This chapter deals with the more practical aspects of HeNe laser power supply design including circuits for providing the HeNe tube operating voltage (AC line and inverter types), starters, regulators, and modulators. There are many options for each subsystem and it is often possible to mix and match as desired!
These may run off of low voltage DC or the AC line but in the latter case convert the AC into DC first and then use a high frequency chopper and small transformer to generate their output.
Modulation inputs may also be provided to permit the transmission of audio or data over the HeNe beam to enable external closed loop control of beam power.
Some more sophisticated commercial power supply designs provide a variety of 'soft start' and other features to maximize HeNe tube life. Others enable 'instant start' for applications where the HeNe tube must be switched on and off frequently. These sorts of advanced forms of regulation are not really needed for general applications - which is just as well since the circuits tend to be proprietary and not available. Some of it may have been Marketing Departement driven specsmanship - how else to distinguish YOUR HeNe laser power supply from everyone else's virtually identical units? :-)
Note: Throughout this document, we use 115 VAC as the nominal line voltage in the U.S. However, the actual measured voltage may range from about 105 to 125 VAC and still be considered to be within acceptable limits by the utility company. For this single-phase system, using both Hot legs of the line will then result in a nominal 230 VAC which may actually range from about 210 to 250 VAC.
The transformer output generally feeds a half wave rectifier or 2 diode 2 capacitor doubler and filter capacitor stack.
Either a parasitic voltage multiplier or pulse (trigger) type starting circuit can be used with these designs.
Compared to inverter type power supplies, line operated units are easier to construct (no custom transformer is needed) and troubleshoot (there are no transistors to blow by the bucketload). Of course, they are not nearly as portable in two ways: the power transformers you are likely to find are usually quite heavy and there is that annoying line cord to drag around!
However, most of the components are readily available or can be constructed from common parts including the high voltage diodes and capacitors:
The only problem may be the power transformer which is typically 600 to 1,200 VRMS at 20 mA or so:
CAUTION: Do not be tempted to increase the high voltage output of a power transformer by more than 30 percent or so above it rated value (by either driving its primary with a higher than rated voltage or by adding booster windings in series with the primary). Even this may be excessive depending on its design margins. At some point core saturation will result in a dramatic increase in input current, overheating, meltdown, smoke, 6 foot flames, etc.
In addition, the insulation ratings may be inadequate for the increased high voltages now produced by the secondary.
Thus using a 115 V transformer on 230 VAC to obtain double the output is probably not a good idea though I know people who have done this and lived to tell!
For example, using two 380 VRMS transformers in series will result in over 2,000 VDC without playing games and 2,200 to 2,500 V with one of the booster techniques described above.
These are suitable for intermediate size HeNe tubes - 5 to 20 mW.
What this means in the end is that although the open circuit voltage (full wave rectified and filtered) may approach 7 kV (for a 10 kV unit), voltage is down about 30 percent at 6 mA. This is no problem if used on a Variac with current monitoring though but the practical upper limit on operating voltage is only about 4 kV.
Although, as noted, the current limited behavior is not linear, it can be approximated by assuming an internal current limiting resistor and an ideal transformer. For a 10 kVAC RMS, 20 mA transformer, the internal equivalent series resistance would be about 500K ohms across the entire winding or 250K ohms per side (center tapped). When used for the intended application - be it producing an arc to ignite an oil burner flame or powering a neon sign (using a luminous tube transformer which is basically similar), no ballast resistor is needed to protect the load or transformer. Unfortunately, this doesn't help us since the rectifiers and filter capacitors come between the transformer and HeNe tube :-(.
Also, because the open circuit voltage is much higher than the actual operating voltage, there will be a current spike through the HeNe tube at the instant it starts of many times the normal operating current. However, for a typical design, the total energy in this pulse is not very large even at the upper limit of the power supply. For example, with a total filter capacitance of .25 uF, an open circuit voltage of 7 kV, and an operating voltage of 4 kV, the energy delivered to the HeNe tube by this pulse will be only about 2 to 4 w-s (J). I do not know if this will result in significantly shortened tube life in the long run under normal usage (reasonable number of starting cycles).
See the section: Sam's Mid-Size Line Powered HeNe Laser Power Supply (SG-HL2) for a sample design using an oil burner ignition transformer.
The characteristics of these neon sign transformers are similar to those of oil burner ignition transformers (see above) but typically have higher voltage and current ratings, and are thus much larger and heavier. They also may have a more constant current characteristic - delivering nearly their rated current up to a good percentage of their no load output voltage (to handle a variable length of neon tubing). For example, a 12 kV, 30 mA unit may behave like an equivalent 30 kV source in series with a 1M ohm resistor for loads resulting in an output voltage of up to 6 or 8 kV.
If you have the option of obtaining a slightly higher voltage transformer than you actually need, go for it. However, the key word here is 'slightly' - not something 10 times too big! Then, if you acquire a higher power tube in the future, you will be all set. For now, it will just require a larger ballast resistor or Variac to run at reduced input voltage.
It might be worth trying a TV or audio equipment repair shop - they may have spare transformers from old tube sets laying around gathering dust. These are ideal and can probably be had for next to nothing.
Another option is an electronics surplus supplier - I have seen suitable transformers at some of these in the past but don't know what is currently available.
A 3 or 4 stage voltage multiplier could be used to boost the output of a lower voltage transformer if a suitable high voltage transformer cannot be located. However, to obtain the needed current, the capacitors would need to be quite large - perhaps 1 uF at 1,000 V or more. Also, you would then probably need to use a pulse type starting circuit as a multiplier type starting circuit may not be able to provide enough output with a reasonable number of stages since the available p-p input voltage will be less with this approach.
I have recently been using the power transformer from a long dead tube type TV both for testing a higher power commercial HeNe supply board and as the basis for a power supply of my own. See the section: Sam's Small Line Powered HeNe Laser Power Supply (SG-HL1). There was a selector for line voltage adjustment built into the transformer. With this set for lowest line voltage (and thus highest output) and the filament windings connected out of phase, it produces over 900 VRMS at 115 VAC input and over 1,150 VRMS using a Variac that goes up to 140 VAC. This translates into a doubled DC voltage of between 2,500 and 3,000 VDC - more than ample for most HeNe tubes up to 10 mW.
_ F1 S1 T1 or T100 Hot o------- _---------/ -------+------+ +--------o X 1 A Power | | ||( R0 / +---+ ||( 47K \ )||( / )||( | Primary )||( HV Secondary IL1 +|+ )||( NE2H |o| )||( Power On |o| )||( +|+ +---+ ||( | | ||( Neutral o---------------------------+------+ | +--------o T | Ground o----------------------------------------+------o Tube- (HV Return) _|_ -Note that the fuse is shown as the first component after the line cord. This provides the most protection where the fuse is located on the panel next to the cord entrance. However, it may be more appropriate to put the power switch first if the fuse is located on a circuit board or other distant location. Both arrangements are common in commercial equipment.
Important: Use a grounded (3 wire) line cord and connect earth ground to the case (if it is made of metal), transformer core, and high voltage return of the tube (Tube- on the schematics below). This will assure that the tube housing is grounded and that no fault (like a short inside the power transformer) will result in any user accessible parts becoming electrically live as long as the line cord is plugged into a properly grounded outlet. The alternative is to double insulate everything but this may be impossible if you are using a commercial laser head where the tube cathode is already connected to its metal shell.
CAUTION: This is probably over 700 V with significant current available. Take care. Make a note of your reading and then disconnect power.
Alternatively, you can probably safely achieve up to a 25 or 30 percent boost using a separate low voltage power transformer to provide your booster winding. (Start with step (2).
For example, with a 24 V transformer, a 26 percent increase in output voltage will result - this is probably about the limit before you risk core saturation with a typical transformer but your mileage may vary.
CAUTION: On transformers with dual primary windings (to support 115 or 230 VAC power), it is possibly in principle to use one of these to drive the supply and the other as a booster on the secondary side. I Do not recommend this approach as the insulation between the two primary windings may be inadequate.
Several types are possible:
However, self oscillating designs are generally not as efficient as driven ones (see below) and may be unstable under certain load conditions.
With any of these, the starting circuit can be separate (a voltage multiplier or pulse type) or built in as part of a high compliance design.
For DIY projects, it is best to run the inverters from low voltage DC. While it is also possible to build inverters that operate directly from the power line (commercial power supplies often do this) with just rectification and filtering, I DO NOT recommend this as an option here for two reasons:
For an example of (3), HeNe laser drive circuitry is briefly covered in a Linear Technology Corp. application note: AN-49, p.13. This is a low voltage DC powered circuit using an LT1170 chip for fully automatic starting and feedback control of operating current. It is essentially a constant current supply with a voltage compliance range of 10 kV. HeNe tube power requirements are also discussed. Unfortunately, the special transformer may not be readily obtained.
Although there is somewhat less of a shock hazard with an inverter running from low voltage DC, grounding the metal case of a laser head and other metal parts is still desirable unless they are totally isolated from user contact (e.g., everything is in a plastic enclosure).
The main difficulty from the hobbyist's perspective in building an inverter type power supply from scratch may be in obtaining or constructing the required high frequency ferrite transformer since these aren't the sort of thing that can be purchased at Radio Shack. However, I have successfully wound my own transformer (from a bare core and bobbin) to repair a commercial power supply (see the section: HeNe Laser Power Supply from HeNe Laser Pointer (IC-HI3)).
Another option that may or may not work (I have not tried this) is to use a fluorescent backlight inverter (for/from a laptop or other LCD) or even a battery powered lantern as the basis of a HeNe laser power supply. Since there generate up to 1,000 VAC or more with a few mA available at 10s to 100s of kHz, the addition of a rectifier or doubler, and starting multiplier may be all that is needed. However, some commercial designs are too smart for their own good (at least for this application) and may shut down if the exact conditions they expect are not met.
However, with modern implementations of switchmode power supplies utilizing digital control techniques, catastrophic failure due to external faults or user abuse should be a thing of the past. With traditional analog control, pulse width modulator ICs, op-amps, and discrete components are used for the drive of the switching transistors or MOSFETs, and for fault detection. These schemes are often ad-hoc and testing for all possible fault conditions is not really possible. But with digital control, a microprocessor implements the feedback equations in firmware and generates (or at least directly supervises) the switchmode drive signals, and monitors for over-current, over-voltage, arc-faults, and other error conditions on a cycle-by-cycle basis. It can then instantly shut down the supply long before any damage to it or the laser can take place. Explosive decomposition of a potted module due to accdidentally powering up with a shorting strap still attached would be a thing of the past. They would also be able to recognize that a laser wasn't starting and modify the output voltage appropriately. For example, by cycling it between 0 and maximum rapidly to take advantage of the dV/dt of the voltage to help ionize the gas. It could also detect a laser that was having trouble staying lit at the selected current and modify the dynamic ballast resistance as required, or shut down if that didn't help.
By now, the "Smart HeNe Laser Power Supply" should be standard practice in the industry but old habits take time to die. :)
eBay and eBay Stores always have a wide selection of high voltage components, some at attractive prices both individually and in large quantities. And searching is easy and quick. However, keep in mind that many of these parts are imports (China, Russia, and elsewhere) so quality cannot be guaranteed and may vary from batch to batch, and even if there is a warranty, returns may not be worth the hassle and postage.
More on specific types of components below.
These types of devices are generally much more expensive on per-kV basis compared to parts like the 1N4007 as well. ECG, NTE, and SK also have a variety of replacements of various ratings (probably even more pricey). See the section: Mail Order - Electronic Components and the document: Troubleshooting of Consumer Electronic Equipment for contact info.
(Note: As of January 2001, NTE has purchased the assets of ECG so the ECG parts listed below may no longer be available, use the NTE or SK equivalent which usually has the same number with the 'NTE' or 'SK' prefix.)
However, there are alternatives to ordering special high voltage rectifiers from electronics distributors or surplus sources, or even ripping apart the family microwave or TV to build your laser power supply (though salvage from these sources - after they are certifiably dead and thrown out! - is well worth the effort). Other sources include: electronic air cleaners, igniters, bug zappers, and photocopier and laser printer HV power supplies. Modules from these devices with nice HV components may be available surplus at ridiculously low prices.
A series string of dirt cheap 1N4007s can be used to construct high voltage rectifiers for line frequency power supplies and seems to be acceptable up to a few kHz for inverter based power supplies. Equalizing components do not appear to be needed - at least where these diodes come from the same lot number. Modern devices are matched closely in terms of leakage current and capacitance. However, the construction DOES need to take into account requirements for high voltage insulation - adequate spacing and the rounding off and possibly coating of exposed wires/connections especially where the very high starting voltages are involved. A variety of mounting techniques can be used including soldering end-to-end installed in a plastic or glass tube, and on or between pieces of perfboard.
The common 1N4007s cost only a few cents each in quantity (typically between $.01 and $.06, though much higher from Radio Shack). Thus, it is generally possible to construct high voltage rectifiers at significantly lower cost than buying microwave oven, TV, or similar commercial types. And, it is trivial to construct whatever size you need! I recommend derating by 30 to 50 percent just to be on the safe side. So, where you need a 5 kV rectifier, use 7 or 8 1N4007s in series.
I have used this approach for power supplies that I have built. See the sections: Sam's Small Line Powered HeNe Laser Power Supply (SG-HL1) and Sam's Inverter Driven HeNe Laser Power Supply 1 (SG-HI1). In addition, in order to repair a particular HeNe laser power supply - the one described in the section: Aerotech model PS2A-X HeNe laser power supply (AT-PS2A-X) - after accidentally shorting its output and blowing most of the original HV diodes in the multiplier, I replaced them each with strings of four 1N4007s. There were much much cheaper than the exact replacements and seem to work just as well. However, where really high voltages are involved - like the later stages of the starting voltage multiplier - it is essential to smooth out the additional connections and coat or pot these diode assemblies to minimize the tendency for corona and arcing.
Note that you still may want to consider microwave oven rectifiers since these are readily available for $2 or 3 in single quantities from service parts suppliers like MCM Electronics. At 12 to 15 kV, the cost is higher than for a string of 1N4007s but the convenience of wiring a single part rather than 15 or 20 may be worth it. However, in their normal application, these is no more than about 5 or 6 kV across the device. Where you are using them close to their PRV ratings, potting would probably be a good idea.
For high frequency inverters (e.g., 10s of kHz or more), fast or ultrafast recovery type rectifiers must be used since at a frequency equal to 1/(2*Trr), diodes turn into short circuits. For example, a nice 1 kV, 1 A part like the 1N4948 has a reverse recovery time (trr) of 500 ns. Thus, it should have acceptable performance up to at least 200 kHz (my rule of thumb for maximum frequency of a diode: around 20% of 1/trr). Note that putting multiple diodes in series scales the junction capacitance by 1/n (where n is the number of diodes in the string) but does NOT affect Trr.
(From: Kim Clay (firstname.lastname@example.org).)
All of my HV diodes have been made using 1N4007s from Dan's Small Parts without any equalizing components. I make my HV diodes in separate assemblies using strips of perfboard cut with an Xacto razor saw to just one hole wide whatever length I need to mount all the 1N4007s. I cut 2 of them for each diode assembly and set the diodes in one strip first (first one anode up, next down, next up, etc...) then slide the other strip over the other ends of the diodes. Then, bend over all the middle interconnecting leads, trim very short, and solder with a nice round ball on each connection to minimize corona. I leave the full length leads on the first anode and the last cathode and now I have a 'custom' diode assembly of whatever voltage I need!
