are three sheets:
Some of the circuitry of the SP-117 and SP-117A is similar and the designer of the SP-117A clearly had access to the SP-117 schematics. However, substantial portions have been totally redesigned rather than simply tacking on the amplitude stabilization and modulation input as a minor addition. The low voltage power supply including the on-board regulation, the heater driver, the HeNe laser power supply, and at least some of the logic/timing circuitry are very similar. But in the SP-117A, the negative voltage source is provided by an a 555 rather than an ICL7660, different op-amps types are used in the analog sections, and the basic design of the control loop has also changed. Of most significance in terms of performance, unlike in the SP-117A, there appears to be no pure integrator stage in the analog chain of the SP-117. Note how much simpler the control loop (analog schematics pages) is for the SP-117 compared to the SP-117A! So there will always be an offset due to the finite and relatively modest gain between the photodiodes and heater response. This means that the exact position of the locked lasing line on the neon gain curve will have a dependency on ambient temperature and initial conditions when the SP-117 is turned on or restarted. And, it will drift somewhat until thermal equilibrium is achieved, which may take considerable time after the Locked indicator comes on. Exactly this behavior has been observed with the SP-117 and wouldn't make sense had the design been similar to that of the SP-117A with a full PID control loop. These anomolies were bothering me and in fact were the original reason I decided to trace the SP-117 circuit. My confusion was your gain. :)
This laser seems to be interesting in another respect: While for the typical ordinary HeNe laser, the modes roughly follow the profile of the gain curve as they traverse it, with this tube, the mode on one side will mode hop - disappear and reappear on the other side of the gain curve relatively abruptly - rather than varying in smoothly in amplitude decreasing to zero or near zero. I don't know whether this behavior is a peculiarity or a feature but it seems like it could be beneficial. ;-) I've seen similar mode behavior on at least one other HeNe laser tube intended for a stabilized laser - the Nikon NKL-85 (and the only other one I've tested in this manner so far!). That laser is always single mode with relatively constant mode amplitudes (variation of less than 20 percent) and abrupt mode hops. In common HeNe lasers with a similar cavity length, two modes with continuously varying amplitudes would be present over a large portion of the mode sweep cycle.
The SP-117B is essentially a SP117C installed in a genuine cheezy Spectra-Physics rectangular case with the addition of an output polarizer, half waveplate, and mechanical shutter. (Geez, this is such a high power laser!) Power/status is via a round 8 pin connector. The beam bounces off of two 45 degree mirrors whose main function must be to increase its height and center it as there appears to be no rational reason for turning the beam around 180 degrees so it exits the opposite end compared to the SP-117C. The half waveplate rotates the polarization from 45 degrees (the way the tube is mounted to match the orientation of the beam-splitter/photodiodes monitoring the waste beam, reason unknown) to vertical. Several photos of an SP-117B can be found in the Laser Equipment Gallery (Version 4.03 or higher) under "Spectra-Physics HeNe Lasers".
The description below applies to what's common to both the SP-117B and SP-117C lasers.
The SP-117C is a single unit mounted on a solid baseplate (with exposed high voltage!). It is designed to install in a cabinet painted with the decorator colors of your choice. :) See Spectra-Physics Model 117 OEM Stabilized HeNe Laser Assembly. The SP-117C has no output polarizer so both polarized modes appear in the output beam. However, the they are at +/-45 degrees - neither is vertical or horizontal as most users would probably desire. So, either a polarizer and 1/2 waveplate, or a pair of polarizers (the first at 45 degrees and the second vertical or horizontal) would be required for most applications. (The latter arrangement would result in an additional reduction in output power.) However, in a homodyne interferometer, a single linearly polarized mode at +45 degrees or -45 degrees might be what's used, which suggests that perhaps that was the original intended application for the SP-117C
On my sample, there was a separate box with a +/-12 VDC switchmode power supply and lighted power switch as the only user control. (This is probably NOT from SP, but a user/home-built unit.) Its operational behavior is similar to the other SP-117s, though the warmup is faster - under 10 minutes. Locking is then abrupt with little overshoot or ringing. Locking following a power interruption of a few seconds occurs in under 1 minute. When to switch from warmup to lock mode is probably detected by a complete mode cycle taking more than around 16 seconds.
The HeNe laser tube looks the same as the 088 used in the other SP-117 systems except that the thin-film heater connections are at the cathode-end instead of anode-end of the tube. And like the tubes in the SP-117/A, it's probably from Melles Griot by its relatively thin-walled construction. The 12 VDC input HeNe laser power supply brick is hidden underneath. The PCB generally resembles the one in the SP-117A and 05-STP-901 controllers with many of the same part numbers, though there are also many differences and it has clearly been substantially redesigned. The timing is now done using 12 bit binary counters instead of multiple monostables. The majority of the discrete resistors have been replaced with resistor packs. There is also a pair of resistor packs in sockets for reasons unknown. The input is +/-12 VDC (rather than just +12 VDC), supposedly being regulated to +/-9 VDC on-board according to the test point labeling. But the resistor that sets the voltage on the sample I have has been selected to produce +/-8 VDC instead of +/-9 VDC and that works fine. There is no oscillator to generate the negative voltage of the SP-117 and SP-117A controllers, so the associated PWM clock must be produced in some other way. There are pads for four large series diodes with jumpers in their place. These would be used to reduce the DC voltage to the HeNe laser power supply if more than 12 VDC were used for the positive power supply. Small MOSFETs are used to control the Enable line of the HeNe laser power supply and the Locked signal, as well as some internal signals. There are fewer test-points. Those that do exist have different numbers than on the SP-117A. (The most relevant are now TP7 and TP9 for the PD preamp signals (with R26 and R25, respectively, for thier gain adjustments), and TP8 for their difference.) And, in case you're wondering, I have absolutely no intention of reverse engineering this unit the way I did the SP-117A! At least there is no microprocessor or ASIC. So it doesn't run Windows, sorry. ;-) However, since I did need to troubleshoot the control PCB on one, I did determine that 3 sections of U11, a TL084, are used for these functions. (The op-amp used for the difference was dead.)
But I have determined most of the external connections to the 14 pin header visible in the upper left corner of the above photo based on how it is wired and the obvious PCB traces:
Pin Function -------------------------------------------------------------------------- 1 +Va - Positive analog power, +12 to +15 VDC. 2 +Va? 3 Mode control?? Input to NOR gate, pulled high. 4 Analog Ground 5 Locked Status (Unlocked: +Va V; Locked: 0 V, will sink 0.6 A). 6 Reset? 7 -Va - Negative analog power, -12 to -15 VDC. 8 -Va 9 Digital Ground 10 Digital Ground 11 Digital Ground 12 Digital Ground 13 NC 14 +Vd - Digital power, +12 VDC (+15 VDC with all diodes installed).
Gary Turner says that Pin 6 is connected to a front panel "Reset" switch on his controller. The manual says that you can use the Reset to clear a flashing "Lock" status, which can occur if there is a temperature/power issue.
The Locked signal originates from an IRFD210 MOSFET which can sink 0.6 A, more than enough current to drive an LED - or a bank of them. :) The +/-12 VDC for the unit I have comes from a small switchmode power supply in a separate box. The analog and digital positive voltages (+Va and +Vd) are the same. I added a Locked LED there and will install a switch if the unidentified control signal does something useful.
A reset function could be useful (and one was found on a commercial system using an SP-117C). I have seen the Locked LED start flashing during testing, probably due to a loss of lock resulting from back-reflections, air currents if the laser isn't covered, or a power glitch. However, a momentary interruption in the photodiode signals does NOT trigger the flashing condition, probably because they both go away and the error signal doesn't have a chance to increase significantly. While the laser re-acquires lock automatically, knowing that it did something bad could be desirable. But without a reset function, it's necessary to power cycle to clear the error so the Locked LED remains on solid.
That commercial system coupled the output of the laser into a single mode APC fiber. An optical (Faraday) isolator and adjustable fiber port were mounted on the same base-plate as the SP-117C laser assembly. The input polarizer in the isolator blocked one of the pair of longitudinal modes present when locked so the output was pure SLM.
I am in need of a user manual for the SP-117C (and/or SP-117B) including info to confirm what I have on the 14 pin header is correct and to determine the function of the unidentified control signal. (Grounding it either during operation or prior to power-on has no obvious effect.) If you have any info, please contact me via the Sci.Electronics.Repair FAQ Email Links Page.
The spec'd output power for the original SP-119 is a whopping 70 microWatts (µW) at 632.8 nm, the normal red HeNe wavelength! With only temperature control, the frequency stability was spec'd to be +/-75 Mcps/day. This is before Hertz was used for cycles per second (cps)!. With the optional Servo Option 259-002 for active cavity length control using the Lamb dip for stabilization (which is actually more sophisticated than modern mode-stabilized HeNe lasers) it goes down to +/-5 Mcps/day.
And the price loaded for the 119 laser head and 259 exciter with the 259-002 Servo Option in 1964: $5,775.00. That's about the same cost as a his and hers set of Chevy Impalas. :)
The specifications for output power and stability were subsequently upgraded to 100 µW (WOW!) and +/-1 Mcps/day. The latter IS impressive and represents a stability of +/-2 parts-per-billion (ppb), which is better than most modern stabilized HeNe lasers.
