“DK Plus”

A Second Simple Nitrogen Laser
Using “Doorknob” Capacitors
in a Voltage-Doubling Circuit,
Primarily Intended as a Research Tool

Abstract:

This page presents the design and construction of a nitrogen laser that can be built by a Do-It-Yourselfer. It is a scaled-up version of a smaller and simpler previous design.

The laser is constructed, as much as possible, from materials that are available at hardware stores and hobby shops: plastic “lumber”, aluminum rulers, steel rods, compression fittings, pieces of basswood and spruce, etc. I used a commercial spark gap, but it is eminently feasible to build one. The two items used in this design that are not easy for a do-it-yourselfer to construct are the “doorknob” capacitors, which are available from time to time on eBay; and the fused silica windows, likewise an eBay item, though it is possible to use fused silica microscope slides, which I buy from EMS.

This laser uses either nitrogen or a mixture of nitrogen and helium at low pressure (roughly 30-90 Torr each), runs at 25-30 kV, and puts out pulses of more than 1.6 millijoules, lasting ~8 nsec FWHM. This corresponds to average pulse power of 160 kW and (assuming a vaguely sinusoidal pulse shape) peak power of roughly 230 kW. It is enough power to make small sparks if you focus the beam onto a metal surface:

[For anyone who is reading this before the prototype is completed & tested, those numbers are placeholders that I have achieved with the prototype, and will be replaced by higher numbers when I get the prototype optimized better.]



CAUTION

Construction and operation of this laser (or any high-performance pulsed laser, particularly one that emits output that is invisible) should be attempted only by people who have been trained to handle lasers and high voltages. The term “training” includes relevant and appropriate DIY experience; but if you have never built a working nitrogen or TEA CO2 laser before, this is not the right place to start. Even the previous design, referred to above, is a bit difficult for a beginner.



General Plan

(30 July, 2005; 4 August, 2005; 7 October, 2005; May and June and September, 2006; October 2009.)

This laser is intended to operate at reduced pressure, though there is some chance that it will work with a full atmosphere of helium inside the head. The channel is roughly 3 mm deep (the actual depth depends on discharge parameters, as well as the width of the electrodes) and about 22 mm wide. It has active length of about 600 mm, but the electrodes are about 914 mm long (this may help to reduce the amount of sparking at the electrode ends). The overall length of the head is approximately 950 mm.

The switch is an EG&G GP-32B spark gap, which is rated up to at least 30 kV in air. Although such gaps are available on eBay from time to time, it is eminently possible to construct your own triggered spark gap if you cannot buy one or even simply prefer not to. I must, however, stress the fact that a triggered gap with low inductance is almost certainly required for good performance. Any gap with excessive inductance is (in the words of an earlier era) unlikely to give satisfaction.

Here is the schematic diagram of the laser:

The preionizer is a coating of 100-mesh silicon carbide powder embedded in a thin layer of epoxy on the inner surface of the upper sidewall of the head. It is connected to the electrodes by steel rods that also serve to space the upper sidewall away from the electrodes. I have indicated a charging inductor here, but you may find that a resistor works better; it is a good idea to try it both ways.



Materials and Construction

Here are some of the things you will need if you want to build this laser:

(Two aluminum rulers, 1/8" thick and at least 35" long; two wooden yardsticks, about an inch and a half wide; two pieces of spruce from the hobby shop, 3/16" square; two more pieces of spruce, either 1/4" square or, if you can’t find that size, 1/4" x 3/8" or even 1/4" x 1/2"; one or more tubes of silicone rubber caulk. I have found that the “XST” paintable version, shown in the photo, works better for me than other versions I’ve tried.)

In addition to what you can see here, you will need a brass kickplate from the hardware store (I used a large one, 8" x 34"); 16 laser-grade “doorknob” capacitors (I used capacitors rated 2000 pf and 40 kV); some sort of spark gap switch and a trigger circuit to operate it (I used an EG&G/PerkinElmer GP-32B, which I got on eBay, but you can certainly make your own gap); some plastic quick-connects for gas and vacuum (you can use metal fittings if you can’t find plastic ones); a pair of fused silica windows; a mirror; some silicon carbide grit to make a preionizer (unless you decide to build a more traditional preionizer, a subject I will try to discuss later, or you can email me if you have questions); various nuts and bolts; one or two types of epoxy cement; and a certain amount of high-voltage insulating varnish. Also, of course, various hand tools, preferably including a jeweler’s saw, a high-speed rotary tool, a set of small files, an X-Acto™ knife or equivalent, and so on.

Oh, yes: you will also need a high voltage power supply and a vacuum pump of some sort, if you actually expect to operate the laser.



Notice that the electrodes are aluminum rulers, which will be cut down from their original 48" length to 35". These provide both good stiffness and a clean straight edge, though you have to choose them with some care — I have seen many bent rulers at hardware stores during the development of this design.