Microwave ovens include a capacitor of about 1 uF rated for at least 3 kV, and TVs and computer monitors may include 1 or 2 low uF, 1.6 to 2 kV ceramic caps. Higher voltage (but low uF value) caps can be found in equipment like: electronic air cleaners, igniters, bug zappers, and photocopier and laser printer HV power supplies. Modules from these devices with nice HV components may be available surplus at ridiculously low prices.
In some cases, a stack of regular capacitors will suffice but not always, and using a single capacitor with the proper ratings will be best, especially in areas like the voltage multipliers of HeNe laser starting circuits. See the section: Series Banks of Capacitors.
For modest values (up to a few thousand pF), homemade capacitors may represent a low cost alternative to hard to locate expensive 'real caps'. However, these will always be larger, bulkier, and more problematic (unless submerged in oil or completely potted) than their commercial counterparts.
While not very practical for high uF value caps, most coaxial cable can withstand more than 3 kV (at DC). RG58 (50 ohm) is mostly 100 pF per meter while RG59 (75 ohm) tends to be 67 pF meter. Coax using high density polyethylene (rather than foam) may be able to withstand 15 kV or more A similar type is used to attach HeNe laser heads to their power supplies (and it must withstand the starting voltage). So suitable lengths of coaxial cable (all that old network or video cable sitting above the ceiling tiles at work!) might be useful for experimenting. However, wrapping 1,000 meters of old cable inside your HeNe laser power supply to obtain the needed capacitance probably isn't too great an idea! :)
To achieve high capacitance, you need metal plates separated by as thin an insulator as possible - but one that won't break down under the stress of the maximum voltage applied. Capacitance is proportional to plate area and dielectric constant of the separator material, and inversely proportional to the distance between the plates.
Common printed circuit board stock - Fiberglass Epoxy - is good for about 1 kV/mil (1 mil = .001 inch) of thickness. Plexiglas acrylic has a puncture voltage of between 450 and 990 V/mil depending on quality. Materials like plate glass, ceramics, and other plastics have similar ratings. In all cases, the quality is extremely important - a single microscopic pinhole, bubble, or other manufacturing defect can render these ratings meaningless!
The dielectric constants of these materials can vary significantly even within the same product family (see below). Therefore, selection must take this into account as well. Less of a higher dielectric constant material may be needed even if its puncture voltage is lower (and it thus needs to be thicker).
The Information on Building Capacitors Page (part of the Tesla Coil Mailing List Web Site has information on the dielectric constants and puncture voltages of common materials as well as useful equations for designing home-built high voltage capacitors.
(Portions from: Dustin Lang (email@example.com).)
The basic parallel plate capacitor formula is:
A * K * 8.85 pF/m C = ------------------- * (N - 1) dWhere:
As an example: For a 1000 pF, 10 kV capacitor (typical of what might be needed for the output filter capacitor of a small high compliance inverter type HeNe laser power supply), you will need an insulator at least 10 mils thick. Using a piece of 1/8" thick Plexiglas (about 3 mm), we have: d = 0.003 m, C = 1.0E-19 F, assume K = 3.0. Solving for A we get:
C * d 1000 pF * 0.003 m A = ---------------- = ------------------- = .113 square meters (K * 8.85pF/m) 3.0 * 8.85 pF/mwhich means you'll need foil about 34 x 34 cm on a piece of Plexiglas slightly larger to provide a border to prevent air breakdown along the edges (only about 25 V/mil for air!). Or, several smaller capacitors in series (you can share adjacent plates by interleaving connections - e.g., 5 plates (17 x 17 cm) on 4 pieces of slightly larger Plexiglas).
Using a piece of .010" Fiberglass Epoxy instead (good for 10 kV), the required area would be reduced by more than a factor of 10. This material, which is readily available from PC board manufacturing companies, is excellent for these capacitors and for general high voltage insulation as well. It can be procured bare in which case you add your own aluminum foil plates, or copper clad, in which case the perimeter border needs to be removed. This can be accomplished by etching (messy chemicals) or by scribing along the edge of the desired plate area (taking care not to damage the underlying material) and then pealing away the unwanted copper.
To prevent accidental contact, cover it on both sides with additional LARGER insulating plates and clamp or tape the entire assembly!
Connections to copper can be made by soldering, taking care not to damage the substrate (raise a tab if necessary). For aluminum foil, fine wires can be inserted between the foil and insulator which will remain in place once the protective covers are fastened in place.
There is supposed to be a type of Mylar (from Dupont) that has a dielectric breakdown of 18 kV - over 20 times that of your typical plastic or glass!
(From: Chris Chagaris (firstname.lastname@example.org).)
That is a somewhat higher breakdown rating than is commonly tossed about. This rating would likely be for the specific material that Dupont produces especially for capacitor construction, which may be very difficult to obtain in small quantities. Although, I have seen a similar material available (surplus) in 32 gauge (0.00795" I think) thickness, but only at 3 inches in width. The more commonly available Mylar (polyethylene terphthalate) is usually considered safe to hold off 7,500 volts per mil. How thin you can actually find this material, that would also be commonly available, would be interesting to find out.
Polyester and other non-electrolytic capacitors may be more readily available in lower voltage ratings as well. Failure of these types is also possible (though probably less spectacular).
(Note that microwave oven high voltage capacitors represent a low hassle alternative to piles of smaller capacitors where modest capacitance (around 1 uF) at 2,500 to 3,500 VDC is required. Of course, these can also be wired in parallel or series to provide increased capacitance or voltage ratings. For series banks, these are treated as non-electrolytic types.)
When capacitors are connected in series, if one fails shorted, the rest will likely follow in rapid succession. I hope your power supply is fused! This somewhat undesirable behavior is a result of two effects:
Current balancing resistors are added to compensate for unequal DC leakage currents.
Matching of the capacitance of the individual parts may be necessary to minimize unequal AC voltage drops.
Note that if the dominant voltage across the series combination is only DC or AC, it may be adequate to only worry about balancing that. For example, a filter capacitor charged from a medium to high impedance source (compared to its reactance) will see mostly DC - the AC will be small. Therefore, only current balancing resistors are required.
A reasonable design approach is as follows:
CAUTION: Make any leakage current measurements in a manner that will prevent damage to your multimeter should the capacitor decide to short out. For example, measure the voltage drop across a high value resistor in series with the capacitor rather then directly with the meter in series.
If locating capacitors with this tolerance is not possible (i.e., your parts supply is limited), additional derating may be necessary - use additional capacitors in the series bank to achieve the needed voltage rating.
When electrolytic capacitors sit around unused, they deteriorate and their leakage current increases. Reforming is accomplished by slowing bringing the voltage across the capacitor up to its rated value (and possibly slightly beyond) while limiting the current to a safe value. Initially, there may be high leakage through the capacitor but as it reforms, this will drop down to near zero. If current limiting is not provided, this may also result in a bomb or smoke grenade.
(From: Charles Mosher (email@example.com).)
In series, each must be shunted by an appropriate equalizing resistor, in order that the variations in their leakage currents will not cause problems with grossly unequal voltage division across them.
Form up well all the capacitors to 525 Volts if you can, and then check them individually for leakage current at 450 Volts. Choose the shunting resistor value to pass perhaps 10 times this current. Watch out for cooling in the resistors; since little heat sinking will be provided, and in light of the applied voltage and the nominal voltage rating of the resistors, you may want 2 W molded carbon resistors, run at no more than perhaps 1/2 watt dissipation, in order that they be reliable.
You may have to discard some of the capacitors due to excessive leakage current or complete failure to reform.
Remember to place insulating sleeves on the cans of the capacitors whose negative sides are not grounded, as there is really no insulation provided between the capacitor negative electrode in the capacitor guts and the can itself.
Such a capacitor bank may have to be brought up to rated voltage slowly after prolonged non-use because of uneven deforming of the electrolytics. You might have to repeat the whole selection process. I built such a thing about two years ago, and after two years of non-use would not now just bang it on.
Not only can arcing take place across inadequately rated resistors, internal damage can occur which may not show up immediately. Unfortunately, catalogs may not list voltage ratings of resistors.
There are basically two options to obtain resistors with adequate voltage ratings:
Note that at the instant the tube starts, there may be much more than the steady state voltage across the resistors. This actual value will depend on how much capacitance there is in the starting circuit and how it is distributed but may be as much as the maximum starting voltage minus the tube voltage. However, everyone ignores this and it doesn't seem to matter in practice. :)
Some types of resistors are better than others in standing up to both the high voltages and high peak currents at startup. The author of the short paper "Wedding Lasers to Power Supplies"  who is a cofounder of LaserDrive, Inc. (a major manufacturer of laser power supplies), recommends Hot molded carbon composition type but suggests that some wirewound types also work well. He suggest avoiding non-hot molded carbon composition, highly inductive wirewound (which will increase dropout current, and film types (which may degrade due to the transient current of the starting pulse). On the other hand, the HeNe Laser Manual by Elden Peterson suggests that wirewound types are best. In reality, if you use conservative design, many types will be just fine. Additional ballast resistor info can be found in Power Technology's What is a Ballast Resistor and Why Should I Use One? technical note.
Some additional comments:
(From: Greg Menke
You can get really capable HV resistors, I imagine they can end up being very
expensive however. I looked at a few databooks we have here, and there is a
wide selection, but they appear mostly custom or manufactured on demand type
Stringing resistors together is a good alternative but can be a little tricky
because you must stay within the dielectric rating for each resistor or it
breaks down, plus you have to manage the dissipation as well. I'm using 2
parallel sets of 30 series resistors as a HV bleeder on my 30 kV capacitor.
At 30 kV, each resistor should be dissipating .4 watts, they are rated for
1/2 W, so I *shouldn't* need an oil bath for them. Actually it was pretty
hard to find resistors to withstand >= 1000 VDC each, regular resistors are
rated for 200 volts or so. Farnell had them, 'metal-glazed', good to 3 kV
You can get really capable HV resistors, I imagine they can end up being very expensive however. I looked at a few databooks we have here, and there is a wide selection, but they appear mostly custom or manufactured on demand type devices.
Stringing resistors together is a good alternative but can be a little tricky because you must stay within the dielectric rating for each resistor or it breaks down, plus you have to manage the dissipation as well. I'm using 2 parallel sets of 30 series resistors as a HV bleeder on my 30 kV capacitor. At 30 kV, each resistor should be dissipating .4 watts, they are rated for 1/2 W, so I *shouldn't* need an oil bath for them. Actually it was pretty hard to find resistors to withstand >= 1000 VDC each, regular resistors are rated for 200 volts or so. Farnell had them, 'metal-glazed', good to 3 kV each.
(From: Gernot Stoffel (Beamchief@gmx.de).)
In many cases it is recommended to connect mains transformer output of AC HeNe power supplies a well defined way, and do some additional grounding efforts: First, I suggest grounding both the transformer iron core AND the negative (cathode) tube connector, each by 220K resistor or so; or at least couple them by a 470K resistor.
Second, the tranformer pin connected with the *inner* part of secondary coil ALWAYS should be connected next to cathode ground potential of the entire circuitry! (E.g., when using a typical voltage-doubling Greinacher circuit, the two diodes should be connected to the outer end of coil, not the inner one, because this diode potential is oscillating with double transformer peak voltage relative to cathode ground, while the other pin, between those two chains of elcaps, stays just at a stable DC potential, apart from some ripple they work against.)
Reason is that before glow discharge gets ignited, the extended high voltage (up to 10 KV or so) tends to "spray" into the air from any part that's not properly insulated (mostly: ballast resistors). This parasitic discharge circuit gets closed by complement spray discharges - mostly inside the transformer. (Effect might continue even at regular tube working, on a much lower "parasitic" level.) Thus, it's ALWAYS better to connect it the described way, keeping voltages between core and coil as low as ever possible, and to make defined potentials by using the recommended grounding resistors. (Note, the little switching-transformers of DC drivers normally have similar preferences, too.)
Not knowing this, my personal laser GAU was in the early nineties when all of a sudden my fine 10 mW "handmade" NEC was striking, just five minutes before an important presentation on a varnishing in Bonn. I found the big's mains high voltage trafo output coil being completely dead. Removing iron core (fortunately, it was not a :(( welded type), then 5,000+ turns of (too) fine wire (fortunately, it nearly was *not* baked together with PU fluid; just that tiny little bit I had to use a saw against, to get secondary coil finally removed VERY sensitive), I found its innermost filaments highly eroded, in an absolute ridiculous way, if that had not been so bad. Nevertheless, trafo had a good, solid two-chamber insulating frame made of thick glass whisker composite nylon that was not damaged as much one could see! (That's no criterion; for sure.) What had happened???
I use a hand-made case from clear Plexodur (a slightly yellowish derivate of Plexiglas, but without its severe disadvantages, that is brittleness and an outstanding bad chemical sensitivity, e. g., getting into handy pieces by alcohol). It's *absolutely* nice-looking, for that machine is covered with well-done technics and mechanics up to the very top (take you some exciting photo seria as far I've chosen my next digicam, what's a *long* story on its own); and it's handy, too. But naturally, I had to darken its 14-inch laser tube, to prevent discharge from blinding the whole audience when running; for I don't build tricky laser deflectors just for *that*. So, totally coated it with two layers of matt black car spray paint a very careful way and baked it in. (Remember, just the naked tube once made a 1000 $ charge; meaning an entice car load of raw eggs was peanuts against). Tested and calculated before, that this would not effect tube temperature behavior in an intolerably bad way. (Could be proved later-on; while self-adhesive black foil is much worse, by causing heat build-up; by not covering the whole thing; and by highly uncertain long-time stability risks. My beloved Contax all goes into parts just because its glue re-transforms to cosmical dust.) What I did *not* know then, nor expect in any way, was that these black paint sprays *all* are conductive a tiny little bit, by using soot particles for pigment. (Tried nearly ten brands of spray varnish later to avoid this disadvantage; but it's all nearly the same, beside a factor of three or so.) But learned it quite fast, for tube ignition behavior was lousy; so I freed the zones directly around tube-connector passes from paint and extra-insulated them by epoxy-filled, black caps, taken from wonder-glue flasks. (So, this is just another thing best *never* thrown away.) Igniting behavior grew much better, but by far was not perfect; noticed that tube surface still got highly charged during ignition. Hmmm. Fixed the whole problem the "sledgehammer way" then with an extra circuitry, made by a tiny, gas-filled 350 V / 5 kA spark discharger item and a "bullet-proof" 10 nf 630 V capacitor, contacting it by a five-bucks-rhodium-plated-gold-blinking spring contact pin and holder to the tube's central copper collar I made there, fairly under the paint, for proper temperature taking, and ground. During ignition, capacitor now got charged within a second or so, TICK!, discharger ignited with some pretty hundred Amps and little blue spark, and kicked the whole tube discharge a proper way (ask Sam why it did). Truly clever, I thought. Even a touch better than using a piezo lighter module a similar way. Worked fine.
Until the transformer was dead, several years later, the described way. Then it was not clever any more, not at all.
The simple, very obvious and primitive truth is that the metal ends of the tube are connected to the inner voltage(s) as well - lately by the ion discharge! -, especially that one near the anode pin; and so the conducting tube surface was still rather contacted with it. AND the truth was that the transformer output coil by natural means was switched just the opposite way than it would have been able to bear that problem. (As usual, Murphy had done his very best.)
