Dual-Discharge Current Regulator Design and Construction Notes.
I designed and built this dual-discharge current regulator for a couple of reasons. First, the unregulated DC power supply that I had built in 1980 (along with the CO2 laser) was a bit underpowered, and the ballast resistors wasted so much of the available voltage that I could not get stable operation at currents below about 30 mA. Second, I was thinking of building a new three-electrode laser that could at least double the power output using operating voltages no greater than those across my existing laser.
Active current regulation is much more efficient than ballast regulation, primarily because it can operate with a relatively low fixed voltage across it. More of the overall power supply voltage is therefore available to operate the laser discharge. In addition, negative feedback ensures that the set current is highly stable; remaining constant even with large changes in supply voltage, discharge voltage, gas pressure, gas composition, and temperature.
The basic regulator circuit concept is simple:
This is a “cascode” connection of a power field-effect transistor (the actual constant-current source) and a large, high-voltage vacuum triode. The triode drops the excess voltage (that is, any difference between +HV and the voltage across the laser) while passing the laser current on to the drain terminal of the FET. The operational amplifier ensures that the voltage across the resistor R is equal to VC.
Then, the laser discharge current I = VC/R. That’s it! You vary the current by varying VC.
My unit contains two regulator circuits as shown above, together with meter circuits, external modulation inputs, protection circuits, and some convenience features.
The following description refers to regulator A, one “channel” of the unit. Regulator B functions identically.
The control voltage VC is the sum of two inputs; one of them being the manual setting on the LASER CURRENT/PULSE BASELINE A knob and the other one a signal applied to the EXT PULSE INPUT A connector. The knob has a 10-turn dial calibrated directly in mA.
The external input is specified as a “pulse” on the front panel marking, but it may be a continuously-variable analogue signal. The input’s sensitivity is 16.6 mA/volt; it has this “odd” value because current source resistor R78 is a 60 ohm part rather than, say, 10 or 100 ohms. I designed the circuit using parts I had in my junk-box!
The inputs are additive. For example, if the knob is set for 10 mA and + 2V is applied to the external input, the actual current will be 43.2 mA. In this way, a laser tube may idle at some small quiescent current and then either be pulsed or turned on continuously using the external input.
Op-amps U4A and U4D buffer the inputs, and U4B performs the addition. U4C is the feedback element in the current regulator itself.
The “gate-stop” resistor R82, in conjunction with the input capacitance of the FET Q1, introduces a low-pass filter component that prevents the regulator circuit from oscillating at high frequencies.
The triode is an Eimac 3-500Z, which can dissipate 500 watts at plate potentials greater than 10 kV. It requires a 5V, 14.5A AC source for heating its thoriated-tungsten filament cathode.
The currents that may be regulated in this circuit range from a few hundred microamperes at the low end (the voltage across the gas discharge soars at low currents, rapidly overtaking the available power supply voltage) to over 80 mA at the high end (as determined by the tube characteristics, at that operating point where the difference in potential between the cathode and the grid is zero). So, while this device has been designed to drive a big CO2 laser, it can also run the tiniest little He-Ne laser!
The voltages on the laser anode and cathode connectors are reduced by a factor of 2000 by voltage dividers, such as the one consisting of the resistors R75 and R23 and trimmer potentiometer R22. Note that the input resistors in each of these dividers must be rated to withstand the maximum input voltage of 20kV. I used TRW RB-4 parts.
The laser anode (input) voltage is measured directly. The buffered voltage from the divider is converted to a current by op-amp U6C and applied to analogue meter M1, which is calibrated directly in kV. The discharge voltage is measured by subtracting the laser cathode voltage from the anode voltage in difference amplifier U7C.
The discharge current is proportional to the voltage across R78.
While there are two separate regulator circuits, there is only one set of meters. Relay K1 determines which regulator’s voltage and current to display, as selected by the METER A/B switch.
I originally used analogue switches (CD4066 CMOS transmission gates) for this meter-select function, but they proved to be unreliable and so I replaced them with the relay.
U10 and its associated components comprise a filament warm-up timer. When the FILAMENT A switch is thrown, power is applied both to the filament of V1 and to this circuit, which is a simple astable multivibrator. It causes the FILAMENT A lamp to blink at approximately two cycles per second, indicating that the filament is warming. Meanwhile, capacitor C25 charges down through resistor R42, and after about 30 seconds it pulls down the /RESET pin on U10. The lamp then stops blinking and remains on continuously, signaling the user that it is now safe to switch on the high voltage.
I designed and built this circuit before selecting the 3-500Z as the current regulator triode. This particular tube warms up in seconds, and so the timer circuit is not even necessary.
