Joss Institute Projects:

A Straightforward TEA Nitrogen Laser for the Do-It-Yourselfer

(A “My First Laser” Project
That Evolves into a Higher-Performance Laser)

[Started on April 8, 2011.]

This photo shows a version of the laser in operation. The output is not visible to the eye; the fluorescent objects on the left and right indicate the presence of the beams. (Because there are no mirrors, the laser produces two.)


Amateurs have been building lasers since fairly shortly after the laser was invented. Several laser projects even appeared in the late (and much lamented) Amateur Scientist column in Scientific American, which is now, fortunately, available in its entirety on CD-ROM. There are also various pages on the Web that provide information about DIY lasers of various sorts, and I provide links to some of them at the end of this page.

Unfortunately, I see quite a few videos on YouTube in which someone has bought a little laser module and hooked it up to a battery; they then proudly claim that they have built a laser. That’s pretty sad, especially when almost any of them actually could have built a laser. This page is for you if you really want to build a laser, and not just buy one.

The lasers I describe here are TEA nitrogen lasers. (TEA stands for Transversely Excited, Atmospheric [pressure].) That is, they do not involve either vacuum or compression. The basic design is sufficiently straightforward that it can be built by a high school student who is particularly interested, or possibly even a middle school student who is truly determined. A laser of this type that is constructed with some care and is properly adjusted should put out more than enough power to drive a small dye laser, as you can see in the addendum near the end of the page. There is also an upgrade path, which can become quite challenging.

Before we get any further along, we need some safety information and a disclaimer.

!!   WARNING   !!

If you build this project you do so on your own responsibility, and at your own risk.

These lasers use high voltages, and capacitors that can store lethal amounts of energy. They put out invisible ultraviolet light that can damage your eyes and skin. It is extremely important to take adequate safety precautions and use appropriate safety equipment with any laser; and it is crucially important with lasers that involve high voltages and/or produce invisible beams!

In addition, these designs use open spark gaps, which will damage your hearing if you do not use adequate ear protection. I strongly suggest that you acquire and use at least a pair of sound-protection earmuffs of the type used by shooters at rifle and pistol ranges; they look about like this:

Figure 1: Hearing protection

(These cost me $35, and they are definitely worth it.)

Earplugs can also help, but by themselves are probably not sufficient unless they decrease the volume by about 33 db and you put them in correctly; I suspect that only special ones that are made to fit your own ears are really good enough.

If you are not using enough hearing protection, you will probably get a nasty headache if you run the laser for a while. Take that as a warning, and get better protection! You can make a new spark gap, and you can make a new laser; but you cannot make new eyes, ears, or fingers.

A Preliminary Look

In the process of working up this page I built several versions of the laser, increasing the complexity each time I revised it. Here (Video 1a) is an informal video that begins with a simple laser, similar to the first version I present on this page, though with a different spark gap design. The laser was not very sophisticated, but as you can see in the video, it worked. (In retrospect, though, I will point out that in the later part of the video it is firing too often.) Here (Video 1b) is a video in which I assemble and operate a somewhat more sophisticated version. [Note, I did not compress this video, and the filesize is >40 MB.]

It is important to note two things about this. The first is that building a simple machine provides you with experience that helps you build more advanced versions. (No surprise there, I trust.)

The second is that this first simple machine can easily evolve, in stages, into a considerably more advanced laser with far better performance. You don’t have to throw it away and start again from scratch, because you can create a more advanced version by modifying it, as I do in the course of the first video.

Of course, if you want to start over again, nothing prevents you from doing so. The version that you see in (for example) Figures 22 and 23 was a complete rebuild, as is the one in Video 1b. That’s another handy thing about these lasers: after you have built several, you will probably find that it takes you only an hour or two. (Don’t expect your first few to be that quick, though; it does take practice.)

Although there are, fortunately, lots of good ways to build TEA nitrogen lasers, there are also lots of bad ways. It is particularly important to remember that if you try something and it doesn’t work, you need to document it carefully anyway, because you will almost certainly need the information later on, in order to figure out something that does work. You will probably also need the information in order to avoid repeating the same error[s]. It is a great relief (and sometimes a large surprise) to return to your notes, possibly months or years later, and find something you did that you may have forgotten about, and to have at least some of the information you need in order to understand how it worked ...or didn’t.

Parts and Materials

In order to build one of these, you will need the following things:


1. The Power Supply

For ease and convenience, I have taken the power supply out of an old electronic air cleaner. Here is the air cleaner before I disassembled it:

Figure 2: The air cleaner

This is what the right side looked like with the cover off:

Figure 3: Interior of the air cleaner

The power supply consists of a transformer and a small circuit board. Here it is on its new base, with a switch so I can control it:

Figure 4: The power supply

(Please note that this is just a temporary setup. Although the high voltages are insulated it is not safe to have line voltage exposed as it is here, and I will eventually enclose this supply in a well-ventilated insulating box.)

I made a voltage divider by putting a 100-million-ohm high-voltage resistor in series with an ordinary 10,000-ohm resistor, and I used the divider to measure the output voltage from this supply. The schematic diagram on the bottom of the case says that the supply puts out 5500 VDC at 0.3 milliamps, but that turns out to be a description of one polarity, not both: I measured roughly 5980 volts on the positive terminal and roughly 6390 volts on the negative terminal, for a total of over 12 kV. This is the open-circuit voltage; when the supply is providing current to a load, the voltage is lower.


2. Parts

I have been using two bases for this laser; both of them are glass, and I got both at thrift stores. (I used glass because the lower electrical plate of the laser is at high voltage, and I wanted to keep it isolated from the table. In addition, glass is generally flat, which is important.) The glass pieces I found are intended for use in the kitchen, and they have pebbly top surfaces; I decided to live with that, but if you find one that has an extremely flat bottom surface you may want to turn it over and use the flat surface as the top. Alternatively, a plain piece of window glass will work, but you should make sure that the edges are not sharp. If you prefer to avoid glass, various kinds of plastic sheet can be used to insulate the high voltage from the bench or table; just be certain that the upper surface is clean and flat.