Here are the specifications for the SP-119 laser head with the SP-259B exciter from the Spectra-Physics 119 Gas Laser Operator Manual. I've tried to use the original units and terminology:
SP-119 Laser Head
---------------------------------------------------------
Output Wavelength: 6,328 A
Output Power, uniphase, single freq. >100 µW
Beam Diameter (1/e2) at laser aperture: 1 mm
Beam Divergence (1/e2), no beam expander: 10 mrad
with 3 mm beam expander: <0.3 mrad
with 6 mm beam expander: <0.15 mrad
Long Term Frequency Stability - deviation from
neon-20 emission center after warmup assuming
a maximum ambient temperature change of +/-1 °C.
Without servo control: +/-75 kc/day
With servo control: +/-1 kc/day
Warmup from Cold Start:
Without servo control: 3 hours
With servo control: 45 minutes
Warmup from Standby:
Without servo control: 30 minutes
With servo control: none
External Modulation:
Maximum Deviation (10 to 3,000 cps): 1,200 Mcps p-p
(20,000 cps): 200 Mcps p-p
Sensitivity: 12 Mcps/V
Servo FM Deviation: <5 Mcps p-p
Servo FM Frequency: 5 Kcps
Laser Head Dimensions: 8-1/2" (D) x 6-3/4" (W) x 4-1/2" (H).
Laser Head Weight: Approximately 10 pounds.
SP-259B Exciter
------------------------------------===-----------------------
Plasma Tube Current: 4-10 mA at approximately 2,000 V.
Current Regulation: 0.1% for +/-10% line or load changes.
Exciter Dimensions: 12" (D) x 16-3/4" (W) x 5-1/2" (H).
Exciter Weight: Approximately 25 pounds.
Input Power: 115/230 VAC, 50/60 cps, 250 VA maximum.
The Lamb dip isn't something that goes with mint sauce or cheese and crackers. :) It is a depression at the center of the gain curve that may occur under the proper conditions in a laser with an inhomogeneously broadened gain medium and is the result of hole burning or depletion caused by the lasing process in a standing wave cavity. W. E. Lamb was an early laser researcher who first predicted its existence in: "Theory of an Optical Maser", Phys. Rev. 134, pp. 1429 (1964). A. Szoke and A. Javan described it in a more useful form in: "Isotope Shift and Saturation Behavior of the 1.15-µ Transition of Neon", Phys. Rev. Lett., vol 10, issue 12, pp. 521-524, June 15, 1963. (Letters are published much more quickly than full papers which is why I assume this seems to be acausal.)
In a HeNe laser, the Lamb dip is the result of hole burning or depletion in the inhomogeneously Doppler-broadened neon gain curve. It isn't present in all such lasers, but under some conditions, mostly determined by the design and construction of the laser, it will appear as a small depression at the exact center of the neon gain curve where its peak should be, as shown in The Lamb Dip in a Helium-Neon Laser and Longitudinal Modes of Short HeNe Laser with Lamb Dip. (The Lamb dip is not something that comes and goes, though the health of the laser tube and thus its gain does affect it.) In a nutshell, the explanation is as follows:
The Doppler-broadened neon gain curve really respresents a distribution of atomic velocities, with zero velocity being at the center. Atoms moving toward a photon traveling down the (Z) axis of the laser have a higher frequency at which stimulated emission can occur. Atoms moving away from a photon traveling down the (Z) axis of the laser have a lower frequency at which stimulated emission can occur. So, a photon traveling in the +Z direction will only be able to produce stimulated emission if an excited atom that it encounters is moving with the specific velocity needed to Doppler shift the neon gain center frequency by the appropriate amount so it equals (within the natural line width) the photon's frequency. If the photon's frequency is above the neon center, then the atom must be moving toward the photon with a velocity of, say, -V. However, the exact same conditions will be met by a photon traveling in the -Z direction and an atom traveling with a velocity of +V. And this is exactly the same offset from the center on the opposite side of the Doppler-broadened neon gain curve. Thus, the result is two depressions until the cavity tuning is such that the F-P resonance condition is at the very center and the lasing process is drawing on the zero velocity population.
In the diagram, the cavity tuning is increasing in frequency (the cavity is getting shorter) clockwise for each successive residual gain curve (1 to 5). The "Unsaturated Gain" is present when there is no cavity to enhance stimulated emission, or below the lasing threshold. The "Saturated Gain" is present while lasing. And the "Output Power" is the useful beam from the Output Coupler (OC), the partially transparent mirror. (The amplitude of the dips in the diagram are somewhat exaggerated compared to what is typical in practice, and nothing is guaranteed to be to scale!) Though the dip pairs can't be observed in the output of the laser, they would show up if the single pass gain were measured using a probe beam from a tunable laser passed down the bore of the tube. Energy is being extracted from the gain curve to produce the intra-cavity (and output) beams through stimulated emission. So, what's left will be reduced in the areas where this takes place. (And there's an entire scientific field called "Lamb dip Spectroscopy" that involves the finer points of this phenomenon.) (A more detailed explanation of the Lamb dip can be found at the end of this section.)
Under the proper conditions (more below), the Lamb dip will result in a very pronounced variation in laser output power as the cavity is tuned, and this can be used to accurately lock the laser to the center of the neon gain curve. The locking technique is actually quite simple: The cavity length - and thus the optical frequency - is periodically varied, or "dithered" by a small amount (a few MHz) via a PZT to which the HR mirror is attached. This results in a corresponding variation in laser output power based on the profile of the neon gain curve on either side of lasing mode location. The optical power is sensed via a photodiode monitoring the waste beam from the HR mirror. The lasing line can then be maintained at the exact center of the neon gain curve where the Lamb dip is located by satisfying two conditions:
(1) locks to an inflection point and (2) forces it to be a local minima.
The electronics is designed to generate a low speed correction signal to drive the PZT to maintain these conditions. In modern terminology, the circuitry to do this would be called a lock-in amplifier, phase-sensitive detector, or synchronous demodulator. Even using 1960s technology, it isn't very complex.
The SP-119 system can still be used with manual control of wavelength (Lambda, or equivalently, optical frequency), which varies the DC voltage on the PZT via a 10 turn pot on the front panel of the SP-259. After a short 3 hour warmup :), the wavelength (and thus optical frequency) will be fairly stable as a result of the thermal regulation of the resonator. But once the approximate gain center has been found by monitoring the servo error (on the built in meter), switching to the "Lock" position should maintain the laser on the center of the neon gain curve (center of the Lamb dip) indefinitely. Should the system lose lock for any reason, even momentarily, an "Error Alarm" indicator will latch on.
A number of requirements must be satisfied to result in a pronounced Lamb dip (or any Lamb dip at all) whose center frequency is relatively independent of tube current and output power, and it's not generally observed in common commercial HeNe lasers. The most important of these are probably:
For the HeNe laser, the homogeneous linewidth is about 100 MHz compared to the 1.5 to 1.6 GHz for the full inhomogeneously Doppler-broadened gain bandwidth (FWHM) of neon.
The SP-119 uses isotopically pure 3He and 20Ne which results in a Doppler-broadened neon gain curve where zero velocity actually corresponds to the very center or peak. This is required for the symmetry condition as noted above. With mixed isotopes and a smeared out gain curve - or one with multiple peaks - the merging of the symmetric dips would not be distinct or coincide with the neon gain curve center
The SP-119 has a conventional linear Fabry-Perot (standing wave) resonator.
The SP-119 uses a (nearly) hemispherical resonator configuration with the rear (HR) mirror being planar and the front (OC) having a RoC just slightly greater than the cavity length (to guarantee resonator stability). This results in the diameter of the intra-cavity mode volume near the HR mirror being very narrow which fully saturates the gain in that region and results in a reliable Lamb dip regardless of overall gain, which changes as the tube ages.
The SP-119 resonator is just under 10 cm long corresponding to an FSR of over 1.5 GHz. So, the nearest adjacent longitudinal mode is 1.5 GHz away and well into the tail of the neon gain curve.
In fact, even very short barcode scanner HeNe laser tubes like the Melles Griot 05-LHR-007 (mirror spacing of 110 mm, 1.36 GHz FSR) show no evidence of a Lamb dip. Though these tubes typically have a long radius hemispherical resonator and satisfy most of the other requirements, they may not have isotopically pure gases.
And note that while longer tubes like the SP-088 (or Melles Griot 05-LHR-088) may produce a very distinct valley when one mode is near the center of the neon gain curve as shown in Plot of Spectra-Physics 088 HeNe Laser Tube During Warmup (Detail), this is NOT the Lamb dip, but simply a consequence of the relative amplitudes of the other modes that are oscillating. However, the SP-259 should have no trouble locking a longer laser tube to that valley. I may try that using a one-Brewster HeNe laser tube with the OC mirror on a PZT, configured to be about the same length as an 088. But, this would not result in a single frequency laser unless the cavity were somewhat shorter as there would probably be two other weak modes lasing on the tails of the neon gain curve. Since all modes have the same polarization orientation due to the Brewster window, it's not possible to suppress these unwanted adjacent modes. See the section: Using the SP-259B to Control Some Other Stabilized HeNe Laser.