Well - the rest was easy work: Insulated those tube ends, by first another time carefully removing the white silicone caps as I did several times before, e. g. for fine-optimizing the resonator alignment the adequate way it *can* be done at operating NEC tubes if you really know what you're doing, then by using short sections of black shrinking tube covering the removed ring zones of paint, and re-monted the end-caps. Dark as hell. AND I re-assembled the transformer, after professional efforts to improve wiring-cabin (frame) insulation: knowing the primarily-coil number of turns (had been cool enough measuring ten testing turns, manually threaded to that dead transformer before slaughtering, so there was no need to remove/move the core twice, thus putting 75 pieces of thin metal at alternating orientations), by that simply fixed the number of needed secondary coil turns, then calculated optimum wire diameter (0.15 mm) to fill the frame (3.3 times better and twice stable than it was before (0.1 mm)); attended 0.2 kg = 1,220 m of this stuff (needing about 850 m, I knew); and *then*, turn beside turn, layer onto layer, with some additional cross-section insulating efforts I made that new coil manually, 5,375 turns, meaning 21,500 well-guided frame corners, forty layers, just assisted by a little mechanical turn-counter I had put to my wiring "machine" made of 'Fischertechnik' construction kit. Lasted two days or so. Blocked the whole transformer frame with thin, clear epoxy sealing compound at the end, using good vacuum to get any gas away during that procedure, and then re-assembled transformer core. So now it will work properly even at the bottom of Marian trench, if cable is long enough. Since then there is no problem any more.
OK: Cold tube does not ignite very proper (lost function of my fine ignition "sledgehammer", by finally insulating the tube coating from discharge voltages an effective way); but lately *does*.
(Conclusion: HeNe tubes definitely *don't* like conducting black paint, in terms of their ignition behaviour, but just want to show all they got.)
Question: Which wire diameter 'd' shall I take to fill a chambered transformer frame cavity quite proper, when 'n' turns are wanted? A needless thin wire type is bad, for increasing resistance, heat-up etc., and often is uncomfortable to practice, too; and too thick wire type does not fit the problem, but is one of the ugliest ways to restart work. And guessing is no method for getting proper results here.
Answer: When the wiring cabin (frame) cavity is 'w' wide and 'h' high (at any side), usable area is A=w*h. Total (!) maximum wire diameter then is d:<=1,075*sqrt(A/n), when coil gets wound turn by turn, layer onto layer. By winding a transformer coil not that exact way (but not extremely crowded!), a safety margin of about -5% in diameter is quite sufficient (because space grows as the square with diameter; so this will yield about 10% additional space: 0.952=0.9025). So you then get d:<=1.02*sqrt(A/n). Or make it easy: d:<=sqrt(A/n). Wire diameter usually gets specified without the insulation coating. A single coating adds about 0.03 mm to the wire diameter, a double coating adds 0.05 mm. Thus, you have to subtract this from your total result, and then take the next available diameter same or below this value. You always will be on safe ground.
(Taking all dimensions in mm, the result is in mm, too.)
[Proof: With a cross-sectional area 'A' and 'n' turns, each one has maximum space 'A/n' to use. Nearly all wires are in a hexagonal grid (if no left-handed gorilla made that coil), where wire diameter is the side-to-side diameter 'd' of such hexagon. A little geometry sketch shows, on just a hundred different ways, that hexagonal area is then sqrt(3)*d2/2. So with A/n>=sqrt(3)*d2/2 you get d:<=sqrt(A/n)*sqrt(sqrt(4/3)). This factor is about 1.075.]
Important: If you wind high-voltage coils, an endless job requiring perfect precision all the time, you may need additional insulating foils between layers. Then substract another 10% from wire diameter before selecting it, and it's best to use double-coated type if you can get it.
Question: Which length of wire do I approximately need with that nearly-optimum diameter?
Answer: When frame chamber kernel is x*y, average length of any turn is (2*(x+y)+h*pi. ('h*pi' is the net turn length around all the four corners.) By this, total wire length is about n*(2*(x+y)+pi*h).
(Taking all dimensions in mm, result is in mm, too.)
Note: This is the absolute maximum length which can occur, beside of some additional cm for connecting.
The input side doesn't present too many problems:
However, there are special concerns for the secondary side. For higher power (higher voltage) HeNe tubes, these are even more critical for both safety and consistent performance:
Keep components adequately spaced from each other, and grounds or the case, where voltages differ substantially. Maintain separation of at least 1 inch per 10 kV though more is better. Or, include additional insulation such as a sheet of Plexiglas or other plastic, or Fiberglass-Epoxy printed circuit board material (without copper or copper on only one side if that side is grounded).
These are both significantly overrated for the laser applications but are also quite robust and resistant to high temperatures and mechanical abuse.
There is special RTV Silicone formulated for high voltage applications. The normal bathtub caulk or window sealer may corrode the circuitry (if it smells like vinegar) or may degrade (decompose, become discolored) after some amount of time resulting in conductive paths. I have used the clear GE stuff apparently without problems on the HeNe laser power supply I routinely use for testing but your mileage may vary.
CAUTION: Know the limits of your power supply. Ironically, being too successful at insulating the starting voltage can overstress marginally specified components which would normally be protected because the voltage never got high enough to exceed their ratings!
WARNING: This resistor does NOT reduce the danger from the supply - it is to protect the equipment. The available current can still be lethal.
There are special resistors designed for use in high voltage circuits and these would be best. However, series strings of lower value resistors should be acceptable with proper construction techniques. See the section: High Voltage Resistors for more information.
You can extend the cable all you want as long as:
Although usually specified by tube manufacturers to be 7 kV to 12 kV depending on size, the actual starting voltage is often much less, typically only 3 to 5 times the operating voltage. However, to be sure, adhering to the minimum starting voltage specifications is still a good idea unless you intend to only use one tube and test it before constructing your power supply.
Here are many possibilities for starting HeNe tubes:
A voltage multiplier can be constructed from common diodes and capacitors relatively easily. Alternatively, a 3X or 4X type may be purchased as a replacement for the type found in some color TVs and monitors - you may even have one gathering moss in your junk box!
See the section: Voltage Multiplier Starting Circuits.
The spark generating modules used in electric stove or furnace igniters, and pocket lighters that use a capacitive discharge approach would also work with few additional parts (a pair of HV diodes). See the section: HeNe Starter Using Electronic Ignition (or Other Similar) Modules.
These approaches work particularly well for hard-to-start tubes because up to 16 kV or more is available and mostly enclosed (inside the flyback potting material if it has an internal rectifier). Thus, problems with corona from external wiring are minimized.
See the section: Starting Circuits Using Pulse or Flyback Transformers, for more information.
On some internal mirror HeNe tubes or laser heads, you may find a wire or conductive strip running from the anode or ballast resistor down to a loop around the tube in the vicinity of the cathode. (Or there may be a recommendation for this in a tube spec sheet.) This external wire loop is supposed to aid in starting (probably where a pulse type starter is involved) in a similar manner to what is used for an external capillary (above). However, with an internal mirror HeNe tube, the capillary is usually isolated from the outside envelope. Therefore, I wouldn't expect there to be much, if any, benefit especially when using a modern power supply, but it might help in marginal cases. And, running the high voltage along the body of the tube requires additional insulation and provides more opportunity for bad things to happen (like short circuits) and may represent an additional electric shock hazard.
For a basic AC line powered supply, I recommend the parasitic multiplier approach (1) as it can be constructed from readily available inexpensive components: 1N4007s (about 5 cents each) and .001 uF, 1,000 V ceramic disc capacitors (about 10 cents each). Depending on the number of stages in the multiplier, between 6 and 28 of each of these components will be required.
Of course, you can also use higher voltage diodes and capacitors to simplify the construction but they will probably be much more costly. See the section: Sam's Small Line Powered HeNe Laser Power Supply (SG-HL1) for a tested design using inexpensive parts.
Alternatively, a commercial voltage multiplier block can be used. See the section: Color TV or Monitor Voltage Multiplier.
A voltage multiplier can be constructed from common diodes and capacitors relatively easily. Alternatively, a 3X or 4X type may be purchased as a replacement for the type found in some color TVs and monitors - you may even have one gathering moss in your junk box! See the section: Color TV or Monitor Voltage Multiplier.
A typical design with 7 diodes (3-1/2 stages) is shown below. For small tubes, fewer stages can be used. Going much beyond n=7 or 8 (4 stages) is probably not useful as the losses from diode and stray capacitance and leakage will limit output.
R1 C1 C3 C5 C7 X o---/\/\---||------+-------||------+-------||------+-------||------+ 1M, 1W D1 | D2 D3 | D4 D5 | D6 D7 | +--|>|--+--|>|--+--|>|--+--|>|--+--|>|--+--|>|--+--|>|--+--o HV+ | | | | Y o----------+-------||------+-------||------+-------||------+ C2 C4 C6 HV- o---------------------------------------------------------------------o HV-Where:
With n diodes, HV(peak) is approximately (X(peak) * (n + 1).)+ Y and HV(average) is (X(peak) * n) + Y.
Note that an even number of diodes may be very slightly better since there is less ripple on the output when the tube is running. However, with a high value R1 (10M) this is so small anyway that it really should not matter and an odd number of diodes saves components but results in nearly the same peak starting voltage.
For use in HeNe laser starting applications where no real current is required, R1 limits power to the multiplier once the tube fires. Power is then drawn from point Y through the string of diodes.
Multipliers can be used with both line operated supplies and high frequency inverters but since the capacitors must be larger at the (lower) line frequency.
The voltage ratings of the diodes and capacitors must be greater than the p-p output of the transformer.
Because the capacitors used in the multiplier are so small, they cannot really supply much current. Once the tube fires and current starts to flow, the ladder just becomes some series diodes. Then you're back to the basic power supply output (rectifier or doubler and filter capacitors), with a few diodes in series.
And to emphasize, there really is virtually zero added ripple from these starters once the tube is running even if the apparnet impedance of the components fed from the HV transformer is relatively small (e.g., a few M ohms). It's not only that the current through the capacitors is small, since this could still result in 100s of volts of ripple across the ballast+tube or regulator. More fundamentally, it's because the forward voltage drops across the HV diodes are nearly constant as a function of current as long as they are conducting. In fact, it's only the left-most HV diode of the multiplier that really matters since most of the ripple current will flow through it to get absorbed by the main filter capacitor at the input to the multiplier. The only added ripple will be due to the non-zero slope of that HV diode's V-I curve at the operating point, probably less than 0.1 V/mA. On a power supply where the starter could be turned on and off with a switch while the tube was running, the difference in voltage ripple at the output of the multiplier was totally undetectable with all the ripple originating from the main filter capacitor. However, this does mean there should be no added resistance placed between the starter and the main filter capacitor, as that would result in increased starter generated ripple due to the IR drop, which is not constant.
There are two types of nodes:
In order for no current to flow through the diodes:
Assuming a peak-to-peak amplitude of 2 units (just to keep the diagram simple), the voltages at each node will be:
R1 C1 +1(AC) C3 +3(AC) C5 +5(AC) C7 +7(AC) X o---/\/\---||------+-------||------+-------||------+-------||------+ +0 D1 | D2 +2 D3 | D4 +4 D5 | D6 +6 D7 | +--|>|--+--|>|--+--|>|--+--|>|--+--|>|--+--|>|--+--|>|--+---o HV+ | | | 1| Y o----------+-------||------+-------||------+-------||------+ C2 C4 C6 G o----------------------------------------------------------------------o HV-Where:
In practice the actual output will likely be somewhat less due to stray capacitance and other losses.
Since these voltage multiplier blocks are designed for horizontal deflection (e.g., ~15.7 kHz), they are ideal for drive from an inverter. I have also tested two different multipliers on my line powered (60 Hz) supply without any problems. In fact, one of those I tried was removed from a TV because it had a short in the focus divider network but this didn't affect HeNe tube starting performance at all!
These are usually 3X units though some 4X types are also available. Hooking them up is very straightforward and the entire multiplier is well insulated (totally potted in Epoxy) with a high voltage output lead (just remove the CRT suction cup connector). The unneeded terminals (e.g., F, CTL) can be ignored.
Some of these have a capacitively coupled drive input. If this is NOT the case, the addition of a high voltage capacitor (.001 uF, 3 kV typical) in between the X and IN terminals is required to achieve the full multiplication factor. If you do not know if there is an internal capacitor, include C1 - it will not hurt anything. Since you do not want much current from the voltage multiplier, insert a 1M to 10M series resistor in this path as well.
HV Multiplier R1 C1 +----------------+ X o----/\/\----------||--------| IN HV |----------o HV+ 10M, 1W .001uF,3kV | ECG535 F |-- NC Y o----------------------------| DC CTL |-- NC +----------------+ HV- o--------------------------------------------------------o HV-ECG535 is just one typical 3X model. Almost any HV multiplier of this general type will be suitable. However, if you have a choice, obtain a 4X multiplier as this will provide a bit more margin (though a 3X model should be adequate for most HeNe tubes). The IN terminal may be called GND, REF, or COM on some models - or there may be a pair of terminals with two of these names. If so, they should be tied together. F and CTL are Focus and Control respectively. These and any other special terminals can be left unconnected.
Note: These may be called either pulse or trigger starting circuits and I use the terms somewhat interchangeably.
For some types of specialized or older HeNe (and other gas) lasers where there is a separate gas reservoir and the surface of the capillary is accessible, the output of a true pulse starting circuit (single shot, not a charge pump or inverter since a high dV/dt pulse shape is required) may be applied to an external electrode in a similar manner to that used for triggering a xenon flashlamp. The pulse circuits below may be suitable (but leave out any HV diodes). However, this type of laser construction is not that common and this approach will probably not work for the inexpensive mass produced sealed HeNe tubes with a wide glass envelope surrounding a narrow capillary since it will be difficult to ionize the gas inside the bore in the center of the tube.
While the current required to start a HeNe tube is negligible, the energy needed to achieve sufficient voltage given the stray capacitance of the wiring, HeNe tube anode, and other connected components is not. With a voltage multiplier type starting circuit, this just means it takes a little time to charge up to the required voltage but as long as the leakage is small (it usually can be ignored), the tube will start eventually. For a pulse (trigger) type staring circuit, the energy must be provided in one shot or a charge pump must be used to accumulate smaller energy packets.
Thus, pulse starting circuits are more effective if the wire length (and thus the capacitance) between the power supply's final rectifier and HeNe tube anode is minimized.
Locating a large enough pulse transformer to start the tube in one shot may prove to be a challenge. However, a charge pump (really, just an extra high voltage diode or two) can be added to any of these circuits so that the energy in each pulse can be accumulated. The tiny trigger transformer from a disposable pocket camera will not work very well by itself but should work with a such a charge pump. See the section: Pulse Starting Circuit Using Small Flyback Transformer for the charge pump configuration.
A flyback transformer from a small B/W TV or computer terminal does work quite nicely with a charge pump. An automotive ignition coil would probably be large enough for single shot operation.
Replacing the pulse button with a simple transistor oscillator results in an inverter based starting circuit. See the section: Inverter Based Starters for more information on this approach.
The first three circuits that follow differ only in the type of trigger transformer used but operation is otherwise identical. These have not been tested.
The final circuit uses a flyback transformer from a long forgotten video display terminal and has been tested using a manual pushbutton.