I used a piece of single-sided circuit board as the baseplane of the laser (it would ordinarily be the ground plane, but because of the way the power supply is configured, it is definitely not at ground potential), largely because I have several pieces, acquired on eBay some time ago, that are of an appropriate size. Also, although it is quite thin, this material is just stiff enough that the pebbly surface of the glass underneath it is not a problem. (The board is so thin, in fact, that it can be cut with a pair of scissors. I don’t recommend doing that if you have a tool that is better suited to the job, however, because cutting circuit board isn’t a very nice thing to do to your scissors.)

Brass shim stock is a good alternative that you can get at some hardware stores, and also on eBay. It works at least as well as the circuit board. Shim that is 0.003" or 0.004" thick should be suitable; if it is any thicker it starts to become more difficult to cut, so you may want to obtain a pair of tinsnips if you don’t already have one.

The next step is to find a piece of plastic that can serve as a dielectric. I had originally intended to use overhead projection transparencies, but then I went to an office supply store and priced them: $40 for 100 sheets. I eventually found some at a thrift store for a much more bearable price, but they are only 0.004" thick, and that isn’t enough to handle the full output of the power supply, so I went to the hobby shop and bought some styrene sheets that are 0.010" thick and about 18" long. These work quite well, but it is important to remember to get the long size if you are using long electrodes: the brass sheets that I used as the upper plates of the capacitors in my initial versions of the laser are 12 inches long, which meant that I needed a dielectric sheet at least 13 inches long, and preferably longer. High voltage will jump across insulators if it can, and you need to provide a margin of more than half an inch all the way around the capacitors. Likewise, when I use brass sheets that are 4" x 10" (another size that I can get at the hobby shop) it is easier to deal with the length, but I need to use a piece of plastic that is well over 8½ inches wide, and because of the spark gap, which I have positioned at one edge of the laser, I really need at least 10 inches of width.

You will need a spark gap to conduct electricity from one of the capacitors to the ground plane. I made my gap out of a pair of 1/4-20 carriage bolts, as you can see from the photos, below. (Figure 9 shows the initial version.)

Note: It is not necessary to position the spark gap where I did. You can put it almost anywhere you want, provided it doesn’t interfere with some other aspect of the design. There are some people who claim that it has to be in or near a corner of the capacitor, particularly if you want to achieve what is referred to as travelling-wave excitation; but if you think about that claim you will notice that in order for it to be valid, the gap would have to switch in a rather small fraction of a nanosecond. That’s considerably faster than is physically possible for a design of this type. (If you actually have a fast photodiode and a fast oscilloscope, you can check this for yourself. You will find, as I did and as you can see in Figure 26, below, that there is exactly zero chance of it being accurate. OTOH, it is typically possible to wedge the electrode spacing so that you get most or all of your output from one end of the laser.)


3. The Laser Itself

(I’m going to presume that you have acquired the parts and materials you need, and that you have a power supply.)

Start by rounding the corners of the brass sheets that will serve as the top plates of the capacitors. (In this version, they also serve as the electrodes of the laser.) Sand the edges at the corners and ends to smooth them — sharp edges can cause sparking where you don’t want it, and can sometimes even make it easier for the high voltage puncture the dielectric, which immediately causes the laser to fail until you put a new dielectric into place. You can also make sure that the edges that face each other and serve as electrodes are straight and parallel to teach other, and it’s a very good idea to smooth them and then polish them. Here are two views of one of the electrodes from the first version I built:


Figures 5 & 6: Electrode edges

As you can see, the laser can be made to operate even if you don’t smooth the edges of the electrodes; but it won’t work as well, and it won’t be as easy to adjust. Here is a detail of two new electrodes that I have smoothed and started to polish:

Figure 7: Polished electrode edges

Because the brass pieces are formed by stamping them out of larger sheets, one face of each is slightly convex and the other is slightly concave, and the sheets are almost never really flat. It’s a good idea to put the convex faces down, as this prevents air from being trapped under the sheets. Also, if there are sharp edges it holds them a tiny distance up above the dielectric, which helps avoid punctures.

The fact that the sheets are not fully flat is another reason for using weights. In addition to the slight edge-to-edge curvature imposed by the manufacturing process, they can also be slightly bowed from rough handling. If the middle of the sheet is high, you will want more weight there. (That was what I found with some of mine, but “your mileage may vary”. If the ends are high, you can either put more weight there or very cautiously bend the strips so that the middles are slightly higher than the ends.)

You need a way to conduct high voltage from one sheet to the other, so that they both charge correctly. I originally used resistors, but the first ones I tried had too much resistance, and the channel sparked every time the laser fired, so I changed over to an inductor that I made by winding a few turns of high-voltage wire around a surplus ferrite core. (You can see this inductor in several of the photos; it is the blue toroid with the red wire wrapped around it.) Later I returned to resistors, but I used a lower value. Either method can work, but when I tried 200 ohms I found that it is not enough; a value that low will steal some power from the discharge. A combination of resistance and inductance works better than resistance alone, and if the inductor has enough turns resistors are not even necessary.

If you use resistors at all, btw, it’s a good idea to make sure that they are rated for high voltage; alternatively, as you can see in some of the photos here, you can use resistors that are encased in ceramic envelopes and are rated to dissipate several watts: they seem to withstand the voltages involved. As I say, though, if your inductor is big enough you won’t need resistors.

For your inductor, you will need something like 25 or 30 turns around a nonconducting cylinder that is perhaps an inch and a half in diameter. (I haven’t taken the time to find the minimum viable number of turns; it may be less than 25, and it will depend on the diameter of the cylinder and the turn-to-turn spacing of your coil.) For convenience and stability you can wind the wire around anything strong enough to hold it, perhaps a piece of PVC plastic plumbing pipe if you want to be relatively fancy about it. I am not fancy; I use the cardboard cylinder from a roll of toilet paper, which I stiffen by soaking it with cyanoacrylate adhesive [“superglue”]. I glue the ends into place with cyanoacrylate adhesive. This works well, at least on the type of wire I use.