More on the Lamb dip:
The following much more detailed explanation of the Lamb dip was excerpted from Professor Tony Siegman's book, LASERS, Chapter 30, page 1,199: "Hole Burning and Saturation Spectroscopy".
(Forwarded by: Confused2.)
At the beginning of the chapter, he starts with saturation of homogenous media and notes that saturation at a particular frequency within the atomic linewidth results in a reduction or shrinkage everywhere of both the real and imaginary susceptibilty curves. He emphasizes this with "a strong saturating signal even well out in the wing of the atomic transition will, if strong enough, saturate the entire transition uniformly across its line shape." Thus, hole burning cannot exist in homogenous media.
In section 30.6: "Inhomogeneous Laser Oscillation: Lambs Dips", he goes on: "In an early and widely read analysis of the gas lasers, Willis Lamb predicted, and experimenter soon confirmed an unexpected aspect of Doppler-broadened gas laser oscillation. If we tune the resonance frequency of a single oscillating cavity mode across a Doppler- broadened gas laser transition, the curve of oscillation power output versus cavity frequency shows a comparatively sharp and narrow dip in output power when the oscillation frequency coincides with the center of the Doppler broadened line." He notes that the Lamb dip only occurs in standing wave cavities and "is a consequence of saturation and hole burning effects in the Doppler-broadened line, caused by two oppositely traveling waves in the cavity."
"Physical explanation: The signal field inside a standing wave cavity can be divided into two oppositely directed traveling waves which we have referred to as +z and -z waves. Any single atom with axial velocity v thus sees two opposing traveling waves, for which it has equal and opposite Doppler shifts, even thought the two waves are at the same frequency. This leads to double whole burning effects" and thus a Lamb dip. "Consider a laser with an inhomogeneous Dopper-broadened transition oscillating in a single frequency standing-wave axial mode resonance, with the frequency w of this resonance detuned from the atomic line center by several inhomogeneous linewidths or whole widths.... The traveling +z wave component of the standing wave cavity fields will interact with and burn a hole in only those atoms in the velocity class given by v/c-w0-w/w0; while at the same time the fields in the - z traveling wave component will burn an equal and opposite hole in the symmetrically located velocity class at opposite value of v/c. Whenever the cavity frequency is well away from line center on either side. Therefore, two symmetric holes are burned, and in essence the laser is able to extract power from two separated set of atoms or velocity classes in the atomic velocity distribution8on."
"Velocity class" is jargon for a small range of velocities (never thought about why the term "class" was used, but that was the usual terminology from the beginning). As usual when considering continuous distributions, "small" is not precisely defined, but is to be taken as meaning a range small enough that all the atoms within it act more or less the same - in other words, somewhat smaller than the velocity spread that would create a Doppler frequency shift larger than the atomic linewidth of those atoms. You can start off thinking of a discrete number of velocity classes, which you sum over; then make these classes narrower and more numerous, until you're really integrating rather than summing, in which case each velocity class takes on in fact a differentially small range."
"If, however, the cavity frequency is tuned exactly to the line center, both the +z and -z waves can interact only with the v=0 velocity class in the Dopper distribution. This velocity class is therfore saturated twice as heavily as either of the separate velocity classes in the off- resonance situation because it sees two signal instead of one. But this means that the laser need only oscillate roughly half as hard to produce the same degree of saturation needed to reduce the gain to equal the cavity losses. In essence the two symmetric holes coalesce into one, and the laser power is taken from the single velocity classes v=0. In the inhomogeneously broadened single frequency laser this results in the slight, but definite dip in laser power at the line center known as the Lamb dip."
Since the SP-119 locks on the center of the gain curve using the Lamb dip rather than on one side of it, the resonator needs to be somewhat shorter than those of dual polarization mode stabilized lasers to guarantee that adjacent longitudinal modes - which have the same polarization and thus can't be separated from the middle one - are are far enough away that they don't have enough gain to oscillate within the 1.6 GHz Doppler-broadened neon gain curve. The active discharge length is 7 cm while the distance between the mirrors is 9.7 cm corresponding to an FSR of a bit over 1.5 GHz. This very short cavity length is intended to guarantee that at most, only a single mode can lase. However, as a practical matter, it's quite possible a somewhat smaller FSR would be adequate, and provide more output power in the lasing mode. In fact, with such a large FSR, there may be no lasing at all over a portion of the mode sweep cycle - where two adjacent modes are on the tails of the neon gain curve. One paper that mentions the SP-119 suggests that 10 cm RoC mirrors are used to assure a single spatial mode (TEM00) beam profile. (But I don't know if these are exactly the RoC used in the standard SP-119 laser. See "Pressure Shifts in a Stabilized Single Wavelength Helium-Neon Laser", A. L. Bloom and D. L. Wright, Proceedings of the IEEE, vol. 54, no. 10, Oct., 1966, pp. 1290-1294.) With the very short cavity, the output power varies significantly over a mode cycle and the beam may disappear entirely over a portion of it, especially with a weaker tube.
There are no real adjustments for mirror alignment as this is determined by the precision ground resonator assembly. (Though, tightening the various screws that hold the resonator sections together have some effect!) However, there are bore centering adjustments, which essentially align the bore to the mirrors. :) The HR mirror is on a PieZo Transducer (PZT) for cavity length control. The active part of the tube (the bore and Brewster windows), mirrors, and PZT, as well as a heater and sensor for thermal control, are all enclosed in a Mu-Metal cover to shield the gain region from stray magnetic fields.
I first acquired an SP-119 laser head (Model 119, S/N 3143-532) but no controller. It appears to be mostly complete though the output beam expander was missing. That was probably the only thing the previous owner considered useful after completing experiments with the laser!
The wiring is rather overly complex with 3 separate cables that run between the laser head and controller. (In fact, the fancier version that goes with the servo lambda control unit has 4 cables; this one lacks the cable for the photodetector.) The cable and connector for the external HeNe laser power supply is HUGE, a bit excessive considering that it's basically a sub-1 mW-class tube! This is pre-Alden though.
I had absolutely no doubt that this tube (call it #1) would be completely dead and up to air. However, upon removing the top cover, I was almost dazzled by the getter spot, which was huge and nearly like new. Something wasn't right here. I'd expect that with a modern hard-seal tube, but not something presumably from the 1960s. With a modern HeNe laser power supply hot-wired to the tube directly, it lit instantly with a bright stable discharge, but no sign of an output beam. See: Spectra-Physics 119 HeNe Laser Tube 1 With Good Complexion. The large silver/black getter spot with just the slightest evidence of contamination around its periphery is visible near the top of the photo. The white block contains the ballast resistor normally used with the SP power supply. The cylinder on the left is the Mu-Metal and thermal cover for the tube bore, optics, and PZT.
Then I noticed a sticker that had fallen off the side of the tube:
HeNe 9:1 @ 4.0 Torr 2-14-86 Isotopes 3 & 20 Mfg. by El Don Engineering
El Don Engineering is apparently a company founded by the brother of the owner of Jodon, Inc., but now defunct as Google could find no reference to it.
So, this was a rebuilt or replacement tube manufactured in 1986, and with a decent sealing technique as there has been almost no leakage.
There are photos of the Spectra-Physics 053, the tube that was probably the original one from the SP-119 laser in the Laser Equipment Gallery under "Spectra-Physics HeNe Lasers". Tube #1 in my SP-119 laser head may have a larger gas reservoir but is otherwise similar with the short two-Brewster bore mounted on the side far away from the gas reservoir, though the main bore and Brewster windows are hidden by the Mu-Metal cover Note how short the bore is - the 7 cm active gain region is similar to what is in a 1 mW tube with internal mirrors, and there will be significant losses through the Brewster windows, so the output power of this tube will be lower. (As noted above, the spec'd output power is only 70 µW.)
Why wasn't it lasing? I would have expected it to be burning holes in the wall. :) Sure, alignment could be messed up and/or the optics could be dirty after 20+ years. But then I carefully looked in the ends) and at first thought there were no mirrors! The discharge was clearly visible and bright with no hint of the blue coloration or reflections that would be present with 633 nm mirrors. Someone ripped the mirrors out for other projects? Ridiculous! That didn't make any sense. Not only would it be rather substantial effort to get to the mirrors to remove them and then put everything back in place without even any missing screws, but why bother? Aren't they just ordinary red HeNe mirrors?
Then something occurred to me: That 4.0 Torr is a rather high pressure for a 632.8 nm HeNe laser and the discharge was rather bright and more orange than usual, which would be consistent with the higher pressure. Normally, it should be 2 to 3 Torr for a visible HeNe laser. And a 9:1 He:Ne ratio is also rather high - 5:1 to 7:1 would be more typical. I didn't think the use of isotopically pure gases would be used routinely in common HeNe lasers, at least not back then. (However, as it turns out, genuine SP-119 laser tubes all do use isotopically pure gases, needed to obtain a Lamb Dip. More later.) But all three of these would make sense if someone wanted to do experiments with an IR stabilized HeNe laser! So I dug out my IR detector card. At first I didn't see anything. But in a dark room, there was just the faintest evidence of a lasing spot. Yikes! A 20+ year old tube still lasing on a (likely) low gain IR line in a 40+ year old laser! Not only is this laser still functional, it is a most unusual specimen!