Some other options using salvaged parts (or at least salvaged designs) from household devices include:
Operation is similar to a repeating strobe trigger.
Z o--------------------+-----------+ | | R5 o / R1 +---/\/\--+------+---------+ \ 10M 4M 1W | | ):: +---o Y / NT1 | __|__ SCR1 )::( | NE2 | _\_/_ TIC106 )::( Q1 R8 | +--+ R3 T | / | )::( MPSA43 +--/\/\--+---+--|oo|--/\/\--+--|---' | +-----+ ::( | 100K | | +--+ 1K | | | | ::( Tube- R8 |/ C / _|_ / / | _|_ C2 ::( o-+--/\/\--| R2 \ --- C1 R4 \ \ R6 | --- 1uF ::( | 1K |\ E 1M / | .5uF 1K / / 600K | | 600V ::( CR1 HV+ / | | | 200V | | | | +--|>|--o \ R7 +--------+---+--------------+--+------+---+ o / 200 | T2 | | o-+----------+ HV-The voltage divider formed by R1 and R2 charges C1 from the high voltage power supply (Z, a lower voltage tap if possible to reduce the dissipation in R1 and R2). At the same time, C2 charges from R5 and R6 (this time constant is faster than that of the relaxation oscillator). Once the voltage across NT1 (NE2 neon tube) reaches about 90 V, NT1 breaks down dumping C1's charge through the gate of SCR1. This turns on and discharges C2 through T2 generating a 5 to 10 kV pulse in series with the high voltage power supply ionizing the gas in the HeNe tube. CR1 must be a 15 kV or greater high voltage rectifier and prevents reverse voltage from appearing at the tube.
Should the HeNe tube not fire on the first pulse, the process repeats at about a 2 Hz rate until current starts flowing in the tube. A current of about 3.5 mA through the HeNe tube and R7 results in a voltage drop of .7 V across B-E of Q1 turning it on. This short circuits the relaxation oscillator shutting off the starting circuit.
Component values can easily be adjusted to accommodate the specifications of your specific power supply voltage and HeNe tube current.
These are all basically capacitive discharge ignition systems and indeed, an automotive ignition coil may be satisfactory for the trigger transformer where isolation is not needed!
Operation for both is as follows:
The trigger capacitor, C2, charges through the voltage divider formed by R1 and R2. When a pulse is input to the gate of the SCR via the high voltage coupling capacitor, C1, it triggers dumping C2 through the primary of the trigger transformer.
WARNING: The voltage rating on C1 must be adequate - with a safety margin - for your power supply.
The input, T, comes from the autotrigger circuit which is part of the circuit shown in the section: Pulse Starting Circuit - Trigger Transformer with Isolated HV Winding.
Y o----------------+-------+-----------+--------+ | | | T2 | R1 \ __|__ SCR1 +-+ +-+ 1M / _\_/_ TIC106 o )::( o \ / | )::( C1 | | | )::( T o----::----|----+ | )::( .01uF | | | )::( 2kV | R3 / | )::( | 1K \ | +-+ ::( | / | | ::( | | | C2 | ::( +----+--+------)|---+ ::( | | 1uF ::( R2 / _|_ 600 V ::( 1M \ --- C3 ::( / | .01uF ::( | | 600V ::( CR1 W o----------------+-------+ +---|>|---o HV+
Y o----------------+-------+---------------+--------+ | | | T2 | | | +-+ +-+ | | o )::( R1 \ _|_ C2 )::( 1M / --- 1uF )::( \ | 600V )::( | | )::( | | )::( | | +----+ ::( | | | ::( | | SCR1 __|__ ::( | | TIC106 _\_/_ ::( | | / | ::( C1 | | | | ::( T o----::----|-------|---------+ | ::( .01uF | | | | ::( 2 kV +-------+----+ R3 / | ::( o CR1 | | | 1K \ | +---|>|---o HV+ R2 / C3 _|_ | / | 1M \ .01 --- | | | / uF | +----+--+ | 600 V | W o----------------+-------+
Higher input voltage, more or fewer turns on the flyback, or a different capacitor may improve response - your challenge!
The primaries on the flyback are ignored and a new one is added - 5 turns of #20 or thicker insulated wire wound anywhere on the core where it will fit. As long as the original primary windings are not shorted, they will not interfere with circuit operation.
Locate the HV return and guess at the polarity - it will work properly only one way since the output is a huge spike. Reverse the input connections if you cannot get the tube to fire and you think everything else is correctly wired.
I used a separate 50 VDC power supply to drive this circuit but you can use a tap on the main filter capacitor of the main supply as well.
R1 + o-------------/\/\---+------------+ +----o Y 1K | | T2 | | | +--+---|>|---+ | | o ::( CR2 | | +----+ ::( | _|_ C2 )::( | 50 VDC --- 20uF 5 T )::( | | 100V #20 )::( | | )::( | | +----+ ::( | | S1 | ::( o | CR1 | _|_ | +------------+---|>|---o HV+ - o--------------------+-----o o----+ Start
Many modern gas stoves, ovens, furnaces, and other similar appliances use an electronic ignition rather than a continuously burning pilot flame to ignite the fuel. Well, guess what is in one of these modules? That's right - a high voltage pulse generator which is ideal for starting HeNe tubes!
The typical design is remarkably similar to that of the circuits described in the previous sections starting with: Starting Circuits Using Pulse or Flyback Transformers. It is powered either from 115 VAC or 24 VAC and generates of a series of sparks at a rate of perhaps 1 or 2 per second. These modules should be readily available as replacements at your local appliance repair shop or parts supplier.
Internally, a typical 115 VAC unit consists of a simple rectifier/filter or doubler, neon bulb relaxation oscillator triggering an SCR which dumps a capacitor's charge into a pulse transformer - very similar to trigger circuit of a repeating strobe.
All that is needed to convert one into a HeNe starter is a suitable power source and a pair of HV diodes to form a charge pump (as described above). Depending on design, it may be possible to run these from a DC power supply or one derived from your HeNe operating voltage.
The Harper-Wyman Model 6520 Kool Lite(tm) module is typical of those found in Jenne-Aire and similar cook-tops. Input is 115 VAC, 4 mA, 50/60 Hz AC. C1 and D1 form a half wave doubler resulting in 60 Hz pulses with a peak of about 300 V and at point A and charges C2 to about 300 V through D2. R2, C3, and DL1 form a relaxation oscillator triggering SCR1 to dump the charge built up on C2 into T1 with a repetition rate of about 2 Hz.
Based on the 10 kV or so output of this module, I estimate that T1 has a step-up turns ratio of about 35:1 but have no easy way of determining the exact number of turns since it is potted in clear plastic. Fortunately, the primary side of the circuit was accessible after prying off the bottom cover plate. My totally wild guess is that the primary and secondary are about 25 and 900 turns respectively.
C1 A D1 T1 o H o----||----------------+-------|>|-------+-------+ +-----o HVP+ .1uF D2 1N4007 | 1N4007 | | o ::( 250V +----|>|----+ | +--+ ::( | | | )::( +---/\/\----+ | #20 )::( 1:35 | R1 1M | C2 _|_ )::( | R2 / 1uF --- +--+ ::( | 18M \ DL1 400V | __|__ ::( | / NE2H | _\_/_ +-----o HVP- | | +--+ | / | | +----|oo|----+----|----' | SCR1 | C3 | +--+ | | | S316A | .047uF _|_ R3 / | | 400V | 250V --- 180 \ | | 1A | | / | | R4 2.7K | | | | | N o---/\/\---+-----------+------------+----+-------+For use on DC, remove C1 (and D1, D2, and R2 if you like - they won't affect anything) and provide about 300 V from your HeNe operating supply (typically from point A of the schematics shown in the chapter Complete HeNe Laser Power Supply Schematics. Just add a pair of HV rectifiers (D3 and D4) to form a charge pump and current bypass, and an optional auto-shutoff circuit (not shown).
The resulting HeNe starting circuit is shown below:
T1 o D3 Z o-------------+-----------------+-------+ +----------+--|>|--o HV+ | | | o ::( | 15kV | | +--+ ::( | | | )::( | | | #20 )::( 1:35 | | C2 _|_ )::( | R2 / 1uF --- +--+ ::( | 18M \ DL1 400V | __|__ ::( D4 | / NE2H | _\_/_ +--+--|>|--+ | +--+ | / | | 15kV +----|oo|----+----|----' | SCR1 o C3 | +--+ | | | S316A Y .047uF _|_ R3 / | | 400V 250V --- 180 \ | | 1A | / | | R4 2.7K | | | | N o---/\/\------+------------+----+-------+Some types of lighters generate a spark in a very similar manner. The guts from these may be pressed into service as HeNe starters (a much more noble cause than their original application, I might add!) with the addition of a couple of HV diodes. One type uses a 12 V battery, 330 uF, 16 V capacitor, trigger switch, and tiny pulse transformer (less than 0.5" x 0.5" x 0.4"). There are about 12 turns on the primary and several thousand turns on the secondary producing an output of more than 5 kV. The circuit is basically similar to that shown in the section: Pulse Starting Circuit Using Small Flyback Transformer.
The trigger portion of an electronic photoflash or strobe (not the main energy storage/discharge components and xenon tube) may also be used to start HeNe tubes. These run on either batteries or the AC line (sound familiar?) and produce HV pulses of between 4 to 10 kV. You may already have a dead pocket or disposable camera laying around (or your expensive, neglected Nikon) which has such a module. See the document: Notes on the Troubleshooting and Repair of Electronic Flash Units and Strobe Lights for detailed sample circuits as well as schematics from some common pocket and disposable cameras and separate flash units.
Here is a circuit which I assume is for an electronic air cleaner or something similar. I received this thing in the mail (no markings). I did check to make sure it wasn't a bomb before applying power. :-) A photo is shown in: Air Cleaner HV Module. This one is nice because all parts are accessible so modifications are easy to make. Or, just strip the HV diodes and caps from a couple of these modules and build your own voltage multiplier type starter. Complete air cleaners may use a potted circuit that looks like a mini-HeNe power supply brick but isn't. These can still be used for HeNe laser starters but your options are more limited.
The AC line powered driver and HV multiplier are shown in the two diagrams, below:
D1 T1 o H o--------------|>|----+---+--------------------+ +-----o A 1N4007 | | Sidac __|__ SCR1 ::( | | R3 D2 100 V _\_/_ T106B2 ::( AC C1 | +--/\/\---|>| / | 200V ::( Line Power .15uF _|_ 1.5K |<|--+--' | 4A o ::( 350 ohms IL1 LED 250V --- _|_ | +-------+ ::( +--|<|---+ | C2 --- | | )::( | R1 | R2 | .0047uF | | | .1 ohm )::( N o---+--/\/\--+--/\/\--+ +-----+--+ )::( 470 3.9K | +--+ +--+--o B 1W 2W | | R4 | +--------------------------------+---/\/\---+ 2.2M (Remove, see below)The AC input is rectified by D1 and as it builds up past the threshold of the sidac (D2, 100 V), SCR1 is triggered dumping a small energy storage capacitor (C1) through the primary of the HV transformer, T1. This generates a HV pulse in the secondary. In about .5 ms, the current drops low enough such that the SCR turns off. As long as the instantaneous input voltage remains above about 100 V, this sequence of events repeats producing a burst of 5 or 6 discharges per cycle of the 60 Hz AC input separated by approximately 13 ms of dead time.
The LED (IL1) is a power-on indicator. :-)
The transformer was totally potted so I could not easily determine anything about its construction other than its winding resistances and turns ratio of about 1:100.
A o C3 | +------||-------+ (-5 kV) R5 R6 D3 | D4 D5 | D6 R7 R8 (+5 kV) Y o---/\/\---/\/\--+--|>|--+--|>|--+--|>|--+--|>|---/\/\--+--/\/\---o HV+ 10M 10M | C4 | 220K | 10M +------||-------+ | D3-D6: 10 kV, 5 mA _|_ _|_ C3,C4: 200 pF, 10 kV --- C5 --- C6 C5,C6: 200 pF, 5 kV | | B o--+----------------------+The secondary side consists of a voltage tripler for the negative output (Y) and a simple rectifier for the positive output (HV+). I assume this asymmetry is due to the unidirectional drive to the transformer primary.
From my measurements, this circuit produces a total of around 10 kV between HV+ and Y, at up to 5 uA. The two output voltages are roughly equal plus and minus when referenced to point B.
The only modification required for our needs is to remove R4 to isolate the HV secondary from the AC line and earth ground. Once this is done, it may be necessary to add insulation and/or reroute some of the wiring on the PCB to prevent arcing as the HV attempts to find its way to ground via the AC since the negative of the main HV power supply (i.e., Tube-) should be grounded.
Then, connect the high voltage outputs up like that of the other inverter based starters using a HV bypass diode (at least 10 kV) to isolate the HeNe tube anode circuit from the main power supply. A 1 nF, 10 kV capacitor across the diode may be needed to reliably start tubes on short cables (which don't have much capacitance). Using a pushbutton switch to activate the circuit is probably easiest though leaving it on all the time will probably do no harm since its output current is so small. For the purist, an opto-triac or similar device can be added to disable the pulse discharge circuit once current flows in the HeNe laser tube.
Where there is no access to R4 (e.g., the unit is totally potted), it probably won't be possible to obtain the full output voltage from the positive or negative leads to ground (only between them). This is because tying one of the outputs to ground (as would be required if the cathode of the HeNe tube is earth-grounded as it should be unless totally enclosed and insulated) will partially short the internal voltage multiplier via R4 to the AC line (the Neutral wire is grounded at the service panel). In this case, leave the negative output of the air cleaner module floating. This may still result in enough starting voltage for smaller HeNe tubes (up to perhaps 2 mW). Don't even think about attempting to isolate the HeNe tube or starter from ground unless as noted, it is totally enclosed!
A quick check to see if one of these modules has enough output voltage to start your HeNe tube is to connect it up as follows:
+-------------+ + 100K +-------------+ H o-----| |------/\/\---|- |-|-----+ AC Line | Air Cleaner | +-------------+ | 115 VAC | HV Module | - HeNe Tube | N o-----| |-----+ | +-------------+ |<-- (Connect only if | | R4 is removed) | G o-------------------------+---------------------------+
The HeNe tube will glow faintly (actually pulsing at the AC line frequency) if there is enough voltage to start. Assuming the tube is good, there may even be an indication of lasing, but with very low (and possibly erratic) power - maybe as much as 1 microWatt. :) So, one of these modules can be used to perform a quick check of a HeNe tube as well.
Alternatively, with a bit more rewiring, the multiplier can be used in-line with the HeNe laser operating voltage supply. Or, the high voltage diodes and capacitors (probably from a pair of these units) could be removed and used for a parasitic multiplier driven from the AC line or inverter based power supply.
I tested this starting approach with the inverter, doubler, and filter capacitor portion of SG-HI3 (with no starting multiplier, see the section: Sam's Inverter Driven HeNe Laser Power Supply 3 (SG-HI3).) It worked quite well starting instantly every time. The electrostatic air cleaner unit used for this experiment, while similar to the one described above, was totally potted and therefore I had to leave its negative lead floating. But it was still more than adequate to start a 1.5 mW HeNe laser tube.