Here is a photo of a combination that I made; the inductor is 25 turns, and the resistors are 100 ohms each. The wire has fairly thick insulation (0.045", a little over 1 mm); it was the thickest I found at my local hardware store. It works quite well, and my later versions do not use resistors.

Figure 8: Resistor-inductor combination

The upper electrode of the spark gap in my initial version is mounted on a small piece of brass shim stock that is attached to the capacitor plate of my laser with conductive glue, though it is probably sufficient to hold it down with a weight. In the first version of the laser I set the lower electrode of the spark gap down so that it was partly on the ground plane and partly on the dielectric, in a configuration similar to the one that Jarrod Kinsey uses (I have a link, above, to a photo of one of his lasers), and although I used some conductive glue I also put a weight on it to help hold it in place. Here is that original gap in operation:

Figure 9: Spark gap, initial version, in operation

You don’t necessarily need the shim stock: it is probably fine to connect both sides of the gap the way I did the lower side, and hold them both down with weights. That may be trickier to adjust, though — the spark will jump along the surface of the plastic unless you put some sort of blockage in place to prevent it from doing so. (The laser could possibly work with surface sparks in the gap, but probably not very well; and the surface sparks would eventually damage the plastic, after which the high voltage would puncture it and you would have to replace it. I often use a short piece of plastic I-beam from the hobby shop as a block in this sort of circumstance; you can see a piece in Figure 25, behind the gap. It is necessary to glue the I-beam (or whatever you use — a cable tie or a piece cut from one could also work) to the dielectric with something that is a good insulator at high voltages, because otherwise the sparks will just sneak under it. I use “corona dope” (there are several versions, any of which should work).


3A. Assembly

Here are some stages in the assembly process:


Figures 10-12: Assembly

(Note: this sequence shows wider plates, a different dielectric, and a more advanced version of the spark gap than you see in Figure 13. I hope I don’t have to keep saying that there are lots of good ways to build these devices.)

Figure 13: Overview of an early version

(At a later point I replaced the inductor shown here with the combination of inductor and resistors that you can see in Figure 8 and in the photo at the top of the page.)

Note: I connect one side of the power supply to the baseplane, and the other side to the capacitor plate that doesn’t have the spark gap on it; but the laser should work with the connection on either top plate, as long as both capacitors get charged. You can test to see whether one way works better with your laser than the other. You should also test to see which polarity works better. I usually find that I get better operation if I connect the positive output of the power supply to the top plates of the capacitors and the negative output to the baseplane, but your laser may be different.

Here is a picture of the output (the small bright spot), causing a piece of white paper to fluoresce. (The large diffuse area is light from the spark gap.)

Figure 14: Output from the simple version of the laser

A verbal description of the process of putting one of these together is likely to be somewhat confusing, so here is a video to accompany and supplement the photos above. In this video I build a slightly later and more advanced version of the laser. (I had just taken it apart, which is why the voiceover begins the way it does.)

(Video 2)

(Note that the spark gap design you see in this video is different from the ones I show in the photos above. Any of these designs will work.)

Because all of the pieces were already shaped and ready, because I had just disassembled the laser, and because I have a fair amount of experience, it took me only 3 minutes and 30 seconds to get it running reasonably well.

  --> CAUTION <--

Although this video can give you a fair sense of how to assemble a simple laser of this type, it is essentially guaranteed to give you an exaggerated notion of how easy it is to get one of these machines correctly adjusted. For one thing, lasers of nearly any sort almost never work when you first put them together. I had already built this one several times, though, so I had a fair sense of some key parameters — for example, the best distance between the electrodes, which is about two millimeters or a bit less for this design. In addition, I have to confess that I have, on more than one occasion (and as recently as a day before I wrote this paragraph), spent several hours at a time trying to adjust a TEA nitrogen laser so it would work correctly. You need to be patient with the laser, and you need to be especially patient with yourself.

You also need to remember safety precautions when you are tweaking, as there is often high voltage on parts of the laser after you turn off the power supply. Always turn off the power and short out the HV before you touch any part of the machine.

Signs and Symptoms:

Working and Nonworking TEA Nitrogen Lasers

First, here is a view of one section of the channel in normal operation. The tiny white dots on the surface of the cathode are expectable. They probably indicate that the channel is being driven strongly, but they are not necessary for lasing, and I don’t see them as much when the electrodes have been polished.

Figure 15: Normal discharge in the channel

Notice that the discharge is not very bright. This is normal. In fact, it is common to obtain good laser output from a discharge that appears quite dim to the eye. (Remember, we are looking for strong UV emission here, and that does not necessarily correlate with strong visible emission.)

Note, added later: As I continue to work with these lasers and to refine my designs, I get to adjust and observe them quite a bit. In the process I begin to suspect that I get best output from a discharge that is just on the edge of sparking, or is showing occasional bright sparks. My initial guess is that although more preionization produces a smoother and cleaner discharge, it also takes more of the energy that is stored in the capacitors. This decreases the amount that is available to drive the laser channel.

Here are some conditions that typically do not result in lasing:

  1. Electrodes too close together. [You will see lots of thick bright sparks between them.] Sometimes only one end is too close, as in this photo:

    Figure 16: Electrodes too close together

  2. Electrodes slightly too far apart. [You may get a nice-looking discharge, but without any output from the laser. ( Video 3 shows what this looks like.)]

  3. Electrodes much too far apart. [You will see a few thin bright sparks between them. Occasionally only one end is separated far enough for this problem to show up.]

    Figure 17: Electrodes much too far apart

  4. Every time the spark gap fires, there are bright sparks in the channel. If these occur even when you first turn the power supply on, before the laser has a chance to fire, that’s a key to this issue. [Do you have an inductor or resistor (or a combination) between the capacitors? If not, then the channel is the only way the high voltage can get from one capacitor to the other. If you do have a charging path in place, is it working right? If it is, then something very strange is going on; please email me.]