U11 is an analogue multiplier which evaluates the product of V1’s plate voltage and current. If the value exceeds 500 watts, as determined by comparator U14D, the circuit illuminates the OVERLOAD A lamp to indicate that the tube’s power rating has been exceeded.
The multiplier circuit is somewhat of a luxury and the chip itself, a Motorola MC1494, is an obsolete part that requires lots of cumbersome external components (such as the three nulling potentiometers) to make it work. For any new design, I would recommend either using one of the newer “laser-trimmed” multiplier components, or performing the calculation digitally. A microprocessor having a built-in analogue-to-digital converter, such as the PIC16C74, could easily do this and all of the measurement and display functions, too.
The actual HV DC power supply is external. It should be equipped with a voltage control (such as a variable transformer) so that after the laser is operating at the desired current, the potential across the regulator tube may be adjusted to a minimum (around 1-2 kV).
The HIGH VOLTAGE and LASER START switches are intended to control external relays. When actuated, they apply + 24V to their respective terminals on the rear panel.
In my system, I have rigged the HIGH VOLTAGE switch to turn on the HV power supply itself. The LASER START switch simply bypasses the variable transformer; this way, I can leave the variable transformer at its optimum setting and still obtain the power supply’s maximum voltage for striking the laser discharge (strike voltage for a gas discharge is always higher than the sustaining voltage). Another possible approach would be for the switch to actuate some kind of firing circuit, such as a series-connected high-voltage transformer of the sort used to trigger xenon flash-lamps.
The unit is built on a big chassis, which I then mounted behind the front panel using long spacers. Large and/or heavy components (DC power supplies, relays, transformers, etc.) are mounted inside and on top of the chassis, and the electronic circuits are mounted on a perforated board.
Big power tubes like the 3-500Z are somewhat difficult to mount, and they have stringent cooling requirements. The manufacturer recommends using special "air-system" sockets and glass chimneys which, together, can cost as much as the tubes themselves! I improvised somewhat using more conventional Johnson-style ceramic tube sockets. I mounted them below the top of the chassis, leaving large openings for ventilation. A squirrel-cage blower pressurizes the inside of the chassis, and the cooling air flows up through the socket openings and chimneys.
Remarkably, I found that it was less expensive for me to have the tube chimneys made by a local glass-blower than to order them from Eimac!
I needed several high-voltage connectors for this project, each capable of working reliably at 20 kV. Rather than buying them, I made them myself, using PVC pipe, PVC sheet, ordinary banana jacks, and SO-239 (UHF coaxial) receptacles. The body of the connector is made from the PVC pipe, with the banana jack mounted at the rear in a disk cut from the PVC sheet (see the photos). Another disk, at the front, serves as a mounting flange. I removed the insulator and center contact from the SO-239 (easily accomplished by grinding the back off of it) and then mounted it in front of the panel, with the HV connector body behind the panel.
If you cannot locate PVC sheet, it may be easily prepared by longitudinally slitting a length of pipe and then gently heating it in an oven to make it open up and lie flat!
I fashioned the mating cable from RG-8 coax and a PL-259 UHF plug. After removing the center pin, I drilled out the PL-259 so that the insulated center conductor of the coax could pass through it; then I attached the PL-259 shell to the coax shield in the conventional manner, with the appropriate length of center conductor protruding. I soldered an uninsulated banana plug onto the end of the center conductor.
When the cable assembly is mated to the HV connector on the chassis, the banana plug engages the jack at the rear of the connector.
If you use this technique, be sure to obtain RG-8 cable having a hard polyethylene dielectric. There are many cables, otherwise similar to RG-8, that are made with foam dielectrics that cannot withstand the high voltage. I have seen the polyethylene cable used reliably in commercial 40 kV power supplies.
The ARRL Handbook for Radio Amateurs. Newington: The American Radio Relay League (published annually).
This useful reference contains a lot of general design and construction data. It has a detailed discussion of the mounting and operation of large power tubes, such as the 3-500Z that I used in my circuit.
The contents of this handbook often change each year, so I keep a large collection of back-issues. Old copies of the ARRL handbook are most easily found at local “hamfests” (swap-meets conducted by radio amateurs).
Paul Horowitz and Winfield Hill, The Art of Electronics, 2nd Edition. Cambridge: Cambridge University Press (1989).
This is my favorite electronic design text. It was intended for use by scientific researchers, and is totally and thoroughly practical.
The second edition of “Horowitz and Hill” emerged in 1989. I feel that the book is a bit overdue for a third edition, but the authors have expressed no interest in revising it as yet.
Michael J. Posakony, “A High Voltage Current Regulator for Laser Gas Discharge Tubes,” The Review of Scientific Instruments Vol. 43, No. 2, pp. 270-273 (February 1972).
I based my regulator design on the ideas presented in this paper.