I intend to use one of these modules in conjunction with a partially broken Aerotech brick power supply for HeNe lasers in the 5 mW class - model LSS5(L)6.5, 2,500 +/-300 V at 6.5 mA. This particular unit has lost its ability to generate sufficient starting voltage, peaking at only 4 kV (it should be greater than 10 kV), cause unknown. However, it runs just fine. The external starter will enable an otherwise useless brick to be salvaged for about 99 cents plus the cost of a HV diode, HV cap, and pushbutton switch (or opto-triac if I feel more ambitious, or maybe have both manual and automatic start, switch selectable). I have already tested the setup but still need to mount everything in a nice box. The biggest problem was rerouting one of the traces on the air cleaner PCB to prevent arcing once it was isolated with its outputs floating on the HV+ side of the Aerotech brick. :)
The HV output is then placed in series with the HV+ of the main supply as above (Y to HV+) through high value isolation resistors (10M or greater rated for at least 15 kV). These are required for safety and protection: so that the output of the main supply cannot appear on the inverter with any significant current and so that the inverter is not overloaded (partially short circuited) by the tube once it starts. In addition, the resistors prevent any significant ripple on the inverter output from turning the HeNe tube *off* once it starts. While it takes 1000s of volts to start a HeNe tube, it only takes a few volts to shut it off!
A high voltage blocking diode (e.g., one or more 15 kV PRV microwave oven HV rectifiers) bypasses operating current around the inverter once the tube starts. A stack of 1 kV diodes (as many as needed for the maximum starting voltage) can also be used. General purpose (non-fast recovery) types like 1N4007s are fine. Make sure they are well insulated - inside a thick plastic tube, for example. Since 1N4007s are only a few cents each in modest quantity (not from Radio Shack!), this may be much cheaper. It IS more flexible as a rectifier of nearly any desired voltage rating can be custom made. I have replaced a fried (from overvoltage) microwave oven rectifier (in a HeNe laser power supply, NOT a microwave oven!) with a stack of 20 1N4007s to create a (hopefully) more robust 20 kV unit. However, it would probably be better to derate them by 25 to 50 percent to be sure of the minimum PRV rating.
The 1nF capacitor may only be needed when driving HeNe laser tubes or heads attached by short cables or open wiring. Normally, the capacitance of the cable to a laser head will provide enough energy storage to keep the discharge going once it strikes. But where this is inadequate, starting may be erratic. If the output of the power supply always goes though a few feet of coax, it probably isn't needed.
Y o-------+-----|>|-----+-------o HV+ (to ballast resistor) | 15kV | +-----||------+ | 1nF | / / R1 \ R2 \ / / \ \ | | o - Starter + o
As shown above, the inverter is configured to add its output to the operating voltage of the main supply. For this scheme to work, the negative terminal of the source must float at the voltage at 'Y'. This may be possible when using a flyback transformer with your own windings. However, it is almost never possible when using existing circuitry. In that case, the negative can be tied to ground (R1 is omitted) but then the benefit of having the sum of the operating and starting voltages is not available. The added voltage may not matter unless the HeNe tube needs all the help it can get!
To make these fully automatic - to disable the starter once the HeNe tube fires - a very simple circuit needs to be added to the inverter controller. For example, where drive is provided by a 555 timer chip in astable mode, the following circuit will work:
Vcc (555) o | / R4 \ 1K / \ | _ +--------o R - Reset input to 555 (low) | |/ C Q1 Tube- o-------+-----+-------| 2N3904 | | |\ E / | | R3 \ _|_ C1 | This assumes the cathode of the 270 / --- .1uF | HeNe tube was grounded. Current \ | | sense resistor and 'Beam On' LED | | | note shown. +-----+---------+ _|_ -
Once the HeNe tube starts, current of more than about 3 mA through R3 switches the transistor ON which disables the inverter drive by forcing the 555 to the reset state. C1 prevents the inverter from being cutoff if the tube pulses momentarily but doesn't stay on.
For examples of inverter based starters in actual HeNe laser power supplies, see the sections: Sam's Mid-Size Line Powered HeNe Laser Power Supply (SG-HL2), and Sam's Ultrabeam(tm) Line Powered HeNe Laser Power Supply (SG-HL3).
Since it is relatively straightforward to generate voltages up to 15 kV using small readily available flyback transformers (even more using those from color TVs or monitors), and this gets ADDED to the operating voltage, such an approach is quite effective for large and hard-to-start HeNe tubes. In addition, because the inverter does not depend on high voltage AC or DC from the main supply, it is also easier and safer to construct and troubleshoot, and this also provides added design flexibility.
Here are some possibilities:
Although the first of the following options should start any HeNe tube you are likely to encounter (assuming it is not blown), it never hurts to be prepared:
These modules probably cannot be floated on top of the operating voltage supply since the HV return is likely connected internally to the common and the case. However, with over 20 kV typically available, running single ended really should not be a problem. :-)
One particular type (found in a 19 inch workstation monitor) produces 25 kV (more after some tweaking) at up to 1.1 mA (and a bunch of other voltages) from a 26 VDC, 2.5 A power supply. I have been using one of these modules to start larger HeNe tubes. For this purpose, it runs happily at 15 VDC and only draws about 1 A (since output current is negligible). See the document: Salvaging Interesting Gadgets, Components, and Subsystems for a detailed description of this module as well as other unconventional high voltage power supplies (and a lot of other neat stuff).
But, what a pity to waste the potential (no pun...) of such a device on something as boring as starting a HeNe tube (even if it is BIG)!
One key advantage of using predesigned circuitry is that you are less likely to destroy power transistors and other expensive parts - and I have blown my unfair share :-(.
See the section: Sam's Super-starter(tm) for a specific example of this kludge, um, err, approach. :-)
Now, I know what you are thinking about now: "Why not run the HeNe tube from one of these HV supplies entirely instead of just using it for starting?". The problem is that their design provides high voltage (which we only need for starting) but at relative low current - typically a maximum of 1 or 2 mA. Unlike the design used in "Sam's Inverter driven HeNe Laser Power Supply 2" which allows complete control of both the frequency and pulse width of the drive, without serious circuit modifications, your options are likely to be quite limited. Available current will probably be marginal at best even for small HeNe tubes (3 or 4 mA at the operating voltage) and you may blow out components in the attempt.
The entire horizontal deflection and high voltage sections of a long obsolete and lonely ASCII video display terminal were pressed into service for starting larger HeNe tubes. A source of about 12 VDC at 1.5 A is needed for power and a 555 timer based oscillator is needed to provide the fake horizontal sync:
Well, it turns out there was an unused spot on the board ready made for this circuit (well almost, at least there was a pattern for a spare 8 pin DIP! So, once the thing was basically working, I built the oscillator onto the board to reduce the clutter!
CAUTION: Since these power supplies were designed for a specific purpose under specific operating conditions, their behavior when confronted with overloads or short circuits on the output will depend on their design. It may not be pretty - as in they may blow up! Take care to avoid such events and/or add suitable protection in the form of fast acting fuses and current limiting to the switching transistor.
They are based on the spark generating assembly from a disposable butane lighter that uses a piezo element to generate the spark (as opposed to old fashioned flint and wheel) or the starter assembly from a gas grill. These devices are based on the piezo-electric effect which results when certain materials are twisted or otherwise deformed. Even these dirt-cheap lighter or gas grill starters are able to produce more than 10,000 V mechanically with the press of a button.
A simple charge pump using a pair of high voltage diodes completes the starting circuit. These are needed because a single press of the button may not generate enough charge to achieve the required starting voltage. This is true at least of the type from disposable lighters. Perhaps, one from a gas would work by itself. The charge pump enables the charge generated by each button press to be accumulated on the stray capacitance of the wiring, CR2, PE1, and the HeNe tube anode. The diodes should have a PRV rating greater than the starting voltage requirement of your HeNe tube, 15,000 V should be sufficient. These type of diodes are typically used in B/W TVs or monochrome computer monitors. Microwave oven high voltage rectifiers also work but not nearly as well (many more presses of the button are needed) due probably to their higher capacitance and leakage.
Both of the circuits below work but the second one may have an edge since the capacitance being charged is slightly smaller since it doesn't include PE1 and its wiring:
HV+ o CR1 CR2 | Rb Y o------|>|----+----|>|----+----/\/\-------o Tube+ | | | - + | +----|X|----+ |--^ PE1 HV+ o CR1 CR2 | Rb Y o-----+----|>|----+----|>|----+----/\/\-------o Tube+ | | | - + | +----|X|----+ |--^ PE1Complicated, huh?
CAUTION: Since the piezo assembly/push button is connected to the positive terminal of the high voltage power supply, make sure it is well insulated or the only thing being triggered may not be the HeNe tube!
Make sure the wiring is short and well insulated - these high voltage pulses tend to go wherever they want unless properly trained. :-)
I tested this with the piezo assembly removed from a Scripto lighter found in the park. (You will have to agree that this is classic MacGyver!) The required part is self contained and pops off with minimal disassembly (just remove the sheet metal flame guard). With this particular device, the pointy electrode is negative. I do not know if this is always the case. If the HeNe tube does not start after a reasonable number of button presses, try interchanging the connections to the piezo assembly.
Replacement push button starters for gas grills can be purchased at most home centers and may even work better because they are likely to produce a higher voltage higher energy spark. I will not be responsible if your dad cannot barbecue the dogs and burgers because you 'borrowed' the starter. :-)
Long-reach butane "multipurpose lighters" are also readily available almost everywhere. I paid 77 cents for one on a special. As with the Scripto, the piezo mechanism is a self contained module but keeping the entire lighter intact is most convenient since it includes a large trigger button and nicely insulated high voltage wires. The butane tank can't easily be removed since part of it is the support structure for the piezo module and trigger, but it could be drained via the fill valve thing at the bottom outdoors and away from open flames if you can't wait through the required number of cookouts. In any case, I'm sure you'll find a way to use up the butane so that only the much more useful spark remains! ;-)
As with transformer based pulse starting circuits, these are more effective if the wire length (and thus the capacitance) between the first high voltage rectifier (CR1) and the HeNe tube anode is minimized.
For my bare 1 mW HeNe tube, two presses of the button were required. For a laser head with a 3 foot long high voltage cable, 4 or 5 presses were needed but it always started eventually.
However, my initial attempt at using one of those hand-held BD-10 "Oudin" coils intended for vacuum leak testing (among other things) totally failed. (These may also be called Tesla coils or spark coils.) When used at various locations on the glass envelope of a good tube with my adjustable HeNe laser power supply running just below the normal operating voltage so the tube was flashing, there was absolutely positively no effect of any kind except to cause a glow in the gas reservoir. The rate of flashing didn't change. When I used it on my impossible-to-start tube, there was similarly no effect, even when applied directly to the tube anode. In this case, the tube seemed to light down the bore but wouldn't catch. But maybe that wasn't a fair test: While my impossible-to-start tube had been known to run normally - if it could be started - the laser time I tried it was several years ago and who knows how hard to start it is now! (The gas is good though as it would flash with lasing if run with reverse polarity.) So, perhaps I'll try again with a tube I know can indeed run if started.
And, here's another one to be taken even less seriously: Cut a pair of holes in the front and back of the a microwave oven just large enough for some of those longer HeNe tubes to fit! Starting cycle: 3 seconds on HIGH (with a cup of water to act as a load)! Yes, I know, it would look kind of silly to have your HeNe laser tube sticking out of the family microwave! I did say I wasn't really serious, didn't I? :) WARNING: Please don't attempt this stunt - both dangerous levels of microwave leakage and HeNe tube frying arcs are quite possible!
OK, I think that's enough on HeNe laser starters! ;-)
Supply voltage fluctuations do affect operating current proportionately more than would be expected since the voltage drop across the tube is fairly constant. Therefore, nearly the entire change appears across the ballast resistor. The smaller the ballast resistor, the larger the current change for a given voltage change. This may mean that a 2 percent change in line voltage produces a 10 percent change in tube current - still not a big deal for most applications. For experimental use, it is perfectly acceptable to use ower supply without regulation and a Variac to control the input as a way of adjusting output current. The setup I routinely use for testing HeNe laser tubes is a combination of a Variac, an unregulated HeNe laser power supply, and a ballast resistor selector and meter box. See the section: Ballast Resistor Selector and Meter Box. Some commercial HeNe lasers have even used a power rheostat in series with the input of the main HV transformer for this purpose!
However, where stability is required (both short term and long term due to component and tube aging) or where it is desired to be able to switch tubes without monitoring tube current and adjusting the input voltage, a regulator is essential. This also provides an easy way of adding a current adjustment control to accomodate a variety of HeNe tubes.
The regulator must be designed with protection so that it (and the power supply) are not damaged during starting or should the laser head or HeNe tube not be connected (where there is no current but very high voltage) or as a result of fault conditions like accidental short circuits.
The circuits are fairly simple. However, more than one high voltage transistor in a series cascade may be needed to achieve the desired compliance range. If only there were such a thing as an LM317 (set up in constant current mode) that accepted a 3 kV DC input! NPN transistors with adequate voltage ratings (300 to 500 V) are somewhat more common than PNP types especially for higher power types. MOSFETs with suitable ratings are also available and relatively low cost. Manufacturers of HV devices include Supertex and Zetex. There are many more. Universal semiconductor replacement manufacturers like NTE Electronics and others have a few devices that are suitable but the quality and consistency of original name brand parts may be better, more complete specifications are available, and they are often cheaper as well. (Note: As of January 2001, NTE has purchased the assets of ECG so the ECG parts listed below may no longer be available, use the NTE or SK equivalent which usually has the same number with the 'NTE' or 'SK' prefix.)
Some typical part numbers are listed below. Those rated a watt or less are likely to be in some sort of a TO92 package while higher power types are in TO202 or TO220 packages except for these particular >1 W MOSFETs which are in an SMT TO-243AA/SO89 package.
(There are also many high power MOSFETs with Vdss ratings of 500 V or more.)
Although all the linear regulator designs in subsequent sections of this chapter and in the chapter: Complete HeNe Laser Power Supply Schematics use bipolar transistors, you can substitute one or more high voltage MOSFETs (depending on the specific compliance range required) but need to take care that you NEVER violate the maximum ratings, particularly maximum Vgs, which is usually quite low (e.g., +/- 20 V). This means protecting against voltage spikes during power on/off and starting, as well as being fault tolerant of even momentary shorts on the output. However, high voltage MOSFETs do not suffer from the second breakdown failure mode of bipolar transistors. Since MOSFETs have essentially zero gate current, the drive circuitry is less of a problem (the voltage divider resistors can be very large). However, the Vgs threshold voltage is not as tightly specified (and generally higher) compared to the Vbe drop for a bipolar transistor which will affect the design as well. In the past, MOSFETs with similar voltage ratings to HV bipolar were more expensive but this appears no longer to be the case. I don't know if that is the reason I have never seen a linear MOSFET based regulator in a HeNe laser power supply.
For all but the smallest HeNe tubes, the larger (>1 W) transistors will be required to handle the power. In principle, TV or monitor horizontal output transistors (up to 1000 V or more, high power) could be used to reduce the number of required devices in a cascade but they tend to have very low gain (less than 10) making drive more difficult. High voltage high power MOSFETs could also be used in this manner but both these approaches are probably more expensive than using a larger number of lower voltage transistors.
For other gas lasers like the carbon dioxide or helium-cadmium type requiring tube currents of up to 100 mA or more, the use of high power bipolar transistors or MOSFETs may be required if a linear regulator is to be used. Its basic circuit topology will be similar or identical to that used for the low current HeNe laser power supply but much greater attention will have to be paid to heat dissipation at these higher currents!