    If you do not get bright sparks when you first turn on the power, but you do get them every time the laser fires, the real question is whether they interfere with lasing. If there are only a few and they occur only sporadically, you can probably ignore them; but if there are lots of them you will want to readjust the laser. That should get you better output, and it will also help prevent the electrodes from getting too badly pitted. If the sparks always seem to occur at the same locations you probably want to disassemble the laser and clean up any jaggedness or irregularities on the working surfaces of the electrodes (and preionizers).

    In any case, you can often get some output even with a fair number of sparks, as you can see in these photos:


    Figures 18 & 19: Lasing with sparks in the channel

  5. The spark gap won’t fire unless the distance between its electrodes is really small. [Possibly a punctured dielectric, though that usually results in a complete inability to get the spark gap to fire; see below. It is somewhat more likely that you have excessive corona loss from sharp edges, and your power supply just doesn’t have enough “oomph” to overcome the losses and charge the capacitors high enough. Try smoothing the edges of the top capacitor plates, and make sure that all sharp corners have been rounded. This will also help your dielectric last longer. If you can rule out excessive corona, is your power supply working properly?]

  6. The laser stops running, even though the power supply is on. (Version 1) [If your spark gap is not stable, and the distance between its electrodes in it becomes too large, the power supply won’t be able to charge up the capacitors enough to cause the gap to fire. Turn off the power supply, and be extremely careful to discharge the laser before you mess with it!]

  7. The laser stops running, even though the power supply is on. (Version 2) [Punctured dielectric; the laser stops discharging, even though the spark gap spacing has not changed, and occasionally there may be a rapid ticking sound, not particularly loud. It is a good idea to turn off the power supply as soon as possible, so you don’t damage it. Again, be extremely careful to discharge the laser.]

If you encounter problems that I don’t list here, you may want to send me an email message. I can’t guarantee to be able to help, but at least there’s a chance.

Continuing Progress:

Improved Performance and Advanced Versions

There are various changes you can make in order to get better performance from your laser[s]. Some of these involve fairly extensive rebuilding, some don’t.

First: separate electrodes

As you can see in Videos 1 and 2, I added separate electrodes to the laser. This helps in several ways. First, moving the discharge away from contact with the surface should cause it to take longer to damage the dielectric. Second, the edges of the dielectric aren’t always very flat, and sometimes the bumps or ripples obstruct or interfere with the beam. Raising the discharge up, even a little, usually eliminates this problem. Third, having separate electrodes makes it easier to adjust the channel width, because the electrodes are not held to the dielectric by electrostatic attraction, the way the capacitor plates are. Fourth, having separate electrodes makes it easy to test various thicknesses and different edge profiles. (See the addendum about this, below.)

About electrodes: many people who build TEA nitrogen lasers use cylindrical electrodes, but my experience has been that round bars want to roll out of position at the most inconvenient times, so when I built this laser I chose a different path. I acquired some brass strips at the hobby shop instead. They turned out to work quite well. Jarrod Kinsey uses steel rods as electrodes, but he bends the ends, as you can see in the photo of one of his lasers that I link to above (where I discuss the spark gap), partly to prevent them from rolling.

Addendum: Electrode profile

I can’t tell you how your laser will behave, but I can tell you that when I smoothed and rounded the edges of the electrodes I was using in mine I got a significant improvement in performance, and when I polished them I got another slight improvement. The discharge is more uniform and a bit less likely to spark, and the laser is easier to adjust.

Addendum: Angled or wedged electrode spacing

If you’ve built and operated one of these lasers, you will have observed that there is almost always more output at one end of the laser than at the other. You can adjust the electrodes to optimize either end, but (at least on my lasers) one end is usually easier to optimize than the other. (Which end this is can change as you tweak things, though, so you need to be alert.) It is usually possible to adjust the channel so that it is slightly wider at one end of the laser, and you can sometimes get most or all of the output to come from that end.

I should note, however, that my best current designs have not been behaving that way. If I attempt to minimize the output from one end I soon begin to get less output from the other, and in fact I seem to obtain best output [at whichever end of the laser is better at the moment] when the output from the other end is only a little below its best level.

Second: Preionization

If you have already built and tweaked the initial version of this laser, you know that it is difficult to position the edges of the channel to produce a clean and even discharge. If the electrodes are too close together or too far apart (or if at least one of them is not straight), you get arcs and sparks. Finding the “sweet spot” where the spacing is optimal is not easy, and can take quite a bit of time and a lot of fussing.

One thing you can do to make adjustment slightly easier (and improve the performance at the same time) is to create a much smaller, separate discharge that starts before the main discharge and fills the laser channel with ions and UV light. (Hence the name, “preionization”.) Preionization is absolutely crucial for high performance.

If you have separate electrodes, the simplest way to accomplish this is by sharpening the edge of one of the capacitor plates. The best shape is not necessarily obvious, and the positioning of the preionizer is also important, so you are in for some fussing and fiddling.

After some thought and experimentation, I added separate preionizers between the capacitor plates and the electrodes. (Figures 22 and 23 show this configuration.) Although this makes more things you need to adjust, so it takes quite a while if you really want to optimize the laser’s performance, I think it may be worth the trouble. On the other hand, since I first wrote this posting I have had very good results using the capacitor plates as preionizers; they are ideally positioned to create a surface discharge on the dielectric.

I am using a shape that is actually similar to the edge profile that Alfonso Torres Rodríguez uses for the electrodes of his high-performance TEA lasers. I create the shape by rough-sanding the upper edge of one of the capacitor plates at about a 45° angle, but not all the way down — I leave a little bit of the original edge. It is difficult to photograph, but here is Alfon’s image of his electrode profile (red circle, with diagram below it), here is a clearer view of one of his earlier profiles, and here are two views of one of my preionizers, first showing the narrow bottom edge, and then showing the angle above it:


Figures 20 & 21: Edge profile of my preionizer

[Remember that because I am using this for preionization, not as a channel electrode, it is positioned with the narrow edge down, close to the dielectric.]