Next to determine the wavelength. There are only two HeNe IR lasing lines that are likely: 1,152 nm and 1,523 nm. I suppose 3,391 nm might also be possible but I don't think my IR card would show it. There are more than a half dozen other near-IR HeNe lasing wavelengths, but their gain is much lower and I've never heard of anyone doing anything with them except to prove they are possible. A thermal laser power meter barely registered anything, perhaps 20 µW. But the silicon photodiode-based power meter I use for testing HeNe lasers also had a barely detectable response. If the wavelength had been longer than 1,100 nm or so, a silicon photodiode would have been totally blind. So, the lasing wavelength is most likely 1,152 nm. That is also consistent with the color of the mirrors (or lack thereof). Mirrors for 1,523 nm tend to have a slightly pink or tan appearance in transmission, and green appearance in reflection. Mirrors for 3,391 nm also have rather pronounced characteristics - possibly clear for the OC but often totally opaque for the HR.
After playing with the mirror adjustments for awhile (including losing lasing and having to use an external alignment laser to get it back!), I was able to increase output power by a factor of 2 to 3, to somewhere between 50 and 75 µW. For a two-Brewster tube this short and optics that were probably last cleaned over 20 years ago on the weak IR line, that is certainly acceptable. :) I don't know what the HeNe gain is at 1,152 nm but if it's similar to the gain at 1,523 nm, a tube that produces 5 mW at 633 nm will only produce 0.5 mW at 1,523 nm. The tube in the SP-119 is at best good for 0.5 mW at 633 nm if it had internal mirrors. It will be less with the two-Brewster window tube. The SP brochure only specs 70 µW at 633 nm! So a similar output power at 1,152 nm is truly amazing.
Now the question becomes: What do I do with this laser? Retain it in its present form as a something unique in the Universe, and also probably rather useless for anything I'd want to do? Or, replace the mirrors with normal HeNe 633 nm mirrors and have another 0.5 mW laser, but one that's consistent with the original SP-119? This is the dilemma I face! ;-) One thing is certain, I won't attempt any mirror transplants until I've had a chance to examine another (probably dead) specimen of this laser to determine the required technique that would minimize exposure of the Brewster windows since cleaning them is not likely to be a pleasant experience.
For more on the IR SP-119 laser head, see the next section.
I later acquired another SP-119 tube (call it #2) including heater jacket (thanks Kevin!), but no laser head OR controller. :) (At least I assume it's an SP-119 tube since I am not aware of other lasers that used one that is similar and can't imagine that there are any.) See: Spectra-Physics 119 HeNe Laser Tube 2 With Good Complexion. The distance from Brewster tip to Brewster tip is about 3-5/8" (9.2 cm). The actual bore is enclosed by the black cover. I had assumed this was a heater jacket used for thermal control in the SP-119 laser, but it seems rather odd to put it only around the bore. On this sample at least, it is Epoxied in place and definitely non-removable. A pair of wires goes inside with a resistance between them of around 300 ohms, kind of high for a heater in the feedback loop, but I later found this to be used during Standby where the laser tube is turned off. It provides a power dissipation similar to that of the bore discharge, so that the time to stabilize after coming out of Standby is greatly reduced (from 3 hours to around 45 minutes!).
This tube (#2) had no sticker on it but the glassowrk is the same as that of the geniune Spectra-Physics 053 tube so it is probably original, or a non-El Don exact copy. (Stickers don't generally fall off of SP tubes!) It may have never been used (or used for one experiment!) and has a large portion of the getter spot remaining. As can be seen, it lights up nicely and the bottom photo is especially spectacular with subdued lighting. :) The operating voltage is nice and low, it starts instantly, and runs stably at a very low current - down to 3 mA or less. The discharge in the expanded tubing doesn't appear quite as orange as the other one, so it may be filled at lower pressure for the normal 633 nm (red) wavelength, but it's hard to really tell without seeing the exposed bore, and that isn't going to happen. :)
Phil insists that he emailed me repeatedly about the significance of this laser when I first acquired it but I have no mental or email record of that at all! :)
(From: Phil Bergeron.)
I think I did tell you back then but either you were not listening or you did not understand that I was talking about the particular tube you had. Only one IR SP-119 was ever made. El Don described the rebuild to me in detail. Imagine if the line was far weaker and invisible washed out in bore light. :) The customer paid big bucks.
Don said it was the only one like it in the world. It was very hard to get the gas mix right with the short tube. I should have paid more attention when he was going on and on about the tube job and how pure the gasses had to be etc. :) Look him up; see if he is still alive! Don Gillespie near where Jodon is. His brother is John...Jo (hn) don...Jodon. :) In the old days I used to call both Don and John once in a while with laser tube questions and catch them up on how each other were doing. :) I wonder if they ever made up and shook hands? My favorite guy at Jodon was the technician Mike Christians. He gave me some good advice but their tubes sucked. Dirty, leaky. Don was the vacuum system guru.
The 119 tube was contracted for a customer who wanted an IR SLM laser. I do not know the source of the mirrors; they may have been provided by Don or by the customer but certainly did not come in a Spectra Physics 119 head! It was re-gassed with isotropic special high purity gasses in a non-standard ratio to optimize IR output. Special attention was paid to cleaning and placing a new getter to make sure gain was as high as possible since the tube is so short and the line is weaker than the 632.8 nm line gain wise. The rebuild cost several thousand dollars back when that was actually worth something.
The tube was completely rebuilt with a larger gas reservoir. I'm not sure how much of the original tube was retained, possibly only the heater jacket. El Don mentioned that the expected output power of a new red SP-119 laser is between 180 and 250 µW but I don't know if that's from SP or one of his rebuilds.
I do not have any of the short RoC (~10 cm) mirrors used in the actual SP-119 head, so a 20 cm RoC OC (with a planar HR) will have to do. The resonator parameters won't be precisely identical as it will be a 150 mm long radius hemispherical cavity rather than a near hemispherical 100 mm cavity. But hopefully this will be close enough for some Lamb dip to be present.
Here is the mostly completed SP-119 Tube Test Stand. The conical extensions on the SP-119 heater jacket are clamped between a pair of mating rings attached to angle brackets. They can be adjusted slightly in X and Y to center the bore. The planar HR is mounted on the left New Focus mirror mount while the curved OC on the PZT beeper is on the right one. (Its two pin connector can be seen behind the SP-119 tube heater jacket.) The mirrors are recessed so that pieces of tape can be stuck over them to prevent contamination when this thing is not being used (which will likely be all the time once the tubes I current have are all tested). The rational for putting the OC on the PZT is that slight changes in alignment caused by the PZT as it moves will have less effect with a curved mirror. But perhaps this is a fantasy. The decorative hole pattern on the baseplate is from its previous life, purpose unknown. :)
While on the test stand, the SP-119 tube will be powered from a Melles Griot 05-LPL-379 via an adapter that includes a current meter, not the SP-259B. It's just a wee bit more convenient! (The pair of blue wires are for the heater jacket resistors and the thin black wire is the case ground, neither used here.)
(Several weeks pass....)
Unfortunately, this turns out to be somewhat more difficult that at first thought due to many unknowns. One of these is the available gain of the two-Brewster laser tube. After failing to obtain even a single coherent red photon with a tube I believed to be good, I set up the single pass gain test using an SP-117C stabilized HeNe laser for the probe beam. The stabilized laser provides a, well, stable intensity so that turning power to the SPY-119 tube under test on and off will change the output power of the beam after passing through the tube in a predictable manner without worrying about mode sweep. The results are as follows for this tube and another one with a somewhat pink discharge. Both are quite likely at least 40 years old:
Net
ID Power Off Power On Bore light Difference Gain Comments
-----------------------------------------------------------------------------
2 1.179 mW 1.199 mW 0.006 mW 0.014 mW 1.117% Perfect color
5 1.925 mW 1.947 mW 0.002 mW 0.020 mW 1.103% Slightly pink tube
" 1.829 mW 1.857 mW 0.002 mW 0.026 mW 1.140% After 4 hours
(The specific value of absolute power is due to steps taken to restrict the maximum to be within a range where 3 significant digits of precision were available from the laser power meter. So, it's exact value and any changes in value with respect to the specific tube or set of measurements is due to changes in alignment of the probe laser and test laser bore.)
so, unless there is some unidentified optical damage to the Brewster windows, tube #2 should lase. But 1.117 percent doesn't allow much room for loss! However, it's not quite as bad as it sounds since that's 2.234 percent round trip. And for the discharge length of about 75 mm, the value is consistent with the textbook value for HeNe single pass gain of around 10 percent per meter.
And now knowing that tube #2 was just playing hard to get, and after threatening it with living the rest of its life in a museum, I have now been able to obtain a whopping 43 µW when installed in an SP-119 laser head. I'm expecting much more but I was fearing that cleaning two Brewster windows would be a pain, and getting to a low enough level of contamination for such a short low gain tube would be even more of a treat, especially when it appeared to be difficult to access the Brewster windows once the tube is installed. However, then I realized that each window could be cleaned individually without removing the tube, only the end mirror assemblies. And by doing this only at the OC-end, it is now up to around 97µW, which is enough above the spec'd minimum power of 70µW that I won't push my luck. The cleaning technique I use is a single swipe with a new cotton swab dampened with 1 or 2 drops of pure isopropyl alcohol. I have not attempted to clean the mirrors because (1) they appear to be pristine, (2) their coatings may not be as robust as modern ones so leave well enough alone, and (3) I hate cleaning laser mirrors! :) But it's quite possible mirror cleaning (as well as HR cleaning) would also help.