The compliance range for a commercial power supply with a linear regulator is typically between about .75 and .95 Vo where Vo is the output voltage under load. For example, a power supply with an output voltage of 2,000 V will be able to provide between 1,500 and 1,900 V. There is no reason why a wider range could not be implemented - it is just a matter of cost (more transistors) as the maximum allowed voltage drop across the pass transistor or transistor cascade determines the lowest output voltage possible before protection kicks in (or the regulator blows up!).
While I have seen commercial HeNe laser power supplies using these types of transistors without heatsinks, I consider this unwise as they can run hot. A heatsink represents a very small insurance premium!
One advantage of this approach is that the regulator can be designed so that the control circuitry all operates at low voltage. Therefore, no high voltage, likely to self destruct, solid state components or cascades are needed. The regulator transformer provides the LV to HV interface via its turns-ratio. Its insulation provides isolation from the high voltage side components. By using thyristors with phase control or some other sort of chopper circuit for the load, power dissipation can be cut way down compared to a simple linear regulator or linearly controlled load. Overcurrent and short circuit protection are easily added minimizing the chances of permanent damage from most fault conditions.
Certain versions of the SP-261 exciter for the Spectra-Physics 125A large-frame helium-neon laser use this approach. A current sense voltage is fed to a unijunction transistor/SCR phase control circuit which acts as a variable load on the low voltage side of a transformer whose high voltage winding is in series with the high voltage secondary of the transformer driving the main doubler/filter circuit.
In addition to constant current regulation, high and low limits and fault protection can be implemented by sensing input current and load current and voltage.
With a wide compliance design, the starting voltage is within the range of the power supply output voltage and there is no separate starter. Other designs provide a more limited compliance range and use a separate starter. Still others use a combination of these - a moderate compliance inverter and modest boost multiplier.
These are the approaches use in most modern commercial HeNe and other laser power supplies.
+------------+---o Tube- | | R1 |/ C Q1 | Z o----------/\/\------+----| MJE3439 _|_, VT2 5M | |\ E '/_\ 150V 1W | | | _|_, / | VR1 '/_\ +->\ R2 _|_, VR3 1N5241B | |v / 5K '/_\ 150V 11V | | \ | | | | | +---+--+------------+---o HV-CAUTION: If your tube is mounted with its cathode connected to a metal case, earth ground must be tied to Tube- for safety as Tube- may be at a potential of several hundred volts with respect to HV-.
The compliance range is about 300 V so a ballast resistor still needs to be selected to enable the regulation to be effective. Use the procedure described in the sections starting with: Selecting the Ballast Resistor to determine an appropriate ballast resistor value for the desired tube current and a power supply voltage 150 V less than that of your (unregulated) supply. This circuit should then maintain that current constant and permit some adjustment (or modulation if desired) of the current on either side of the set-point.
For greater compliance, a cascade of high voltage transistors can be used as outlined in the section: High Compliance Cascade Regulator. This shows a low-side approach. Also see the section: Spectra-Physics Model 247 HeNe Laser Power Supply (SP-247) which is a high-side version of a similar circuit.
Rb Tube+ +------------+ Tube- Rs IL2 LED G HV+ o---/\/\--------|- |-|----+---/\/\----|>|------+----+ +------------+ _|_ 1K Beam On | | LT1 - R1 / | 120K \ | 2W / | | |/ C Q1 +--| MJE2360T | |\ E R2 / | 120K \ | 2W / | | |/ C Q2 +--| MJE2360T | |\ E R3 / | 120K \ | 2W / | R4 | |/ C Q3 +-----/\/\------+--| MJE2360T | 120K |\ E | 2W | | | | | Rx | R5 |/ C Q4 Z o------------/\/\----------+-----+-----/\/\------+--| MJE2360T 470K | | 10K | |\ E | R6 / |/ E | |100K \<------------| | CR1 _|_, / Adjust Q5 |\ C | 1N4744 '/_\ | 2N3906 | | 15V | | (PNP) +----+ | | +--+ | | | | v R7 R8 | HV- o---+-----+---+-/\/\----/\/\---+ Range 5K 1.5K
The PNP transistor (Q5) buffers the reference voltage so that the very low current source (point Z through Rx tapped off of the main filter capacitor string) can drive the base of the cascade. The way to understand its operation is to realize that Q5 implements an emitter follower with respect to the tap on R6. Without it, the cascade would be driven directly from the 15 V reference point; Q5 drags down Q5's base to the value (+0.7 V) set by R6. I think, I hope. :) If the reference can supply a couple of a few mA (say, from a 50 V source through 10K (Rx), R6 can be reduced in value and the circuit could be simplified as shown below:
| | Rx | R5 |/ C Q4 Z o------------/\/\----------+-----+ +---/\/\---| MJE2360T 10K | | | 10K |\ E | R6 / | | | 5K \<------+ | CR1 _|_, / Adjust | 1N4744 '/_\ | | 15V | | +--+ | | | | v R7 R8 | HV- o---+-----+---+-/\/\----/\/\---+ Range 5K 1.5K
The base resistors, R1 through R4 (more or less) equally distribute the voltage between G and HV-. The respective transistors act as emitter followers and maintain approximately the same voltages across their C-E terminals. Within the compliance range, the voltage across R7+R8 will be equal to the voltage on the wiper of R6 (give or take a diode drop depending on whether Q5 is present).
CAUTION: Make sure that the open circuit voltage of your power supply minus the laser tube and ballast resistor voltages cannot exceed the total breakdown voltage of the transistor cascade (4 * Vceo in this case as bad things may happen)! As drawn, it is certainly suitable for a 2,200 V power supply with almost any tube. However, if you are using this with a 3,000 V supply and a .5 mW tube requiring 1,100 V at 3 mA, the compliance range will be exceeded and the transistors may blow - pop-pop-pop-pop! Then, you get maximum current through the ballast resistor and laser tube which will make them extremely unhappy :-(. A bag of 200 V zeners in series could be used to limit the voltage across the transistor cascade. Or, see the section: Spectra-Physics Model 255 Exciter (SP-255) for an example of a simple overvoltage protection circuit.
R7 sets the overall current range. R6 adjusts the current from a minimum determined by the maximum voltage across R1 through R4 to the limit set by R7. However, regulation should be better when R6 is set near its maximum since small changes in voltage/current will have correspondingly less effect than if you are trying to regulate with just a few mV difference to play with! The adjustable R6 can be omitted entirely if desired (tap off the top of CR1) and R7 used as the current adjust. The only thing undesirable about this is that this control is somewhat non-linear so that the high current portion is all squashed at one end.
Rs is a current sense resistor. Monitoring across it with a volt meter is a convenient way of setting tube current. Sensitivity is 1 V/mA. A current meter can also be put across it and will read tube current directly. The LED provides a 'Beam On' indication and rough measure of tube current as well.
The basic linear cascade regulator should be capable of higher current as long as the transistors are mounted on adequate heatsinks can care is taken to ensure they don't operate anywhere near the limits of their Safe Operating Area (SOA).
Devices like the MJE3439T are often used without any heatsinks in commercial designs so adding one will go a long way to preserving their lives. What may need to be done in addition is to reduce the value of the base resistors so that there is adequate drive based on the transistor's Hfe specifications. To be safe, increasing the number of stages in the cascade to compensate for the additional current and keep the power dissipation of each transistor may be prudent.
For dual (or multiple) tube applications or dual discharge lasers like the Spectra-Physics model 125 (SP-125), a pair of regulators could also be used. (Note that separate ballast resistors will be required for these.) For the SP-125 in particular, a common cathode at the center of the tube is fed from separate anodes at each end. Where a power supply is to be built (or the original one repaired or rebuilt), the total discharge current of 35 mA can easily be split into two currents of 17.5 mA each - easily handled with very minor modifications to the basic design in the section: High Compliance Cascade Regulator. See the section: Spectra-Physics 120, 124, and 125, HeNe Laser Specifications if you are curious about the SP-125.
Another alternative is to parallel multiple cascades and just combine their outputs. However, with proper heatsinking, this shouldn't be necessary for HeNe lasers. But carbon dioxide, helium-cadmium, or other lasers requiring greater discharge current, this and/or the use of higher power transistors may be required. Or better yet, the use of a more efficient high frequency inverter instead!
A simple approach like this works quite well for supplies using high-droop transformers (e.g., oil burner ignition or luminous tube/neon sign types). The no load voltage on these can easily exceed 1.5 times the operating voltage and the excess energy in the filter capacitors would normally be dumped into the HeNe tube when it starts - which is hard on the tube!
A shunt regulator consisting of a stack of zener diodes (or gas tubes, remember those?) prior to the bypass HV diode could be added to reduce this difference by providing a substitute current path - loading down the transformer - until the HeNe tube starts.
In this case, since the regulator is only passing significant current before the tube starts, with careful design, the zeners or whatever don't need to dissipate high power for too long. That is, unless the tube fails to start or becomes disconnected!
For high droop transformers, a dropping resistor is probably not needed at all. However, for normal power transformers, a dropping resistor (Rd in the diagram below) should be used to prevent the transformer from being overloaded (as well as the regulator being smoked. These are able to source much greater than their rated current (at least for a short time) with little reduction in output voltage.
Rd Y o---/\/\---+----o Y' To Starter | _|_, ZD1 '/_\ _|_, ZD2 '/_\ . | . | . _|_, ZDn '/_\ | o HV-When contemplating using this sort of approach for regulation of the operating voltage itself, both the continuous power dissipation of the regulator components AND power supply transformer (or inverter) must be taken into consideration.
We won't even talk about the efficiency of the power supply when running continuously with a shunt regulator but you wouldn't be playing with HeNe lasers at all if overall electrical->optical output efficiency were a prime consideration!
Kim (see the sections: Kim's Mid-Size Line Powered HeNe Laser Power Supply (KC-HL1) and Kim's Flyback Based HeNe Laser Power Supply (KC-HI1)) first suggested this after he somehow acquired a pile of 36 V, 400 mW zener diodes and was trying to figure out a good use for them!
(From: Kim Clay (firstname.lastname@example.org).)
I did try the zener diode shunt regulator a few different times and it worked quite well. I just have to set it up for each tube that I'm using AND those little zeners can put out lots of heat when there are so many in a small area! have a small muffin fan in the 5 kV supply & I made sure there was airflow over them.
Where the internal current regulator of an inverter-type HeNe laser power supply has failed so that there is no longer any way of controlling laser tube current except via the DC input voltage, it's possible to add an external regulator to do just that.
Some HeNe laser power supply bricks tend to fail in this manner, expecially if hooked up incorrectly.
So, step 1 of the procedure to use the external DC regulator technique would be to hook up a suitable model backwards. :) One that is very reliable in this regard - that is, it fails reliably and in the same way every time - is the Laser Drive 103-23, normally rated for 21 to 31 VDC input with a fixed output current of 3.5 mA, suitable for many 1 mW-class HeNe laser tubes. However, when "modified" in this manner, it can provide 7 mA or more. There will still be a maximum continuous power rating, which for the 103 is around 7 W. So, that's still a limiting factor, but now the full range of the power supply will be available.
The very simple circuit is shown in HeNe Laser PSU External Linear Regulator 1. The return for the HeNe laser tube current goes though a resistance, variable from 1.5K to 3.5K ohms to ground. This provides a voltage proportional to tube current across it. The non-grounded side feeds the cathode of an 8.2 V zener whose anode goes to the base of a 2N3904 transistor. Its emitter is grounded and its collector goes to the Adj input of an LT1084 low dropout regulator, wired for a maximum voltage of about 10 VDC (200 ohms and 1.5K ohms for the two resistors). (A common LM317 would also work, but would require a volt or so more headroom.) So, when the voltage across the variable resistance exceeds about 8.9 V, the 2N3904 turns on and limits the output of the DC regulator.
The filter capacitor on the output of the LT1084 is required for stability of this particular regulator. (This large a cap is not needed with an LM317.) But minimizing the capacitance is desirable to reduce the loop delay. and there is also generally substantial filtering inside the brick itself anyhow. The capacitor across the sense resistors assures loop stability. The residual noise and ripple at the output of the HeNe laser power supply should be close to spec'd values. In fact for those supplies with a ripple reducer as the final filter stage, the residual ripple is nearly undetectable.
With these component values, the adjustment range is from about 3 to 6 mA. With a 100K ballast resistance and 1 mW HeNe laser tube, the actual DC input to the brick is about 12 to 17 VDC. For the 1 mW tube at 3.5 mA, about 13 VDC is needed on the input. For a 2 mW tube at 5 mA, about 15 VDC would be needed. Of course, running the 103-23 continuously would also exceed the 7 W rating of the power supply at the higher tube voltage.
I also tried 2 other "modified" :) HeNe laser power supply bricks (all Laser Drive!). Actually, they both had lost regulation on their own, cause unknown, but I didn't need to help them along! The first was a 111-Adj-1 from an HP-5517B laser. This worked without modifications to the circuit at about 7 to 9 V for 3 to 6 mA into the 1 mW tube. A much larger 380T brick, with an output rating of 3100 V at 5 to 7 mA and an input of 22 to 26 VDC, operated in a manner very similar to the 103-23 with the 1 mW tube. I then connected it to a Melles Griot 05-LHR-911 laser head and that worked fine except that I had to add additional filter capacitance across the DC to the supply for it to remain stable over the entire range of input voltage and tube current. I used a 10,000 uF, 25 V capacitor, though that was probably overkill. I then tried a Uniphase 1125P and that worked fine as well. I haven't yet tried it with a high power tube, mainly because I didn't have one handly that was still within the specs of the 380T. It wouldn't even start an Aerotech OEM12R or Uniphase 1145P, though these may have been hard-start and my DC power supply didn't go above 20 V. But if it works (or if I find a higher power brick that's lost its regulation), this would have the potential to be the basis for a universal HeNe laser power supply that would not only be able to be used at the normal spec'd current with lasers from less than 0.5 mW to over 20 mW (using a brick like a broken 380T), but also be able to drive lasers having requirements that may fall through the cracks of major power supply manufacturers. For example, try to find a compatible brick for a tube that runs on 4 mA at 2.5 kV as a standard product! Even for those that are adjustable, the minimum current may be much too high for supplies able to provide 2.5 kV.
The main drawback to schemes like this is that the addition of the linear regulator adds substantial power dissipation, and thus one of the main benefits of using an inverter brick is lost. So, the DC input voltage to the regulator should be selected such that it's only enough above the required DC input voltage to the laser power supply to account for the voltage drop of the regulator, and enough additional to handle variations in tube voltage and current. More optimal would be a variable switchmode DC power supply feeding the external linear regulator. However, if power dissipation is a serious issue, an only slightly more complicated switchmode regulator could be implemented instead.
But this really does give a second life to otherwise damaged power supplies! So, while I can't answer the question of whether laser power supplies can have 9 lives, they can certainly have 2, at least under some conditions!
Note that it should be possible to use this approach to reduce the output current of a properly functioning brick, but I haven't tested that! This would then permit a lower current tube to be run on a power supply with a higher fixed current (no adjust pot).
The major internal functional blocks of these IC are the Sawtooth Oscillator (ramp generator) and Voltage Comparator. Timing components (Rt, Ct) set the (constant) oscillator frequency. The Voltage Comparator, subtracts the error signal (Ve) from the instantaneous value of the sawtooth waveform. Its output is high only if Verr is greater than Vosc. The pulse width is therefore a linear function of error voltage over a fairly wide range.