It is surprising how rough and informal the profile can be and still work, though a smoother and more even profile will give you better performance.

You will have to tweak the various spacings to find what works best. I must caution you (again) that this process is usually quite slow, and can sometimes be tedious. I often find that when I am attempting to tweak at one end, the performance at the other end changes even more.

In the best-performing versions of the laser, btw, I have preionizer profiles on both sides of the channel, and I find that the spacing between them seems to need to be slightly larger than if I use one plain edge and one shaped edge. (When the capacitor plates are also serving as preionizers, the optimal spacing seems to be at least twice the channel spacing, and possibly even more.)

Here are two photos taken during assembly. In the first, the preionizers are in place, in preliminary locations; they may end up somewhat closer together after I finish tweaking for best output. In the second I have added the electrodes, also in preliminary locations. (The next step is to add weights to hold everything firmly in place and ensure good electrical conduction. In addition to the half-brick you see in these two photos, weights of several sorts are visible in several places, including the photo at the top of the page and Figure 13.)


Figures 22 & 23: Preliminary positioning of preionizers and electrodes

Third: the spark gap

The original gap design (see Figures 9 and 13) worked, but I wanted something that was physically more stable and that I hoped would switch faster, so I changed to the design that I show in Video 1. Here is a photo:

Figure 24: A faster and more stable spark gap design

(The spacer you can see in this photo did not work well, and is obsolete. I tried various alternatives, but eventually I changed over to the version of the spark gap that is shown in Figure 29, which is stabilized by its components. See the text for details.)

The spark gap spacing determines the voltage at which the laser will fire. If the electrodes of the gap are too far apart, the power supply will be unable to deliver enough voltage to cause the gap to conduct. If they are too close together, the gap will fire very often, and typically the lasing (if you get any) will be weak. If nothing other than the piece of brass shim stock holds the upper electrode in place, it will bounce up and down when you run the laser. This can cause peculiar variations in the firing rate, but at least in some cases it can actually stabilize operation.

Important: remember to use an insulated tool when you adjust the spacing of the gap, unless the power supply is off and you have shorted out the HV.

Addendum: a “start” capacitor

There is another thing you can do to the spark gap that will improve the operation of the laser. It turns out that an additional capacitor, connected directly across the gap and positioned close to it, causes it to form a better conduction channel and to form the channel more quickly. This has two results: the laser works better in general, and the pulse-to-pulse uniformity is much improved. You can see this capacitor in Figure 25, below; it is the small brown cylinder, just to the left of the gap. (It should be as close to the spark gap as is practicable.)

Because the capacitors that comprise the laser are not very large themselves (perhaps 1500 to 5000 picofarads each, depending on construction details), the “starting capacitor” has to be quite small; I am currently using a 25-picofarad “doorknob” capacitor that I had in my junkbox, but there should certainly be other ways to accomplish this. The precise value is not very important, as long as the value is small compared with the capacitances of the laser. You do need to be sure, btw, that it will handle whatever voltage your power supply can deliver. Here is a photo, showing the gap with a start cap in place:

Figure 25: Revised spark gap design

Here, in case anyone is interested, is an oscilloscope trace:

Figure 26: Scope trace, showing the laser pulse and the light from the spark gap

(My apologies for the electrical noise that is superimposed on the trace; it is difficult to shield the detector and the oscilloscope from the nasty EMP that the laser generates when it fires. The noise, btw, looks very much the same from trace to trace. I guess that the way it is generated and the way the scope responds to it don’t change much.)

In this trace it takes about 24 or 25 nanoseconds for the electrical pulse to reach peak power. (In some others it seems to be as fast as about 16 or 18 nsec.) The laser pulse occurred at about 18 nsec, but the light from the laser had to travel about two feet farther than the light from the spark gap (it went out to the side, to a mirror that reflected it into the detector), so it reached the detector about 2 nsec later than it would have if the pathlengths had been the same. I have put a mark on the trace where it should be. (I find it interesting that even with a simple DIY laser we are in a regime where this is an issue. Making comparative measurements involving really short laser pulses must be a nightmare.)

It would be best for the laser to reach threshold when the electrical power is just reaching its peak, but at least it’s fairly close. I also suspect that it’s better to be on the upslope than on the downslope. Either way, the laser pulse is only about 1 nanosecond long, the electrical pulse is far longer, and it is therefore clear that most of the energy stored in the capacitors is wasted. This is not (and cannot be) an efficient laser.

Special Note:

If you are really interested, you can redesign the spark gap so that it is externally triggered. That will give you far more control over when and how often the laser pulses. It is also likely to improve the performance, partly because it allows the power supply to charge the laser up until you trigger the gap, rather than when the gap fires on its own, and the capacitors store more energy at the higher voltage.

I must, however, note that a triggered spark gap without a trigger pulse is an untriggered spark gap. I am, at least initially, using a commercial trigger unit that we bought on eBay, because I don’t have time to design and build one myself right now. (If I do end up building one for this project, though, I will post the design here. I will note that a triggered gap of this type tends to require a that doesn’t have to have much energy in it, but it does need to be fairly fast — the risetime should be less than 1 microsecond. The pulse from an automotive spark coil is much too slow, though there may be ways to finesse that.)