I have not tested the pink-complexioned tube in a laser head but as can be seen above, it actually has a higher gain than the one I am using. So, the color may be more due to a different gas-fill ratio or pressure than to contamination. But knowing that it should lase, it will probably go on the test stand.
Here is a summary of the SP-119 laser heads so far:
Head Output
Tube ID Wavelength Power Comments
------------------------------------------------------------------------
1 1,152 nm 50-75 µW Unique El Don IR SP-119
2 633 nm 97 µW Appears unused
3 " " 15 µW Good color, high dropout current
4 " " 100 µW Good color after running
Tube #5 will probably go on the test stand since its gain is now known to be plenty high.
Stay tuned.
The laser head is Model 119-3683, S/N 578, and the controller is Model 259-3664, S/N 579. Overall, the system is in very good condition for equipment at least 36 years old. The interior of both the laser head and controller could pass for new, with only minor signs of wear on the exteriors, some rotted rubber grommets in various places, and decayed foam pads cushioning the tube. They must have been well stored as there is even very little dust inside.
The power supply/controller (what SP called an "exciter") for the SP-119 laser head is the SP-259. This one is labeled 259B. I'm not sure what the difference is between the "B" and straight 259 or 259A, if there is one (though the improved specifications seem to be associated with the B version). The only obvious difference is that the 259B has a three position switch for Lambda (frequency) Modulation - Off, 60 Hz, and External, while the original 259 only has a toggle for Off or On, with the External BNC. The modulation (input) bandwidth is at least 20 kHz, though the p-p optical frequency excursion does drop off from 1.2 GHz between 10 Hz and 3 kHz, but only 200 MHz at 20 kHz.
To emphasize how ancient the design of this system really is, the SP-259B uses vacuum tubes in the HeNe laser power supply. A 6GJ5 high voltage beam power tube is the current regulator, controlled by a a 12AX7 used as a differential amplifier with a 0A2 gas tube (basically a big glass 150 V zener diode) as the voltage reference. And the main power supply uses four more 0A2s. Based on the date from the SP brochure (above), the original SP-259 was available in 1964. Or at least SP starting testing the market for the SP-119 laser in 1964! However, my SP-259B has what appear to be date codes on the main electrolytic capacitors of 1973 if I'm interpreting the labeling correctly (and assuming they are original). The latest date of the Operator Manual is 1966.
The SP-259 provides the following functions:
Typical Output Power versus Cavity Length for SP-119 Lamb Dip Stabilized HeNe Laser shows what to expect. Depending on the health of the tube, the "Mode Hop" point (and its surroundings) may actually result in an output power of exactly 0.0 mW.
In "Servo Null", the meter should move in a "Z" pattern as the Lambda control is rotated in one direction, roughly as depicted on the meter face - from near the lower limit to above center, back to below center, and to near the upper limit. The currect lock point is in the middle of the center leg of the "Z". When switched to Lock at the correct setting, the meter needle should barely twitch. If this is done at the wrong setting, the meter needle may swing to one end or the other, and the "Lock Error" light will come on.
It would have been nice to have included a meter mode to monitor the actual laser output power. With the photodiode present in the laser head, this would have been a trivial enhancement, eliminating the need for an external laser power meter to set the operating point when using the Manual Lambda Control, and as a confirmation of correct lock point with the Automatic Lambda Control. Not to mention being able to keep track of how many photons this powerful laser is blasting out the front! :)
Before applying power, I checked the ESR of most of the electrolytic capacitors and they all were reasonable. Even those orange Sprague Atoms showed very low ESR, so I don't know why one of them seemed to have been unhappy in the past. It's also not entirely clear why they need to be rated at 600 V as I only measured about 400 V on them. So, this gave me confidence to actually apply power to this thing. Only the HeNe laser tube connector is plugged in so far.
And, you're not going to believe this, but the laser works, sort of. I'm getting a maximum output power of a whopping 15 µW from this tube (call it #3). Despite the cloud of death getter, the discharge color doesn't look all that bad. I wouldn't be totally surprised if it had been regased, without replacing the getter. Just a chop, suck, and fill job, but better than nothing. I can't say it's perfect color, but certainly not dead. However, that decayed foam suggests that tube might be original. The beam (I'm being generous here!) slowly goes on and off as the very short cavity (which is not yet temperature controlled) expands, and it only lases when a single mode is near the center of the neon gain curve. The on-off behavior is probably normal due to the large FSR of the cavity, close to the width of the neon gain curve, though the low gain exaggerates the effect. I suspect that at least part of the lack of power may be due to contamination on the Brewster windows or mirrors. However, the major cause may still be an old, well used, tired tube. And even if it is partly due to contamination, this doesn't help that much - cleaning optics on this thing will be a real treat! What, contamination after 36 years? No way. :)
The power-on of the HeNe laser tube is itself interesting. Since the current regulation is via a vacuum tube, and that needs to warm up to conduct, the laser tube comes on and sputters for a few seconds, then appears to stay on dimly and gradually increases up to a normal discharge brightness. According to the Operator Manual, the current is adjustable from about 4 to 10 mA, with the normal range between 4 and 6 mA. For some reason, this one refuses to go below about 5.5 mA, and sort of doubles back. At first, I assumed it was a problem in the circuity since during the initial warmup, the tube starts out at much less than 6 mA and seems to remain on steady as the current ramps up. But, perhaps it's really flickering too fast to see. Geez, after 36 years, a bad part, no way! So, initially, I set it at 6 mA and proceeded to other checks.
Aside from the wimpy output power, the only thing I found wrong so far is that the power neon indicator lamp was burnt out, no doubt from the system being left on 24/7 for a 100 years. So I replaced that. :)
Then, figuring, "what the heck", I plugged in the other 3 cables and after actually reading the Operator Manual (what a concept?!) proceeded to go through the power up checklist, checking the meter readings for voltages, that the heater seemed to be working, and that adjusting the 10 turn Lambda pot actually changed the cavity length. All were satisfactory. So, then I switched to "Servo Null", the active stabilization mode. And, would you believe it, the thing actually locks, even with the very low output power! See Spectra-Physics 119 Laser Head with 259B Exciter - Locked. It's very twitchy as it warms up and won't stay locked for long because something is drifting, but that is truly amazing. However, I didn't wait the three hours as stated in the manual. What's interesting is that for this wimpy tube, the output power reaches its stellar value of 15 µW or so whether the tube has run or simply has been in Standby and thus at a similar temperature. So, it's not gas cleanup or something like that but simply the bore temperature. Oh, and the "Lock Alarm" lamp, a GE-334, was also burnt out, so I stuffed a GE-327 into the socket (same electrical specs, slightly larger diameter, only requiring a nano-crowbar to make it fit). Eventually, that may become an LED. Why can't designers learn to run incandescent lamps at reduced voltage?!
Later, I returned to the laser tube current peculiarity where adjustment of the current pot does not result in a monotonic change in current, but has a fold-back characteristic with hysteresis. When first powered on, it could be pulled down to about 5.5 mA before abruptly jumping to 7 or 8 mA, and then going only down to 6.5 mA or so even fully counterclockwise. After being on for awhile, that minimum increases to close to 6 mA. When turning the pot clockwise, it must go past the point where the minimum would have been, and then abruptly jumps to a high current. And, if set at close to 6 mA and powered off for awhile, when powered back on, it might not "catch" and end up at 7 or 8 mA. After trying both tubes #1 and #2 (above), I am virtually certain that this anomoly is associated with the laser tube and not the power supply. (Or, at least, is the result of the I-V characteristics of laser tube #3.) There were no problems adjusting the current on those tubes from 3.5 mA to more than 9 mA with no kinks and no hysteresis. So, suspecting that this tube has problems staying lit below about 6 mA (not exactly surprising for a high mileage tube), I powered it from my test supply, and sure enough, it wouldn't stay lit below about 5.5 mA. Perhaps the power supply does funny things when a current is dialed in that's below where the tube will stay lit, either by chance or by design to prevent continuous restarts or sputtering, which can damage both laser tubes and power supplies. Adding some ballast resistance closer to the tube anode might help some as the main 70K ohm ballast is at least 6 inches away. But there is little point since (1) 6 mA is still an acceptable current and (2) the output power will be even lower at reduced current - power continues to increase to well beyond the 9 or 10 mA maximum!
Then I tackled the drift of servo settings, which resulted in the sensitivity of photodiode output declining and the set-point changing as the system warmed up. Shortly after power-on - in fact about as soon as there's a visible beam - it was possible to lock reliably with only a few µW of output power. Only after the system had been on for awhile did locking become more problematic, even though the laser output power had increased substantially. (Well to 12 or 15 µW!) In addition, the meter didn't respond in the "Servo Null" position, though that function appeared to continue to respond. I assumed that the servo unit was full of germanium transistors and it was all too possible that one or more of them was being affected by heat.