The output may need to be buffered to drive the main switchmode transistor (bipolar or MOSFET). Until the HeNe tub starts, some means must be provided to assure that the duty cycle is limited to prevent damage due to transformer core saturation and excess current.
Rt Ct Voltage +--/\/\--+---||---+ Comparator Buffer | | | Vosc +---------------------+ |/|/|/ |\ |\ | | | Sawtooth Oscillator |--------|- \ _|_|_ | \ _|_|_ +---------------------+ Verr | >------| A >--------o Drive +---|+ / | / +------+ | |/ |/ +---| F(s) |--+ | +------+ | | |\ | Vcs = V(Current Sense) Vcs (+) o-----/\/\---+----|- \ | Vcl = V(Current Level) | | >---+ Vcl (-) o---+-/\/\---+ +--|+ / Typical PWM Drive (Expanded) | | Rcl _|_ |/ _ _ +---+ - Verr Low ____| |___________| |______ Current Error ______ ______ Limit Amp Verr Med. ____| |______| |_ ___________ ________ Verr High ____| |_|The Error Amp may actually be part of the PWM controller IC. External sense circuitry may consist of nothing more than a resistor in the HeNe tube return to provide a voltage proportional to current, and F(s) may consist simply of a RC network to control loop response.
Typical oscillator frequency is 200 kHz. To analyze this circuit precisely would require digital signal processing (DSP) techniques. However, where the loop response is limited to much less than the switching frequency, analog techniques can be used. Consult application notes for the PWM controller ICs for details.
See the section: HeNe Inverter Power Supply Using PWM Controller IC (IC-HI1) for an example of a design using this approach.
It's simple in principle to add under and over-current protection to any HeNe laser power supply. Setting a low current limit would protect against a disconnected laser head, or one that decides to sputter or pulse after it warms up. Setting a high limit on current would protect against arcing or a dead short, or a tube with a very low discharge voltage due to a broken bore. Or, a power supply that decided to deliver much more than the desired current.
There are many ways of implementing the protection circuits but they all need to compare the tube current against low and high limits and disconnect input power if either limit is violated. Different time constants for the low and high limits allow for an occasional hiccup that would momentary decrease the average current but would trip on any overcurrent condition. A reset timer would disable the low limit for a predetermined time to allow for slow-start tubes.
The brute force approach uses a pair of voltage comparators monitoring the voltage drop on a current sense resistor. The AND of their output drives a relay that controls input power. RC networks implement the reset delay and set the low and high time constants.
An alternative if you're so inclined is to use a PIC chip with a few lines of code to do the timing and control, with at most a pair of voltage comparators if the PIC doesn't have at least 2 analog inputs. One could think of all sorts of extensions to this approach if you're so inclined.
My basic solution is to add a sensitive relay that is initially bypassed for starting and then remains energized by the laser tube current. A simple crowbar circuit shuts the supply down if the current goes too high. There is no need to handle under-current explicitly since running a HeNe laser tube slightly below spec'd current won't hurt anything and is an unlikely failure mode anyhow. If the current goes so low that the tube winks out, then the relay turns the supply off.
The schematic for what I built into a pair of modified Spectra-Physics 248 exciters may be found in HeNe Laser PSU Protection Circuit 1. The Opto-22 MP240D2 is a Solid State Relay (SSR) which is guaranteed to turn on with less than 3 mA - low enough for virtually any HeNe laser tube. The AC version has "contacts" (actually a triac) rated at up to 2 A and 240 VAC. The DC version, also shown, using an Opto-22 DC60MP handles loads up to 60 V and 3 A. (A sensitive electromagnetic relay could also be used if a suitable one can be found.) A momentary switch bypasses the "contacts" when starting the laser. For general testing, there's no need for both types: Add a mechanical relay to the AC version powered via the SSR and use one set of NO contacts to control DC supplies or the external interlock present on many lab supplies and laser systems. Another set of contacts on The relay can also switch AC loads to handle higher current and eliminate issues of inductive kickback or capacitive surge current destroying the SSR.
The crowbar is a home-built SCR consisting of a 2N3904 and 2N3906 with each base connected to the collector of the other transistor. The SCR is across the SSR input in series with a pot to adjust the current trip set-point. Should the current go too high, the voltage drop across the relay input and pot will exceed the threshold of a zener diode triggering the SCR, which bypasses input current to the SSR thus shutting off the power supply.
To avoid tripping due to the transient high current that may be present for a fraction of a second on many supplies when the laser starts, substitute a momentary switch with two poles - the second being used to disable the crowbar by shorting the base of Q1 to ground while it's depressed. Or, to avoid modifying the PCB, use the second pole on the switch to insert a 5 or 6 V zener diode across the SSR input. The voltage across the zener will be high enough to allow the SSR to turn on, but limit any excessive voltage due to momentary over-current. (If the SSR isn't allowed to turn on, the output will drop out for an instant as the button is released. For the DC SSR, this would be quite short, but for the AC SSR, it could be up to 1/2 of a cycle, 1/120th of a second for 60 HZ power.) Another useful addition is an anti-chatter feature for the situation where the tube drops out. This avoids the possibility of unwanted restarts causing the power (or interlock) to pulse rapidly. The simplest solution where an electromagnetic relay is used along with the SSR is to route the coil holding voltage through its AC NO contacts. Then, the first time the relay is de-energized - even if for a very short time - these are interrupted and it is off for good. Alternatively, a negative pulsed detector consisting of a capacitor and transistor can be used to trip the crowbar.
I have assembled and tested PCBs for both the AC and DC versions. See Photo of HeNe Laser PSU Protection Circuit 1 Assembled. The only difference is the type of SSR module. (These are available via Sam's Classified Page.)
My lab rat testing unit uses the AC SSR version and a electromagnetic relay. It is shown in Sam's HeNe Laser Protection Rig. The protect PCB is built into a really ugly repurposed box (original application unknown, top panel component locations determined by desire to fill pre-existing holes). The small toggle has OFF/ON/START positions while the knob sets the current - note the professionally labeled panel, or total lack thereof. The yellow LED is the one on the schematic in series with the laser tube current, so it will be on when the laser is running and the protect circuit is active. The neon indicator is lit only before engaging, or when it has tripped. The AC output is via a short IEC cord; the pair of banana jacks connect to a set of electromechanical relay contacts to be used with DC-input HeNe laser power supplies or external interlock circuits. The trip range is approximately 3.75 to 8 mA using a 1K ohm pot (and R5 replaced with a jumper). The DPM reading 6.60 mA is a separate unit built as a high tech upgrade to my old analog current meter. :) It is used to set the trip current more precisely and for monitoring tube current.
Most commercial HeNe laser power supplies have some type of current regulation, so the actual effects are already well below the worst case. Nonetheless, if you're determined to have the cleanest (optical-wise) laser on the block, the next two sections deal with add-ons to linear and switchmode HeNe laser power supplies to virtually eliminate ripple and noise in the laser tube current.
So the first thing to do with a candidate supply is to check for AC in the tube current. This can be done by installing a 1K ohm resistor in the cathode return and monitoring the voltage across it with an oscilloscope. The sensitivity is then 1 mA/V. For a typical laser with a nominal 6.5 mA tube current, 65 mV p-p ripple and noise would be 1 percent which is certainly acceptable for most non-critical applications. 0.1 percent would be readily achievable if it isn't already that good. If it's much worse than 1 percent, the regulator is probably not working correctly, the tube is not compatible with the supply, the line voltage is too low or too high (or not set correctly), or there is no regulator!
At least, most of these supplies have accessible internal circuitry - they are not potted like modern supplies. The exceptions, like some of those from Coherent and Hughes, will need to be dealt with like switchmode supplies as discussed in the next section. Or, find another candidate.
Assuming there is a regulator and it's working properly but you'd like a cleaner supply, the two easiest things to do are:
Alternatively (or in addition), adding a second filter capacitor bank separated from the first by a resistor to form an RC filter section cap provide a dramatic decrease in ripple. For example, a second filter capacitor bank of 2 uF with a 50K ohm resistor in front of it will reduce the ripple by more than a factor of 10. The down side is that there will be a voltage drop across the resistor, which will reduce compliance range of the supply. Thus, a supply with the worst case spec'd leakage current of a higher voltage maximum tube voltage may need to be used.
WARNING: Adding capacitance also increases the danger of electrocution should an unfortunate part of your anatomy come in contact with its output!
I modified a pair of Spectra-Physics 248 exciters in this way to provide ultra-low ripple and noise for driving metrology lasers. See the section: Converting the SP-248 into a Low Ripple Protected HeNe Laser Power Supply.
Where the ripple is found to be excessive (more than 3 or 4 percent) in a commercial switchmode HeNe laser power supply, it's likely that the internal ripple reducer has failed and an external circuit will be needed to restore acceptable performance. However, as long as the DC current regulation is still working correctly, this should be possible.
Since virtually all of these supplies are totally potted, access to anything inside is not possible (or at least not very practical unless you're really determined . So, the basic approach is to add an external circuit consisting of a pi-section filter with two HV capacitors and a single transistor circuit that provides a high impedance to AC current but doesn't interfere with the internal DC current regulator. This is similar to what is used in some commercial HeNe laser power supplies such as the Melles Griot 05-LPM-829 HeNe Laser Power Supply. The "pi" consists of C18, the circuit with Q7, and C17.
The circuit I built is functionally identical, though some part values differ slightly. Switchmode HeNe Laser Power Supply Ripple Reducer 1 cuts the ripple from a Laser Drive 103-23 (1,100 to 1,500 V at 3.5 mA) by a factor of at least 50, from about 2 percent p-p to under 0.04 percent p-p. It goes directly at the output of the power supply - before the laser tube ballast resistor or the Alden cable of a laser head. Note that the largest effect is for the high frequency ripple from the switchmode inverter, but it should also reduce line frequency ripple substantially. However, if not sufficient, the uF value of C2 can be increased. The very cute completed PCB is shown in Photo of Switchmode HeNe Laser Power Supply Ripple Reducer 1. It's about 1.7" x 0.7" x 0.75" (WxDxH). The HV input is at the upper right, the HV output is at the lower right, and HV ground is between the large HV capacitors at the left. There's one extra resistor at the output to limit current in the event of a, well, unfortunate accident. :) The only downside to using this circuit is that it drops about 75 V (at 3.5 mA, proportionally more at higher current), so the supply must be capable of handling the additional output voltage. (Though some adjustment in component values can cut this more than in half with minimal impact on performance.) The input RC reduces the ripple by approximately 50 percent while the transistor circuit looks like a resistor of under 10K ohms for DC, but a much higher impedance for AC since C2 maintains a constant voltage on the base of Q1 over any time scale that matters at the 25 kHz switching frequency of the supply. The NE2 seems to provide adequate protection for Q1, which has survived after a rather large amount of abuse, except when accidentally installed backwards. :( :) I'm not sure how the Melles Griot supply deals with the starting transient. It looks like Q7 there should be obliterated the first time a tube starts. The NE2 in my circuit always flashes at startup and sometimgs one or more times at shutdown indicating momentary high voltage across Q1.
This circuit should work for a tube current up to at least 7 mA with one minor change: Since the DC voltage across the circuit may exceed 60 V (the sustaining voltage for a single NE2) for higher current, it may be necessary to use a pair of NE2s in series. Or, the value of R2 can be reduced to say, 150K ohms to keep the voltage drop below the 60 V sustaning voltage of the NE2, with a very very slight, likely undetectable increase in ripple. Alternatives to the MPSU10 for Q1 include the MJE3439 and 2N3439, as well as the 2SC2271E. The MPSU10 was handy. Nothing is critical. Feel free to experiment but beware of the high voltage on the capacitors even after powering off - discharge with 100K to 1M ohms to ground before touching anything! (The NE2 and LED will flash to indicate discharge is proceeding.)
With the circuit installed, the capacitor isolated HV AC scope probe shows 50 mV p-p of ripple at the top of the 108K ohm ballast resistor driving an 05-LHR-006 laser tube. This ccorresponds to less than 1 uA of current ripple assuming a net positive tube+ballast resistance of 50K ohms. (The actual net ballast resistance is probably slightly higher, which means the ripple is even lower.) Without the ripple reducer, the voltage ripple was 3.7 V p-p for 74 uA of current ripple.
The neon lamp provides a diagnostic function in addition to protecting the transistor. If the circuit is operating normally, it should flash on briefly when the laser tube starts, but then die out over a fraction of a second. If it never flashes at all, or stays on continuously, the transistor has probably been damaged. Even with the NE2 and zener diode for protection, I can't guarantee that really hard-to-start cantankerous lasers won't do bad things.
Of course, if the ripple is still too high, it would be possible to use two such circuits in series. :)
(I have assembled and tested PCBs for these available via Sam's Classified Page.)
However, all modern HeNe lasers are DC-excited and some degree of AM modulation is possible by controlling the discharge (tube) current. However, no modulation scheme of this type will permit a continuous (100%) variation from full on to full off. Depending on the particular power supply/ballast resistor/HeNe tube combination, going below a threshold current - perhaps 50% of the optimal current - will result in the tube going out and needing to be restarted, producing a discontinuity in the beam. The exact dropout current will depend on a combination of the power supply voltage, ballast resistor, wiring and tube capacitance, and tube characteristics. The last one - tube characteristics - change with use over the life of the tube. In some cases, higher mileage tubes in particular will have a dropout current very near the operating current so going lower isn't possible. (And some tubes fail as a result of the dropout current intersecting the operating current, at which point they start flashing on their own!)
The incremental sensitivity of optical output power to changes in tube current is also not linear and always less than 1:1 over the useful range of a stable discharge. Well below the optimal current specification, it is more like 1:2 to 1:5 going to 0 at the point of peak optical output (optimal current). Then, the modulation response will turn negative as increasing the tube current actually *reduces* optical output.
Modulating a HeNe laser beam by controlling the input to the a commercial power supply will probably not work due to the filtering in the power supply output and possible regulation against just such input voltage variations.
However, if you have built one yourself, at least *adjusting* power level in this way is probably possible though any output filtering will severely limit frequency response.
Usually, electronic modulation is accomplished by adding a circuit to the return (cathode circuit) of the HeNe tube to control tube current. In this respect, regulators and modulators may be combined.
It may also be possible to pulse the power to the HeNe tube (at least for low power lasers). See the section: Pulse Type Drive and Modulation of HeNe Tubes.
However, mechanical and electro-optic approaches do have their advantages including a potentially much wider modulation intensity range (from full off to full on, which, as noted, is not really possible with electronic techniques controlling tube current). They also do not require any modifications to the HeNe laser power supply and are thus necessary if this is not possible or desirable.
Several suggestions for implementing power supply output side modulation are outlined below. However, the following should be kept in mind:
For example, a tube rated for 4 mA may actually only be usable within a 3 to 4 mA range. The low limit depends to some extent on the power supply voltage and ballast resistor value - a higher voltage (and ballast resistor) will permit the tube to be stable at a lower current. However, there is still a "dropout" current for the tube below which it is impossible to run stably. The high limit will be dependant on when the gain becomes negative - the output decreases with increasing current.
Thus, the peak current provided by these circuits should be set less than or equal to the maximum operating current for your tube. This involves selection of the ballast resistor and power supply and/or regulator current set-point.