To construct the new upper electrode for the gap, I took a short brass 1/4-20 bolt and drilled a hole down the middle of it, or as nearly down the middle as I could. I then drilled out the hole with a slightly larger bit, so that I could easily fit a piece of glass capillary tubing down it. I am using melting-point capillaries that we acquired on eBay for a different project. The brand is not crucial, and in fact there are even several types of capillary tube that will work, not just the ones for determining melting points. The ones I’m using came in a container that looks like this:

Figure 27: Capillary tubes

One end of each tube is closed, but I already had some that I had shortened by cutting off the closed end. (You need to be quite careful when you do that, as these are thin-walled and very delicate. It’s a good idea to wear gloves. I moistened the outside of the tube, scored the wall lightly, and just pulled. I cut several of them this way, and most of them broke cleanly.) I also drilled a smaller hole, one that the capillary tubing would just fit into, through an acorn nut, as close to the peak as I could get it. Here is the preassembled head, but without the trigger electrode:

Figure 28: Triggered spark gap, head end

NOTE: unless the holes line up very nicely, it is important to attach the tubing to the bolt with a flexible glue. (Yes, I found this out the hard way, by using CA glue at first, and had to remove the remains of the tubing from the bolt after I tried to put the acorn nut on. As you can see in Figures 29 and 30, aquarium caulk is your friend here.) In case it isn’t obvious, you do not want to glue the tubing to both the bolt and the acorn nut, as that would make it difficult or impossible to adjust. You could glue it to the acorn nut instead of the bolt, but the bolt has more contact area and is not going to be exposed to hot plasma. For this reason I didn’t try plastic tubing, though that might also work. Jarrod Kinsey has suggested the thin tubes that often come with (for example) aerosol cans of lubricating oil as a possibility here.

I am using a piece of broken jeweler’s saw blade as a trigger electrode. It was handy, and it is fairly hard steel, so I hope it will last a little while. (This is really only the second time I have built a triggered gap, and the first one was much different, so I do not yet know whether the saw blade was a good idea. I’ve put over a hundred shots on the laser with the new gap in place, though, and it seems to be fine.)

Here is a photo:

Figure 29: Triggered spark gap

As seems to be usual with this style of gap (if I recall correctly, it is called a “trigatron”), I have the positive output of the trigger-pulse generator connected to the trigger electrode and the negative output connected to the acorn nut. I also have the positive terminal of the power supply connected to the upper plates of the capacitors. Initially, I was using only one “starting capacitor”, as you can see in some of the previous photos, but later I added a second one.

(Parenthetical note: I tried to get an oscilloscope trace of the light from this gap, but the electrical noise was so bad that I was unable to do so.)

I eventually changed the gap design again. I am now using a carriage bolt as the top side, just as I do for the free-running versions of the gap; I drill the hole starting at the head end of the bolt, so that it is centered where that’s important.

Fourth: the dielectric and the capacitor plates

I was able to get some acetate sheet (which you can see in Figures 10-12 and 22-23) from a vendor on eBay. Acetate has higher dielectric constant than most of the other plastics I’ve been using, which means that at any given voltage, a capacitor made with acetate will store more energy than a capacitor made with, for example, styrene, provided the sheets are of the same thickness. The capacitance is higher if the dielectric is thinner, and the acetate sheet that I acquired is only about 6.5 mils thick, which gives it another advantage over the 10-mil styrene sheet I was using earlier.

Because this sheet is so thin, however, it can’t handle as much voltage as a thicker sheet, and I have to adjust the spark gap so it fires about 2 or 3 times a second. At one point I allowed the time between pulses to get too long, which let the voltage on the capacitors rise too high; a puncture promptly developed in the sheet I was using, and the laser stopped working until I swapped it out and tightened the spark gap spacing. (Even 2 pulses per second doesn’t really seem to be enough; the voltage has punctured two more pieces of acetate sheet since I first wrote this paragraph, one of them twice. Fortunately the first hole was very close to a corner, and I was able to move the capacitor plates slightly to avoid it. [I put a droplet of corona dope over the hole to prevent sparks from going through it.]

In addition to finding a thinner dielectric with higher dielectric constant, for this version of the laser I changed to capacitor plates that are 4" x 10", so they have more area. (You can see these plates most easily in Figure 11, but they are also visible in Figures 12, 22, and 23.) This also increases the capacitance, and thus the amount of energy the capacitor stores at any given voltage. (With these plates, though, I have to use separate preionizers because the electrodes are 12" long.) The capacitance, btw, calculates at just over 5.5 nf; but when I measured it I found that the actual value was only about 4.2 nf. I am not entirely sure what is responsible for the difference, though it is entirely possible there is still a small amount of air trapped between the plates and the dielectric.

When I built the triggered gap I returned to 10-mil styrene and 2" x 12" plates, because I wanted to be able to pause between pulses. With the thin acetate sheet, that would not have been possible.

Alternative Drive Circuits

All of the lasers on this page use a circuit topology that is probably best described as a simple voltage-doubling circuit, though it does not actually double the voltage. (There doesn’t seem to be a better term for it.) There is also a different topology that you can also build, a Charge-Transfer Circuit.

In the doubler circuit, you charge up both capacitors and then short-circuit one of them to ground (or, in the machines I describe here, to the opposite power supply polarity). For reasons that are described elsewhere, this causes a large voltage to develop across the laser channel, which is between the two capacitors, very rapidly. The flow of current from one cap to the other, across the channel, drives the laser.

In the CT circuit, on the other hand, you charge up one capacitor, which serves at the main energy storage point for the laser. You then discharge it into a second, smaller capacitor through a high-voltage switch (typically a spark gap) that is between them. The smaller capacitor starts the laser channel conducting, and then both capacitors drive the laser.

Each of these circuit topologies has advantages and disadvantages. The doubler circuit is efficient, symmetrical (it has been found, both theoretically and experimentally, that it provides best performance when the two capacitors have the same value), and easy to construct. The CT circuit is not symmetrical, is usually less efficient, and is often somewhat less easy to build; but at its best it can perform quite well. (See the nitrogen lasers built by Alfonso Torres Rodríguez; there’s a link to his site, above.)

The main store of a CT circuit (often referred to as the “dumper” capacitor) can have relatively large value, and it does not have to be as fast as the capacitors in the doubler circuit, both of which must drive the channel directly.