However, after poking around with an oscilloscope, the first problem was that the adjustment of the frequency and symmetry of the multivibrator that generates the dither signal wasn't behaving as expected. I replaced the ancient 2N697s (actually silicon transistors!) with 2N3904s and that helped somewhat for no really good reason, since the specs are similar to the 2N3904. But then I noticed a rogue oscillation at around 10 kHz that appeared only after the system had warmed up. This signal was present everywhere, and even showed up across perfectly healthy filter capacitors. That didn't make any sense. There is not supposed to be any legitimate 10 kHz source and this signal was coming from somewhere other than the servo unit since grounding the input to the photodiode preamp made it go away. On a hunch, I figured that perhaps the cause was plasma oscillation in the HeNe laser tube feeding back to the power supply, or even showing up in the optical signal to the photodiode. I knew that the tube was running just barely above the dropout current, which is where such things tend to happen. And, sure enough, increasing the tube current by 0.5 mA to 6.5 mA made the 10 kHz oscillation totally disappear. Adding ballast resistance near the tube might cure this as well as increasing the dropout current, but that's for the future. And 6.5 mA is still acceptable, and now the laser can be left On or in Standby indefinitely with no noticeable drift. In fact, even with the miniscule amount of laser output power, it's now possible to adjust the electronics for normal meter deflection when adjusting the Lambda pot with the servo unit in the Null position.
So, aside from the wimpy output power, there appears to be nothing wrong with the entire system.
A couple years later, I obtained another intact SP-119 laser head, also with a tube having a "white cloud of death" getter. (Call this #4.) It started out with a sickly purple discharge and of course no output, and I fully expected it not to lase at all. But after an hour or so of running at 5 mA with the complexion of the discharge steadily improving, I practically fell over when a steady stream of coherent red photons began appearing. :) After a few more hours, it's up to 90 µW (when locked) and still climbing. And the serial numbers of this laser head and exciter are lower than for the others, so the system is probably even older. Unfortunately, the label on the tube is hidden underneath, so I can't see its serial number. However, date codes on electrolytic capacitors in the exciter show them to be from mid-1970. So this laser - likely around 40 years old - is operating with an output power almost 30 percent greater than the SP spec of 70 µW. The output power is considerably higher at the maximum recommended tube current of 6 mA, but I'd rather run at a more conservative 5 mA and extend tube life.
The only other circuitry on the main PCB, mostly hidden under the 259-002 on the right, is the all solid state laser head heater controller.
The larger transformer is for all the high voltages and vacuum tube filaments, while the smaller one is for the heaters and low level servo circuits. In Standby mode, only the latter is powered.
Aside from the pots for HeNe laser tube current and Standy heater power accessible from the front panel, the only other electronic adjustments in the entire system are two trim-pots visible at the bottom right corner (dither frequency and symmetry) and the one labeled "Inc" (Servo gain), a 10 or 20 turn trim-pot accessible through a hole in the 259-002 cover.
Unresolved issue:
This relates to the second SP-119 laser head (with tube #3) and the SP-259B exciter as described and shown above:
If anyone has another SP-119 laser head and/or controller gathering dust that they'd like to contribute to the cause, or other information in this antique laser, please contact me via the Sci.Electronics.Repair FAQ Email Links Page. Of particular interest are additional SP-119 laser heads, beam expanders (as these seem to be scarse), as well a an extra 259-002 servo unit.
DC OUT - HeNe laser tube and AC interlock:
This is a large circular bayonet-lock connector with 7 pins:
Pin Function Comments
----------------------------------------------------------------------------
1 Interlock Pins 1 and 2 are jumpered in cable, and are in
2 Interlock series with main power.
3 Heater Return
4 Spare? Second black pin jack in laser head, no connection.
5 Laser Tube- First black pin jack in laser head for tube cathode.
WARNING: As much as -5,000 V when starting!
6 Laser Tube+ Red pin jack in laser head for tube anode, via 70K
ohm ballast resistor from cable. Tube current may
be adjusted from about 3.5 to 9 mA (same reading on
meter) via recessed pot below "On" switch marking.
7 Heater Drive From temperature regulator for 18 ohm laser tube
heater jacket.
J101 - PZT/Standby Heater:
This is a small circular screw-lock connector with 6 pins.
Pin Function Comments
----------------------------------------------------------------------------
1 PZT Shield/Return
2 Standby Heater Bore heater on during Standby. Heater is
3 Standby Heater 300 ohms between pins 2 and 3. Adjustable
from 32 to 42 VAC RMS (65 to 90 on meter) via
recessed pot below "Standby" switch marking.
4 External Oven Null Measured +16 V (Meter is 20 V full scale).
5 Meter Return
6 PZT drive +10 to +215 V via 10 turn Lambda pot
with Servo set to "Off" (or manual).
J106 - Thermistor:
This is a small circular screw-lock connector with 3 pins.
Pin Function Comments ------------------------------------------------------ 1 Shield 2 Thermistor 10K ohms between pins 2 and 3 at 3 Thermistor room temperature.
J201 - Photodiode:
This is a small circular screw-lock connector with 3 pins. The photodiode mounted behind the HR mmirror and cable is only present on the laser head if the 259-002 Servo Option is installed.
Pin Function ---------------------------- 1 Shield 2 Photodiode Anode 3 Photodiode Cathode
The 1-B tube I would use is the Melles Griot 05-LHB-270, which has a narrow bore and is only 222 mm in length (just under 9 inches). With an OC mirror mounted on a piezo beeper, the total cavity length would still be only about 9 inches, similar to an SP-088. A photodiode behind a small aperture (to block bore light) would be mounted behind the HR, well insulated from the anode voltage!
I have already done experiments with a similar setup as shown in One-Brewster HeNe Laser Tube with External OC Mirror on PZT and it does have a very nicely shaped output power versus mode sweep as shown in Effect of Mirror Alignment on Scanning Cavity HeNe Total Power Display. This set of photos was taken for another purpose, but they do clearly show the very distinct valley, also similar to that of the 088 tube. One uncertainly is what the response of the piezo beeper will be at the 5 kHz dither frequency of the SP-259B. However, it is also 5 to 10 times more sensitive than the SP-119 PZT, so a simple filter network may be able to compensate peculiarities in its response. At the very least, the DC sensitivity will need to be reduced.
A separate HeNe laser power supply might be required as the 05-LHB-270 requires considerably more operating and starting voltage than the SP-119 tube. (The latter is probably what would really be the problem.) To keep the internal HeNe laser power supply happy, a short tube or dummy load could be connected, or the vacuum tubes could simply be removed. :)
There should be no problems with the photodiode, but if the gain adjustment on the SP-259B servo unit doesn't have enough range (because the power in the waste beam may be higher than allowed for), a neutral density filter or other means can be added to reduce it.
Note that with a laser based on this length 1-B tube, a pure single frequency output will not likely be possible as two weak modes will probably lase on the tails of the neon gain curve. Since all the modes have the same polarization, there is no way to suppress these. However, if a one-perpendicular window (1-W) tube were used instead (very rare), then the two weak modes would have the orthogonal polarization, and a simple polarizing filter would eliminated them. The closest modes with the same polarization would be around 1.5 GHz away and would have no chance of lasing. It might even be possible to use a somewhat longer cavity and still achieve single frequency operation with this setup.
An alternative to the 1-B or 1-W tube would be to use just the glass tube from a Hewlett-Packard 5501A (without the magnet and optics). This has a relatively short cavity with an internal PZT. The 5501A tube does appear to have a mode shape with a Lamb dip, though I don't know for sure if that's the cause. However, to use a 5501A will require a HV amplifier for the PZT as it needs about 1.5 kV to go through more than two FSRs, and a beam sampler at the output of the tube since the waste beam is blocked by the PZT. And, it's a total joy to remove the glass tube from the magnet assembly! The tube I tested also had a peculiar mode flipping behavior whereby it tended to be polarized in one plane on the forward stroke of the PZT, and the orthogonal plane on the reverse stroke of the PZT, even across multiple FSRs. However, a relatively weak external magnetic field had an effect, and with care placement and orientation of a weak magnet, it could be convinced to act normally. The Lamb dip can be clearly seen in Modes of HP-5501A HeNe Laser Tube 1 With No Magnetic Field along with the mode flipping anomaly, as well as some hysteresis and non-linearity in the PZT response. The flipping quirk wouldn't matter as far as Lamb dip locking is concerned since only the output power is used, but the actual beam polarization once locked might depend on, well, the flip of a coin. :)
Here are some photos (courtesy of Bob Hess):
On the far side of the head enclosure, a blue HeNe laser power supply brick and PCB with a single trimpot are barely visible. I'll have to ask the laser head to turn around for another photo. :)
I was able to contact the original designers (Walter Luhs and Dieter Frolich) of the ZL-150. Here are some comments from Dieter Frolich:
At the time when the ZL-150 was developed (around 1984), I was the owner of a small optoelectronics company that designed most of the ZL-150 optics and mechanics and designed and manufactured the entire electronics. I have no documentation any more, but my memory is still quite OK, so let me make a few remarks:
- The ZL-150 was based on an article by Hall et. al.: "T. Baer, F. V. Kowalski, and J. L. Hall, "Frequency Stabilization of a 0.633-micro;m He-Ne Longitudinal Zeeman Laser," Appl. Opt. 19, 3173-3177 (1980)". (This is the same Hall from Pound-Drever-Hall locking fame. --- Sam.)