However, note that the actual output power at high currents is not very stable (the optical gain is too low) - it will vary as the tube heats. Thus, unless a closed loop control scheme is implemented using optical feedback, this approach may not be acceptable. There will also be added higher frequency optical noise.
Where wide range (approaching 100 percent) modulation is required, external electro optical or electro mechanical modulation may be a more effective solution.
Depending on the type of information you would like to transmit over the beam and its bandwidth or data rate, dual tone or some kind of frequency or phase modulation may be better than simple amplitude modulation.
With an unregulated HeNe laser power supply, there are no issues with fighting an active regulator by attempting to control tube current. However, if doing this with a modern regulated HeNe laser power supply, care must be taken to make sure the frequency range of the modulation is beyond the loop response of the regulator or else they will be fighting and bad things can happen. This usually means the modulation circuit should be AC coupled with a low frequency cutoff. The quiescent current should be set somewhere in the middle of the range over which the discharge is stable..
Note: Some of the following circuits have not been fully tested and tweaking may be necessary.
HeNe tube Rb Tube+ +------------+ Tube- Laser PS+ o---------/\/\----------|- |-|-------+----+ +------------+ | _|_ (see note | - below.) +---+ ||( o---------+ ||( )||( Input from LV )||( HV audio amp )||( )||( Small audio or o---------+ ||( power transformer ||( +---+ | Laser PS- o--------------------------------------------+CAUTION: If your tube is mounted with its cathode connected to a metal case as is the case with typical commercial laser heads, earth ground must be tied to Tube- for safety as Tube- may be at a potential of several hundred volts with respect to HV-. If the tube is enclosed and insulated from the user, this is not necessary.
The circuit below is designed for a 4 mA operating current. With R3 set at 200 ohms, a 1 V p-p modulation input will vary the tube current from 1.5 to 6.5 mA. However, the range of your tube may be much more restricted than this depending on your tube, power supply voltage, and ballast resistor value.
HeNe tube Rb Tube+ +------------+ Tube- Laser PS+ o---------/\/\----------|- |-|-------+ +------------+ | R1 | +---/\/\---+-------+ | 100K | | C1 10uF | |/ C | + o------)|----+--------| Q1 _|_, - + | |\ E '/_\ ZD1 Line level | MPSW42 | | 120V audio / / ^ _|_, R2 \ R3 \<-+ '/_\ ZD2 - o 1.5K / 500 / | | 120V | | | | | Laser PS- o--------------------+------------+----------+--+----+CAUTION: Do not use this circuit with a tube whose metal housing is not totally insulated from the user since the cathode will have a potential of several hundred volts PS- (and the signal return). For such tubes, see the alternative circuit below and the section: Enhancements to AT-PS1 for some possibilities.
The use of a transistor like an MJE3439 or MJE2360 would permit a wider range of control but 200 V is likely to be sufficient. One of these would also be needed for higher current tubes since the MPSW42 is only rated at about 1 W (100 V * 4 mA is .4 W). A vacuum tube triode or high voltage FET in a similar configuration could even be used. The tube would be kind of quaint. :-) See the section: Modulation Using a Vacuum Tube for details.
The Vceo rating for such a transistor may need to be several hundred volts depending on your HeNe tube and power supply, ballast resistor value, and how much modulation is desired. The voltage divider formed by R1 and R2 and emitter resistor R3 set the operating current and voltage across the transistor with no signal.
A small HeNe tube will have perhaps 1,200 V across it and 300 V across the ballast resistor at the recommended operating current. The voltage across the tube (a negative resistance discharge device), will be more or less constant around the operating current (but may increase as the current is reduced). It is the 300 V that you have to play with. If the transistor drops 150 V, the ballast resistor will only drop 150 V and the current will be cut in half.
The compliance range for the circuit shown is about 200 V so a ballast resistor still needs to be selected to enable the regulation to be effective from 0 to maximum modulation.
Use the procedure described in the sections starting with: Selecting the Ballast Resistor to determine an appropriate ballast resistor value for the desired tube current and a power supply voltage 100 V less than that of your (unregulated) supply. This circuit should then maintain that current constant with no audio input and permit modulation of the current on either side of the set-point up to a change of approximately +/-100/Rb mA.
HeNe tube Rb Tube+ +------------+ Tube- Laser PS+ o---------/\/\----------|- |-|-------+----+ +------------+ | _|_ | - + o------------+----------+--+----+ | | | | / / | | Line level R2 \ R3 \<-+ _|_, audio 1.5K / 500 / v '/_\ ZD1 | | | 120V C1 10uF | |/ E | - o------|(----+--------| Q1 _|_, + - | |\ C '/_\ ZD2 R1 / MPSW92 | | 120V 100K \ (PNP) | | / | | | | | Laser PS- o---------------------------------+----------+-------+This circuit uses a PNP transistor to permit the signal input to be near the laser tube cathode and earth ground. Otherwise, it operates identically to the previous one.
If I recall anything about tubes (which was a long time ago) select a cathode resistor to bias the grid negative to set the laser tube operating current (less than its maximum) and then drive the grid from a voltage source like the speaker output of an audio amplifier. The voltage on the plate will then vary resulting in an incremental current change of roughly: delta(Vin)/TC where TC is the transconductance of the vacuum tube at its operating point.
For example, using a 12BH7A medium mu twin triode (you can use the other half for a fabulous audio preamp!), the grid voltage will need to be in the range -0 to -30 V for a compliance range of 400 V. At a plate voltage of 250 V, TC is 3,100 umhos resulting in an overall sensitivity of about 3.1 mA/V. Thus, this circuit would work quite nicely with line level audio signals since a total variation of only 1 to 3 mA should be needed to modulate a typical HeNe laser tube.
In the circuit below, R2 and R3 set the bias point depending on the current requirements of your HeNe tube, power supply output voltage, and ballast resistor. C2 is the cathode bypass capacitor (low frequency roll off of around 20 Hz at the minimum (higher current) setting of R3).
HeNe tube Rb Tube+ +------------+ Tube- Laser PS+ o-----/\/\--------|- |-|------+----+ +------------+ | _|_ |P - C1 --- Signal in o------||------+------------------- - - - G ET1 .1uF | .---. K 1/2 12BH7A 500V | +------+--+ ^ | | | | | R1 / | R2 / F F 1M \ | 470 \ (6.3 VAC or 12.6 VAC / C2 +_|_ / depending on connections) | 10uF --- | | 100V - | R3 / | | 15K \<-+ | | / | | | | | Laser PS- o--------------+----------+------+--+CAUTION: If your tube is mounted with its cathode connected to a metal case as is the case with typical commercial laser heads, earth ground must be tied to Tube- for safety as Tube- may be at a potential of several hundred volts with respect to HV-. If the tube is enclosed and insulated from the user, this is not necessary.
Where to find a 12BH7A? Actually finding a new one may be difficult and expensive (I have heard of a retail price of around $50 - wow! In its day, that tube probably went for $2 - and most of the difference isn't due to inflation.) I just selected the 12BH7A (out of the RCA Receiving Tube Manual, 1970 edition) as a triode with a suitable maximum plate voltage rating. Many many other tube types (likely cheaper also - the 12BH7A seems to be even more expensive than other similar tubes for some reason - will work as well. Or, just go to a hamfest or electronics flea market - people are always unloading old tubes. Or, check out a few garage or yard sales - you can probably pick up several old (but not yet antique) tube type radios for less than the cost of a single new vacuum tube!
Of course, if you really want to impress your friends, use a transmitting tube like a 6146, 1625, 811, or even an Eimac 3-1000Z - that's good for a kW!
A high voltage power supply that produces short pulses exceeding the tube's starting requirements will result in output pulses of light. With appropriate design, pulse duration and repetition rate can be controlled. Depending on needs, a flyback transformer or automotive ignition coil may be suitable. The stability of the beam on such short times scales may be questionable and high repetition rate pulsing may result in shortened tube life as well. See the section: Pulse Type Drive and Modulation of HeNe Tubes.
As with modulation, mechanical or electro-optic techniques are also possible. A simple chopper wheel in the beam path may be perfectly adequate for your needs and a lot less complex and costly than high tech alternatives!
HeNe tubes can be pulsed - there is no law that says DC must be used.
However, to achieve rated (or even just reasonable) tube life and output power, discharge current must be controlled within specified limits. This may not be quite as trivial as suggested below - especially for larger HeNe tubes!
WARNING: The comments below apply to the low power (0.5 to 1 mW) HeNe lasers found in bar code (UPC) scanners and similar products. If you attempt this with a 35 mW HeNe laser, you may blow the tube, burn out the power supply, fry yourself (or all of the above) in the process.
(Portions of the following from: K. Sorter (email@example.com).)
Modulation methods are a primary discussion amongst our group although many are now converting to IR (invisible) lasers and IR sensitive PM tubes and APDs for long range work (Mia and Rayleigh scattering/adsorption effects are much less severe in the atmosphere at IR).
High index of amplitude modulation is not possible by smoothing varying the input to a HeNe tube. As soon as the the supply voltage/current is reduced slightly, the tube winks out. However, it was found that the HeNe tubes could be pulsed at about 30 khz by simply removing the filter capacitors in the small switching power supplies used with low power HeNe lasers (yes, they were not potted!) Presto - the units had a 100 percent index of modulation! These supplies and tubes are common and cheap in rejected scanner units that are upgraded to diode lasers.
Therefore, to get the highest index of modulation (and therefore the most miles per watt), you will need to switch the tube on and off at 20 to 50 kHz. This means running the tube on AC or pulsating DC. Unfortunately, there are a couple of issues that may prevent the successful modification of surplus HeNe power supplies:
This was followed by a brief discussion in which the question of switching the laser on and off (a destructive process) was discussed, and it was determined that there was a big difference between keeping the plasma lit while pulsing the laser on and off AND turning the tube completely on and off and restriking the plasma (as in a cold start).
Note that the tube continues to have internal plasma during the 'off periods', but that the tubes didn't lase until the pulse came along and the input voltage got high enough. A photodiode placed against the tube envelope showed constant output all the time with higher levels when the tube lased. The output of the tube however would go completely on and off (at the output mirror of the tube).
IMPORTANT: In this type of operation, the starting pulse is only applied initially to start the process-if the starting pulse was applied 30 or 40 thousand times per second, the tube would be destroyed in short order.
Some tube manufacturers say that pulses are OK, others suggest raw AC with appropriate current limiting resistors. At the prices of these things (as in: dirt cheap), experimentation is encouraged!
You can also use the laser in relaxation oscillator mode by simply putting a capacitor across it. It will oscillate at a fairly constant frequency, although modifying that frequency (for voice/data transmission) is not so easy. It is useful if you need a constant monotone output. However, whether this will lead to significantly reduced tube life will depend on the actual frequency of the relaxation oscillator as well as the type of starting circuit used in the power supply. Above a frequency of perhaps several kHz, the discharge will restrike without requiring the normal high starting voltage (and the power supply's starting circuit will not be activated). If the time constant is too long (too low a frequency), the tube would need to cold start each time. Repetitive application of a high energy starting pulse may result in rapid degradation of the tube characteristics.
I (K. Sorter) used a .05 uF disk capacitor and ran my 2 mW HeNe tube in that mode for 2 weeks constantly. The starter pulse in the power supply was activated by a relay, so each time I would turn on the supply, it would click. There would be no clicking while running in relaxation oscillator mode. Thus, the power supply thought the tube was running normally. The purist might believe this would lead to degraded life expectancy of the tube, but it hasn't been proven in actual real life.
You then FM modulate the 20 kHz frequency with your desired audio. There are various ways of doing this by injecting the audio signal into the frequency determining circuitry of the pulse width regulator chip. This specifics will, of course, depend on the design of your particular power supplies.
The receiver is tuned to the carrier frequency and demodulates the FM signal. This can be easily accomplished with a NE556 chip and photodiode/low noise op-amp front-end.
The best deal in optical front ends these days is the Burr-Brown OPA210/OPA211 op amp/photodiode in a chip combo at 8 bucks! You add an appropriate external feedback resistor (depending on the desired bandwidth), and you have the whole front-end in a simple chip!
It should be noted that FM is a poor performer if your goal is long distance communications! But it certainly can have near broadcast quality audio which is a big advantage if that's your goal.
We have many users that are using this strategy in communications everyday.
For more information and discussions on amateur laser communications, join the Laser Reflector. See the section: Amateur Laser Communications for details.
A typical circuit is shown below where the characteristics of the tubes are sufficiently similar that separate current regulators are not required.
Ri1 S1 Start+ o---+---/\/\---+ LT1 o + Test1 - o | | Rb1 Tube1+ +----------+ Tube1- | Rs1 | HV+ o---+--|---|>|----+---/\/\---------|- |-|--------+---/\/\----+--+ | | CR2 +----------+ 1K | | | | | | Ri2 S2 | | +---/\/\---+ LT2 o + Test2 - o | | | Rb2 Tube2+ +----------+ Tube2- | Rs2 | | +------|>|----+---/\/\---------|- |-|--------+---/\/\----+--+ CR2 +----------+ 1K | | +----------------------------+ | HV- o-------------------| Optional Current Regulator |--------------------+ +----------------------------+ _|_ -(Note that separate cathode feeds probably will not exist on a center-tapped discharge tube.)
Rb1 and Rb2 are the ballast resistors. Where no separate regulators are used, one or both of these can include a suitable variable resistor and/or trimming resistors can be added to balance the currents. Just be sure to use a well insulated tool or adjust with power off and capacitors confirmed to be discharged!
This approach can easily be extended to more than two HeNe tubes.
The starter can be a voltage multiplier or inverter type. Note that since the HeNe tubes will not start at exactly the same time, Ri1 and Ri2 (the HV isolating resistors) must have a high enough resistance so that the starter supply is not loaded down excessively even when all but the last HeNe tube has started. It may also be possible to use a large pulse type starter applied to points S1 and S2 via HV coupling capacitors.
It is desirable to connect the return (-) of the starter supply to HV+ to take advantage of the additional boost provided by the operating voltage supply. Where the return cannot be isolated, it will have to be connected to ground. CR1 and CR2 are HV diodes to bypass operating current around the starters once the discharge is initiated.
+-----------------+-----o Start Pulse LT1 ____|____ | Rb1 Tube1+ +---------------+ Tube1- | HV+ o--+---/\/\---------|- |-|---------|---+ | +---------------+ | | | | | | +-----------------+ | | LT2 ____|____ | | +---------------+ | +---/\/\---------|- |-|-------------+ Rb2 Tube2+ +---------------+ Tube2- | | o - Test + o | +----------------------------+ | Rs | | HV- o---| Optional Current Regulator |---+---/\/\---+--+ +----------------------------+ 1K _|_ -Rb1 and Rb2 are the ballast resistors. Where no separate regulators are used, one or both of these can include a suitable variable resistor and/or trimming resistors can be added to balance the currents. Just be sure to use a well insulated tool or adjust with power off and capacitors confirmed to be discharged!
If the tube cathodes can be isolated from ground, current monitoring can be accomplished by adding sense resistors (Rs1 and Rs2) and measuring voltage at Test1 and Test2 (1 mA/V) or with a 10 or 20 mA full scale current meter to these. If this is not possible, one alternative is to use a single sense resistor (Rs) and run only one HeNe tube at a time while adjusting current making sure that the input voltage is identical in each case.