One other advantage of the CT design is that you can eliminate the charging inductor or resistor-inductor combination. You are only charging one capacitor, so you don’t need or even want a connection to the other one. (In principle, you could put the connection between the peaker capacitor and the ground plane, to discharge the peaker between pulses; but I have not seen very much difference with a discharging inductor in position, so I usually don’t bother with one.)

Here are photos that show the assembly of a straightforward CT design that is closely related to the doubler designs I’ve presented above. The dumper capacitor is a 6" x 16" piece that I cut from a brass kickplate. The peaker capacitor is a brass strip, 2" x 12". This laser uses separate electrodes, which raises the channel up from the surface of the dielectric and allows the top plate of the peaker to be used as a preionizer. The spacing between the top plate of the peaker and the other brass strip is not optimized in these photos. Note: on the side of the channel that is connected to the baseplane, there is no dielectric. The brass strip on this side, which supports the electrode, should be enough thicker than the top plate of the peaker cap to compensate. (With 10-mil styrene as the dielectric, I used a 16-mil strip as the peaker and a 25-mil strip on the low side. When I changed over to 5-mil Dura-Lar™ [a polyester film that is similar to Mylar] I used a 25-mil strip as the peaker, and a 32-mil strip on the low side. I use the same combination when I have two of the 5-mil sheets under the dumper, because I am still using only a single sheet under the peaker.)

I used a thick glass plate as a base, partly because there is enough weight on the laser when it is assembled that if I didn’t provide a stiffener it could flex the top of the table, depending on where I position it. That would put a curve in the channel and prevent the laser from working properly.

Glass plate in position

Baseplane in position on the glass

5-mil polyester sheet (the dielectric) on the baseplane (See remark, just after Figure 40.)

Upper plate of the dumper cap in place

Upper plate of the peaker cap in place

Ground connection in place

Upper side of spark gap, shown in position

Channel electrodes in place

Weights in place

The laser, in operation

Figures 30-39: Assembling a Charge-Transfer version of the laser

Here is a view of the channel, with two sheets of 5-mil Dura-Lar as the dielectric for the dumper and a single sheet as the dielectric for the peaker. The dumper for this version of the laser was a piece of brass shim stock, 6" wide and about 28" long. (The baseplane for this version of the laser, which I built in February of 2012, is a piece of 12"-wide brass shim stock, also about 28" long.) I measured the dumper capacitance as just about 10 nf and the peaker capacitance as 2.9 nf, but I don’t know how accurate my meter is. The channel was adjusted extremely well, and the unfocused beam easily drove a cuvette of “Optic Whitener” that was at least 8" away from the end of the channel, with the laser pulsing a bit faster than 1 Hz. There are occasional bright sparks, but very few, and they do not appear every time the laser fires. The tiny white sparks (which Jarrod Kinsey refers to as “icicle sparks”) do, on the other hand, appear every time. They are on the cathode (negative) side. There is a dim violet glow in the channel when the laser fires, but it can be hard to see, and the camera doesn’t seem to pick it up very well. (Remember, we are looking to get UV out of the laser, not visible light, so a discharge that appears dim to the eye is not necessarily underpowered.)

Figure 40: Channel of the Charge-Transfer version

With 5-mil polyester film as the dielectric for both capacitors the channel needs to be wider for best operation, and it has more and longer “icicle sparks”. It may also require more preionization. Unfortunately, unless the laser is pulsing several times a second, a single 5-mil film used as the dielectric for the dumper typically lasts only a few hundred to a few thousand pulses, so I am using two sheets.

Here is another Charge-Transfer design. It uses a small Marx bank as its dumper capacitor and provides even better performance, but at the expense of added complexity.

Relevant Milestones

Here are some reference points you can use if you want to.

  1. You have read (and at least mostly understood) this Web page, and possibly others that deal with this subject.

  2. You have acquired all of the necessary parts, and you have a working power supply.

  3. You have built a device; it may or may not be a laser yet.

  4. At least some of the time, you get laser light out of one or both ends of the machine. (This is, obviously, crucial. If you get this far you are owed congratulations, especially if this is your first laser.)

  5. You get laser light every time or nearly every time the laser pulses, and there are relatively few bright sparks in the channel. (Another way of putting this is that you are now learning how to adjust the laser for good operation.)

  6. If you have some fluorescent material that is suitable for use in a dye laser (for example, “Optic Whitener” from Dharma Trading Co., Noodlers’ “Blue Ghost” invisible fountain pen ink, or even ink that you’ve extracted from a highlighter marker), the focused beam from the nitrogen laser is powerful enough to let you use it as a dye laser. At this stage you may need to fill the channel with nitrogen, but as you continue to improve the laser it will eventually reach the point at which it will drive a small dye laser even if the channel is filled with air.

  7. Your laser has separate electrodes, and uses its capacitor plates as preionizers. [This could easily happen before the preceding item; it depends largely on you.]

  8. The unfocused output will drive a dye laser, but only if the channel is filled with nitrogen.

  9. The unfocused output will drive a dye laser even if the channel is filled with air.

  10. The unfocused output will drive a dye laser even if the channel is filled with air and the cuvette of dye is more than a foot away from the nitrogen laser. [As of late 2011 my lasers have not yet reached this level of performance. It may be necessary to rebuild the laser to use higher voltages, and to use a different power supply.]

At some point, you probably start coming up with your own ideas for improvements, or for entire designs. This isn’t something that can be placed in the sequence, so I am not giving it a number.

In closing:

I encourage you to try various ways to improve the performance of your laser or lasers, and I strongly suggest that you keep a comprehensive notebook that includes photos, and you should take photos not only of things that worked but also of mistakes, and of things that didn’t do what you wanted them to or thought they would. It can save you from much headpounding, and from having to learn things repeatedly.


Dye Lasers as Indicators of Performance

One of the most common uses for the nitrogen laser is as a pump source for organic dye lasers. As I mentioned above, in the listing of milestones, this can serve as a rough indication of the output power.