- The electromagnetic actuator not only looks like something out of a washing machine, but basically IS out of a washing machine: a standard component used for electrically operated valves. It was used to dither the tube length at ~10 Hz square wave and for the fast branch (P and D) of the control loop. The actuator only pushed, so we applied a DC offset to get both directions.
- If I remember correctly, the laser indeed locked to a minimum of the beat frequency - in this particular arrangement a narrow dip is superimposed on the broad Zeeman frequency peak - I think that was one of the specialties in Hall's paper.
- Indeed, the stabilized frequency was almost independent of anything, including tube aging. The residual modulation of the optical frequency was in the range of 10 MHz.
- The electronics was not complex. As a matter of fact, I used one of the first single chip microprocessors to do all the jobs plus some power transistors, of course.
- There is not only a bar magnet. The entire metal tube surrounding the laser tube over part of its length is a permanent magnet. I don't remember why S&&H had to use an additional bar magnet.
More to come.
(Mostly from: Skywise (into@oblivion.nothing.com).)
This is a Teletrac 1 mW stabilized HeNe laser with built in interferometer receiver. Going to Teletrac, Inc. redirects to Axsys Technologies, which only has information in their quarterly earnings reports referencing the sale of the company. But I found a user manual for a later model, but similar laser at Teletrac Stabilized Single Frequency Long HeNe Laser or Sam's Backup of Teletrac Stabilized Single Frequency Long HeNe Laser. (This manual appears to be for another Teletrac laser which may be similar to the ones described in the sections: Teletrac Model 150 Stabilized HeNe Laser 2 and Teletrac Model 150 Stabilized HeNe Laser 3. The general information and theory of operation should be similar though.
The electronics for the receiver are totally independent of the rest of the laser and are powered through its connector.
The HeNe laser tube itself has no markings. It's about 8 inches mirror to mirror. According to the user manual I found on-line it's manufactured by Zygo. (Tubes in some other Teletrac lasers are made by Zygo but this one is Melles Griot. --- Sam.)
The output of the OC-end goes through a collimator to get a 1 cm low divergence beam. And it is LOW divergence. I once shot this thing out my window to a brick wall about 1/4 mile away, took a walk and found the beam to have barely grown, if at all.
The HR-end has what is obviously a mode detection assembly, but it's all covered in shrink tubing.
A two terminal device (probably an LM335) is glued face down onto the glass of the tube near the cathode end for temperature sensing.
There are two low wattage filament lamps under the tube for heating.
Unlike most other Teletrac/Axsys lasers, the power input for this is 115 VAC, not 12 VDC. The HeNe laser power supply is a brick made by Power Technology, Inc. However, it's definitely non-standard as the 12 VDC to power the stabiliization electronics is provided by a pair of extra terminals on the brick!
Temperature regulation is done by two fan blades that vibrate, driven by a piezo. The vent is on the bottom of the laser so I have to make sure the 'tail' is sticking out in free space or it overheats and the fan blades really start clattering. The lamps are not used once the laser has reached the set-point temperature and is locked.
From a cold start the laser reaches mode lock in about 11 minutes.
The receiver electronics are dirt simple. Just 3 good op-amps (2 LM6361N and 1 LM353). Everything else is just caps, resistors, and two trim-pots. The board has space for two other 16 pin ICs but the spots are empty with no labeling to infer their function. It looks like the outputs are all analog. On the board the wires going to the detectors are labeled SIN, COS, and INT. I think the SIN and COS imply quadrature output, but have no clue what the INT is. That signal goes to the chip that got really hot. The other two signals go to the other op-amps, and I'm seeing signal there on their two test points.
Here's page with 31 photos and 1 Quicktime movie: It's under the reference section of my Lasers Page but here's a direct link: Teletrac Interferometer Laser.
(From: Sam.)
The HeNe laser tube is from Melles Griot, regardless of what that manual says. It may be a 05-LHR-120 or similar tube, possibly selected to for specific characteristics to optimize it for use in this application. Some older Teletrac lasers like the ones described later in this chapter did use Zygo tubes but not this one.
SIN and COS are the quadrature outputs from the optical receiver. INT is the "intensity" which is proportional to the total output and would be used to compensate for variations in optical power due to tube aging and/or interferometer alignment and losses.
The LED on the back of the laser that changes from red to green as the modes cycle during warmup and then goes out when locked is a nice touch and is present on all subsequent Teletrac (and Axsys) stabilized lasers.
I'm impressed with how simple and clever this system is, though some might describe it in another way - a kludge. :-)
There is more on the likely way this (and other Teletrac/Axsys) lasers are used in the section: Teletrac Model 150 Stabilized HeNe Laser 3. That laser uses an external interferometer and optical receiver which are implemented in the same way. (I've also seen a nearly identical laser to this one without the optical receiver.)
Several photos of this Teletrac 150-IV laser can be found in the Laser Equipment Gallery (Version 3.11 or higher) under "Teletrac/Axsys HeNe Lasers".
It has the same Melles Griot tube as the Teletrac laser (above). The actual model number is 05-LHR-219-106. Apparently the combination of "219" and "106" is custom made for Teletrac/Axsys tested for mode flip behavior and selected for high output power. In fact, the output power of one I measured exceeds 4 mW (!!) but the output power of the laser is only a bit over 1 mW. Even this wimpy output power violates the safety sticker maximum rating of less than 1 mW! There is a polarizing filter at the input to the beam expanding telescope at an angle of +45 degrees looking toward the output of the laser, and a Quarter WavePlate (QWP) at its output. Thus, the actual beam from an unmodified Teletrac laser is a single mode that is circularly polarized. So this can go directly into the interferometer optics without worrying about orientation as the polarizing beam-splitter will separate it into two linearly polarized (REF and MEAS) as required. I was surprised that the polarizer was a cheap filter and not a polarizing beam-splitter cube, which would be of much higher optical quality and have lower losses, resulting in much greater output power. In fact, initially, because the the output had no obvious polarization axes due to the QWP, I thought the polarizer was simply a neutral density filter to cut down on the output power to satisfy the safety rating - and then to enable the output power to be easily readjusted upward (at great expense to the owner!) as the tube aged. :) The latter I found out quickly to be bogus since removing that polarizing filter isn't fun. Without the polarizer, it produces a beam with right and left circularly polarized modes 687 MHz apart (the longitudinal mode spacing of the 05-LHR-219 tube). Or, by removing the QWP, one or two linearly polarized modes depending on whether the polarizer is present. In the latter case, the total output power would be almost 4 mW. Thus, regardless of the original intended application, it could be set up as a nice general purpose stabilized HeNe laser.
The output is actually left-circularly polarized according to a test report that I have for one of these lasers, not that the handedness should matter all that much (though I assume it would affect the sign of each measurement axis). The purity may also not be that high as that particular laser - which appeared to be new or very low miles - had a 20 to 25 percent variation in intensity when rotating a polarizer in the beam. It still had the stickers/seals intact and polarization behavior is not the sort of thing that can change with age or use.
The stabilization system uses a conventional Minco thin film heater wrapped around the tube, rather than the light bulbs and piezo fans. :) The control algorithm is implemented digitally with a PIC, quad digital pot chip, and some other stuff. :) A serial EEPROM/NVRAM stores the calibration information unique to each laser. Unfortunately, this basically means there is no easy way of making adjustments that may be required as the tube ages, or if it is replaced. Once the output power declines by some arbitrary amount, the algorithm will abort with a flashing red failure LED. If the output power in either mode from the waste beam of a replacement tube doesn't match the original, the same thing will happen, or it won't detect mode sweep cycles at all. And it may be that if the heater resistance isn't the same, the warmup period will be too long or too short, or it may even think the tube is at operating temperature when it first starts! And troubleshooting and repair of the Control PCB with no accessible (analog) signals and all its SMT components would not be fun. I really can't imagine that the possible flexibility of the digital control scheme has any functional benefits. And, in fact, digital control may not work as well as a garbage LM358 op-amp implementation. But it probably does enhance the job security of the designers!
After power-on, the controller appears to first check that the mode or modes from the polarizing beam sampler at the HR-end of the laser tube are present and of adequate power. Only then does it turn on the heater at full power (approximately 10 V across a 5 ohm resistance or about 20 W). And based on one sample that refused to do so quickly, the threshold (presumably contained in the NVRAM) must be set within 15 to 20 percent of the tube's power when new. That laser refused to turn on the heater until a couple minutes after a cold-start, and this was long enough that it gave up and flashed an error code. Then, power cycling would usually enable it to start up successfully within a few seconds after that.
Once the heater is powered, warmup is quite rapid, with locking occuring in around 5 minutes. It probably uses the resistance of the heater as a temperature sensor, switching to feedback control once it has increased enough to exceed a stored reference value.
Once locked, the short term stability is quite good, but there is a slow periodic variation in locked output power of perhaps 10 percent p-p. This settles out in several hours once the laser has reached thermal equilibrium. I suspect the cause is insufficient or lack of wedge in the HR mirror and/or lack of AR coating on the HR mirror. This results in an etalon effect, causing variations in both waste beam and main beam power, both intrinsic to the laser tube, and amplified through the feedback since the power of the waste and main beam are no longer in a fixed ratio. And the relative power of the two modes would also vary slightly.