Even if you choose them well, they may not provide the optimum edge profile; different researchers have obtained varying results that seem to depend on obscure details of design, and it is very difficult to predict in advance what the optimum edge shape will be. In any case, it turns out to be important to remove the anodizing (or paint) from the working edge. Anodizing creates a tough surface coating, which is difficult to remove; if you scratch up the edge in the process of getting it off, you will probably want to polish it smooth again. You can run through increasingly fine sandpaper until you get to at least 1,000 grit, which is available at auto-parts stores, or you can get up to 600 or 800 grit and then either stay with a satin finish, or rub with fine steel wool. You will need to sand again if you scratch the edge up in the process of removing paint or rounding the corners.

You will also have to remove the paint or anodizing at the places where the ruler makes contact with the doorknob capacitors. I used a handheld rotary tool (there are several common brands, all of which can rotate quite quickly) with a wire brush wheel. This is fine for the capacitor contact areas, but I am worried that it may be too uneven for the edges, so I didn’t use it there (see photo, when I get a chance to post it). The paint on these rulers, btw, was so thick that making the one pair of electrodes used up the wire brush wheel, and I will have to get another one.

It is crucially important that you make sure there are no gaps in whatever sealant you use to attach the electrodes to the spacers, and the spacers to the sidewalls. It is impressively difficult to find and seal all of the leaks in a new head, and sealant gaps just make things worse. When you have assembled the head, you should test for vacuum leaks (and fix them, and retest) before you connect it to the capacitors on the baseplate — it is a lot easier to reach the underside with the head unattached. After you have made sure it is sealed, if you intend to operate the laser at voltages above about 20 kV, you need to paint high-voltage insulating varnish on the baseplate (visible in the photo that I will add when I get a chance) and on the undersides of the electrodes, to prevent flashovers. Remember to avoid insulating the areas where capacitors will be attached (!).

Test the head again after you insulate it and attach it to the baseplate, as the stresses involved could open old leaks or create new ones.

Because nitrogen lasers need to be fast (see my page about the issues involved if you want more information about this), you should use broad foils to connect the cathode and the baseplate to the spark gap switch. It is best to cover the full width of the doorknob array, but I found that I could get away with 12" and still achieve reasonable performance. The switch itself should have the lowest possible inductance, which means that it should have wide electrodes. I use commercial spark gaps for convenience, typically EG&G GP-15B for voltages of at least 25 kV, GP-32B for voltages of about 15-30 kV, and GP-14B for voltages of 12-24 kV. For smaller lasers, GP-70 gaps are suitable for 12-24 kV. These gaps are currently manufactured by Perkin-Elmer, and may be sold as either EG&G or PE. My prototype of this laser used a GP-32B.



Views

Here are some photos of my prototype head, first parts during construction and then discharge examples. These are not the least bit optimized, and the examples are not photos of the final version; my apologies.

                       

Here is a view of the discharge in the prototype, before I changed from a glass “roof” to a wooden one:

Here is a view of the laser pumping a small dye laser, using a homemade cuvette:

(Again, this is an earlier version of the prototype, with considerably less output power than the final one. Even so, it had no trouble pumping the dye.)



Initial Testing

(19 September, 2006)

When you have the laser assembled and relatively free of leaks, you can begin testing.



Sidebar:
A Word About Power Supplies

(11 June, 2006)

Pulse-discharge devices of this sort produce lots of electrical noise, some of which travels back up the wires to the supply. You need to do some serious bypassing, and put some good chokes in the line, not only on the hot side but also on the ground side.

There is also the issue of rectifiers. Let me run some numbers past you, to make this clear.

My transformer is rated 10 mA at 37.5 kV. I am using half-wave rectification, so I can expect to get a maximum of 5 mA out. HV rectifiers are sufficiently expensive that I would be willing to live with this in any case; but as it happens, one end of the secondary is grounded, so I don’t have much choice.

So: how much voltage-handling capability is enough?

The RMS (average) output of the supply at full input voltage is 37.5 kV. This means that the peak voltage is 37.5 times the square root of 2 (which is about 1.414). Let’s round the result off to 52.5 kV.

Every cycle, however, involves a positive half-cycle and a negative half-cycle, so we must double the Peak Inverse Voltage rating of the rectifier. This brings us to 105 kV.

That, however, provides absolutely no safety factor. One little spike at the wrong time, and you risk destroying your rectifiers. Let’s double it again, to provide nominal safety. Now we are at 210 kV. That seems fairly reasonable, but it actually isn’t really very much. If you are using it to operate a pulsed laser, as I am, you need to put some sort of filter into the output lead, to block the spikes that the laser generates when you fire it.

(Some time later)

I found some 80-kV rectifiers on eBay, and bought three of them, which gives me 240 kV peak inverse rating. With chokes and bypass caps in the output line, this appears to be good enough. I have run the power supply all the way up to its maximum output, roughly 38 kV, and it has behaved perfectly, even powering an 89-nf energy-storage cap connected to a nitrogen laser, which is a strong source of EMP — here’s what it did to our Tektronix 7104 oscilloscope, which was perhaps 2 meters away at the time:




This work is supported by
the Joss Research Institute
19 Main St.
Laurel  MD  20707-4303   USA



Contact Information:

My email address is a@b.com, where a is my first name (just jon, only 3 letters, no “h”), and b is joss.

My phone number is +1 240 604 4495.

Last modified: Mon Sep 23 14:05:27 EDT 2013