The total energy stored in the original version of the laser (10 mil styrene and 2" x 12" plates) was probably close to 200 millijoules. That version could drive a small dye laser if I used a cylindrical lens to focus the beam onto the front of the dye cuvette. Some of the later versions can drive a small dye laser even if I don’t focus the beam. (I’ve mentioned this as one of the progress steps, above.) The photo on the left, taken from in front of the cuvette, shows this with nitrogen in the channel. After I got that version adjusted a bit better I was able to get enough output even with air in the channel. (Photo on the right, taken from behind the cuvette.)


Figures 41 & 42: Unfocused output driving a small dye laser

(The brighter spot is the dye solution using the walls of the cuvette as mirrors. The tall diffuse stripe is lasing without feedback, often referred to as ASE [“Amplified Spontaneous Emission”].)

Although I used a commercial dye and a fused silica cuvette in these photos, Video 4 shows a dye cell I built out of microscope slides, glued together with silicone aquarium caulk. In this video I use the focused output of the nitrogen laser to drive three different commercial fountain pen inks. (I have actually lased five fountain pen inks so far.) I’ve already mentioned “Optic Whitener”, which is another excellent laser dye for the DIYer, and you can even get fluorescent dyes from some highlighter markers, though they are not always optimal for nitrogen laser pumping. (If you use isopropyl alcohol to extract the dye from a Sharpie “Accent” yellow-green marker like the one that you can see in the photo at the top of the page, you get a solution that seems to work fairly well. It is quite concentrated, and you will need to dilute it, just as I diluted the inks I used in the video.)

If I focus the output, the nitrogen laser can also drive a piece of commercial fluorescent plastic:


Figures 43 & 44: Fluorescent plastic sheet, lasing

The edges of the sheet are not smooth or glossy, so its output is very diffuse. I decided to clean up one edge, but polishing it would have been difficult, so instead I glued a microscope slide to it, using cyanoacrylate adhesive (“CA”; “superglue”). (In retrospect, I should have used either a different type of CA, or epoxy. The thin glue is difficult to control, and it got on some places where I didn’t want it.)


Figures 45 & 46: Improved plastic sheet

This resulted in a slightly improved output pattern:

Figure 47: Improved plastic sheet, lasing

(Taken with air in the channel of the nitrogen laser.)

One slide provided enough improvement that I decided to add a second one. When I line up the beam of the nitrogen laser so that it is perpendicular to the slides and the dye can use them as mirrors, this works quite well:


Figures 48 & 49: Plastic sheet with 2 slides


My designs started as variants of designs developed by Jarrod Kinsey; I have introduced a number of changes, partly for ease of construction, partly because of the materials I can get, partly because of the way I think about the issues involved here, and partly to point up the fact that this is not a “one right way” situation. Just because I do something a certain way does not mean you have to do it the same way if you can’t find the part or material that I use, or even if you just want to do it differently. If you decide to do something differently, though, or if circumstances oblige you to, you will want to think it through before you build it, so you’ll have some idea of how your version is likely to work, ...unless you are willing to spend some time and effort on what could end up being a lengthy trial-and-error method. [I am hoping to provide some background information on a separate page, to help you understand how these lasers work and what the important parameters are. I also provide links to pages that other people have written about their nitrogen lasers, for comparison.]

Jarrod’s designs, in turn, are based partly on designs and suggestions from other DIYers including myself (there are no one-way streets here!) and Milan Karakas; partly on his own experiments; and, though not as directly, partly on articles that have been published in the scientific literature.

Just as I made a number of cogent suggestions to Jarrod as he developed his designs, he has made a number of cogent suggestions to me as I developed the designs on this page, and I am indebted to him for his help on this project. I am likewise indebted to Milan Karakas, who asked several key questions and provided crucially important advice relating to the performance of the lasers, particularly with regard to preionization and switching.

The preionization method that I use in the improved versions of the laser was inspired by a feature that Alfonso Torres Rodríguez uses in his TEA lasers. My spark gap design was partly inspired by two high-speed spark gap designs that I encountered in the scientific literature, and partly by what I could get at the hardware store. The power supply is my own idea.

I must also acknowledge my indebtedness to Ernest E. Bergmann, who developed some of the earliest room-pressure nitrogen lasers, and whose papers have been extremely helpful to me; to Professor Mark Csele, who has done superb work with a number of DIY-feasible lasers and has published fine Web pages about them; and to Sam Goldwasser, for his amazing Laser FAQ. (Links, below.) If I have left anyone out, I hope they will forgive me...



There are various DIY laser projects on the Web, including several TEA nitrogen lasers. No two of these are alike, which underscores the fact that there are lots of ways to think about the requirements and parameters of these lasers.

[Although it is somewhat peripheral, you may want to read my rant about the explanation that is in the Scientific American “Amateur Scientist” nitrogen laser project. — There are some potentially useful references at the end.]


Sam’s Laser FAQ is an incredible resource, covering a tremendous variety of laser types. My one caveat is that in addition to a lot of important information there are many opinions, some of which are more credible than others. When you see descriptions of a laser or a project, or specifications of a commercial device, they are generally quite reliable; when you see someone stating that dye lasers are not worth the trouble (because he barely managed to threshold Rhodamine 6G with 200 joules into his flashlamp), that’s not necessarily as reliable. (It is entirely possible to threshold some dyes with just a few joules into a flashlamp, but it is extremely difficult to threshold R6G or any other dye if you drive your flashlamp with capacitors that are not well adapted to fast-pulse service. This clearly includes photoflash capacitors, though they are admirably suited to slow pulse service.)


Here is a page on which I hope to provide additional technical information about TEA nitrogen lasers, the principles on which they work, circuit topologies, and so on. As of late 2011 I am just starting to work on it, so please bear with me if you find it mostly empty. (You can always email me with questions.)

To the top of the LASERs section


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

Contact Information:

Email:, where you can replace a with my first name (jon, only 3 letters, no “h”) and b with joss.

Phone: +1 240 604 4495.

Last modified: Sun May 5 22:33:34 EDT 2013