I was hoping to repair the Control PCB rather than simply salvage the almost new tube for use in some other stabilized laser like a Coherent 200 (which appears to use the same tube, or one close enough). I suspected that the AD8304 quad digital pot chip was bad as the "red" mode input is stuck high. Of course, it could have been something else like bogus data read from the serial NVRAM. The HeNe laser power supply was also dead, and I suspect that my testing with a substitute power supply is what actually damaged the Control PCB, though I'm not sure how. However, an arc from the anode of the HeNe laser tube to the red mode photodiode could conceivably have been the cause. Really? :)
CAUTION 1: Be extremely careful around the anode area of the HeNe laser tube, especially if there is a need to remove the heat-shrink insulation. While contact with the HV probably won't be lethal to you (just a shock and the smell of burning flesh), it is very likely to jump from you to one of the conveniently located cables in the vicinity and kill the controller. It's possible that arcing to the chssis (as with a hard-to-start tube) might even do this. Almost everything on the controller PCB is surface mount which along with the PIC (Microchip PIC16C73A-20/SP) and its serial NVRAM (Xicor X24C44P), makes troubleshooting virtually impossible. I've managed to screw up the controllers on two separate lasers! :( The symptoms then seem to be that the mode LED remains red and no longer responds correctly, with a significant offset and difference in gain (or more), the heater never turns on (even though its LED says its on), and the firmware gives up after a few seconds and flashes the right-hand green LED forever. I suspect that at least part of what's really happening is that the perhaps the NVRAM has gotten erased since removing the NVRAM from its socket results in no noticeable difference in behavior. Or, possibly there's nothing wrong with the NVRAM, but the PIC is unable to communicate with it and/or the quad digital pot chip that connects to the PD inputs (among other things). So key parameters never get initialized correctly. I did reaplace the quad digital pot chip with no change in behavior, but it's quite possible I screwed up the SMT soldering! The PIC remains slive though. And with enough light into the beam sampler to get the modes to switch back and forth from red to green repetitively, the onset of it giving up can be delayed indefinitely. So it's still looking at the inputs, for whatever good that does!
CAUTION 2: When using the X-Y adjustment screws to center the output from the HeNe laser tube in the beam expander, DO NOT use a tool, only finger rotation. They are made of metal and press against the glass of the laser tube, separated from it only by the not very compliant Minco heater. Too much force WILL crack the tube. I also found this out the hard way - after realigning the mirrors on a non-lasing tube to like-new specifications. :(
I did eventually try replacing the AD8304 (quad digital pot) on a bad controller PCB with no obvious change in behavior. My surface mount rework skills are somwhat lacking, but I don't think that was the problem since at the very least, behavior before and after were identical. I later found out that the X24C44P (NVRAM) was dead (or erased). In fact, I now have two of these lasers with bad controllers, due to zapping. One controller PCB seems to be fully functional except for a bad or erased NVRAM. It works normally with a known good NVRAM except that the switchover threshold to feedback control appears to be several degrees higher then another known good controller PCB with the same PIC and NVRAM installed, possibly simply due to normal component tolerances. (The NVRAM in each controller would be programmed for its specific tube, so swapping controllers might not be an acceptable repair technique!) The other seems completely dead even with a known good PIC and NVRAM, except for turning on the red status LED and the heater!), but the mode LED never comes on and the laser never locks.
Replacing tubes in this is laser is problematic due to the digital controller. The NVRAM apparently also stores parameters for the feedback probably including the mode amplitudes and offsets for both photodiodes. Thus installing a different tube results in a change in all of these. With no pots to twiddle :), the controller is likely to be unhappy, resulting in one improper or no state changes during warmup, and inability to lock once it reaches operating temperature. I have one such laser where every test I perform indicates that both the controller and beam sampler/photodiode assembly is working correctly. They just won't play together with the new tube. The original tube was too weak to guarantee it would even reach the threshold to turn on the heater, so I can't go back and test with that. Thus, I am contemplating the construction of an photodiode preamp adapter board with gain and offset adjustments to determine if this is indeed the problem. One complication is that the common for the photodiodes is ground and they feed negative current into the inputs. Thus, any active circuit would need a negative power supply to be added. A quick test using a pair of 100K ohm trim-pots was able to get the mode LEDs to respond weakly and the laser locked. But a proper circuit would be the only reliable solution.
If someone has more information, a working laser whose X24C44P NVRAM they would be able to copy or a dead (or alive!) laser like this they would be willing to offer to the cause, please contact me via the Sci.Electronics.Repair FAQ Email Links Page.
Pin Function ------------------ 1 Return 2 Tube Cathode 3 +12 VDC
Heater:
Pin Function ---------------------- 1 HTR- 2 Sense-? 3 Sense+? 4 HTR+ (+12 VDC)
Photodiode:
Pin Function ---------------- 1 PD1 2 Common 3 PD2
Several photos of this Teletrac 150 laser can be found in the Laser Equipment Gallery (Version 3.11 or higher) under "Teletrac/Axsys HeNe Lasers".
An operation and service manual for what appears to be this Teletrac laser can be found at Teletrac Stabilized Single Frequency Long HeNe Laser or Sam's Backup of Teletrac Stabilized Single Frequency Long HeNe Laser. It even includes schematics! Interestingly, there is information at the end of the manual on interpreting or specifying Teletrac laser part numbers, but these never seem to show up on the lasers! :)
It isn't quite identical though as the output of the laser in the manual is linearly polarized at 45 degrees, while this one is circularly polarized. However, it is probably close enough for government work. :) Here are the specifications from the manual:
Operation
Non-operating
Plasma tube
Test Connector
Pin Function --------------------------------------------------------------------- 1 GND 2 Mode (1.4 to 11.0 Vdc, lock point is 6.2 +/-0.2 VDC) 3 Tube Temperature (Grounding this pin will force heating) 4 Laser READY (External pullup required to +24 VDC maximum) 5 Temperature Set-Point (Equals Tube Temperature at switchover) 6 Servo Drive (1.5 to 10 VDC). CAUTION: Probing these test-points may result in a temporary out of lock condition.
With the separate temperature sensor and likely different heater resistance and thermal response, the Axsys and Teletrac controllers are not interchangeable, and fortunately for those wanting to try, the connectors differ sufficiently to make it difficult (though not impossible!) to do something bad in the process.
Naturally, since this version is emminently repairable, there would be nothing seriously wrong with the sample I acquired. It only had a smashed on/off switch and polarizer with excessive scatter! The tube is like new - instant start, stable run, locks in 15 minutes or so, and well aligned producing over 3.2 mW total out the beam expander in both circularly polarized modes since I never actually replaced the polarizing filter.
That lock time of 15 minutes is somewhat longer than for the Axsys equipvalent, probably due to the fact that the thin film heater occupies a very small portion of the Zygo tube (less than 2 inches) compared to most of its length for Melles Griot tube. However, I bet the life expectancy of the Zygo tube, typically 50,000 hours, is more than double that of the one from Melles Griot.
Why do manufacturers redesign a perfectly functional easy to manufacture low cost PCB for no obvious reason other than to make it more proprietary? This is the third example in this chapter alone (Agilent and Zygo being the others I've seen so far). There's no evidence that the digital controller has any benefits in terms of specifications since they haven't changed. It would be hard to believe that it is cheaper to manufacture or test. But there is no doubt that it is more difficult or impossible for anyone other than the original manufacture to repair or adjust!
Several photos of this Teletrac 150 laser can be found in the Laser Equipment Gallery (Version 3.11 or higher) under "Teletrac/Axsys HeNe Lasers".
An operation and service manual for a similar laser can be found at Teletrac Stabilized Single Frequency Long HeNe Laser or Sam's Backup of Teletrac Stabilized Single Frequency Long HeNe Laser. However this laser is probably the "Short" version since (1), it IS short, and (2) the spec'd output power has to be much lower due to the much smaller HeNe laser tube it uses.
The beam exits out a large hole in the side of the case via a 45 degree mirror on an adjustable mount. However, the front plate is clearly original but has no aperture, so this was almost certainly done by Teletrac, not and end user. And "Beam Exit: RT" is an option listed in the manual! The approximate arrangement of components is shown in Teletrac 150 Laser and Optics. From left to right: Teletrac 150 laser (version 3) with right angle output, optical receiver, linear interferometer optics, and retroreflector on rotary mount. The 4 holes in the top of the optical receiver are for the four (4) adjustment pots. The linear interferometer is basically a miniature version of the HP/Zygo units. The retroreflector is a cut-off cube-corner RTV'd into a bracket that clamps to a ball bearing shaft. See Teletrac Retroreflector on Rotary Mount for a closeup. So, the total travel would have only been a few cm.
The principle of operation for a positioning system using this laser would be similar to that of one using the HP/Agilent or Zygo lasers found elsewhere in this chapter. However, it is what's known as a "homodyne" system since it uses baseband fringes, rather than the "heterodyne" system using the difference (split) frequency of the two-frequency laser. The general approach is shown in Interferometer Using Single Frequency HeNe Laser. The linear interferometer is placed between the laser and the retroreflector on the remote "tool". The optical receiver goes between the laser and the interfer
