Rebuilding (Yet Again) the C5000 Nitrogen Laser Head

This Phase: Doorknobs, Version 2

Preliminary notes:

(06 December, 2008)

This version is a reprise of the one on the second page of this set, and as I said on that page it has been quite a while since I worked on this laser. I have, however, thought about it a fair amount, and I have decided that it is time to try again. (Every time I rebuild this thing I learn something from it, even when it doesn’t work as well as I might like.)

Here’s the topology I am using for this rebuild:

I haven’t shown the chokes in the + and - HV lines, which keep the EMP from destroying the power supply and also help mitigate the effect of shorting the output of the power supply every time I fire the laser. I should also point out that I have not bothered with the small starting capacitor that I usually put across the spark gap, because the main store is small enough and fast enough that it didn’t seem necessary. The bleeder resistor, which is part of the original design, prevents you from getting a nasty shock if you open up the head and touch the peaker caps, which are not completely discharged when the laser fires.

Please note that while this looks fairly straightforward electrically, it is not trivial to implement physically. The spark gap has to stand on its head on top of the cathode rail, which makes it somewhat difficult to reach the trigger electrode, and requires insulation to prevent the trigger signal from shorting to the cathode outside the housing of the spark gap. Moreover, the positive side of the power supply is grounded, and the negative side is the “hot” side.

If you go back to the first page of this set, you will observe the fact that this head is built to a slightly strange design. There are two cathodes, so it generates two output beams. (You can see this in some of the photos, below.) I usually just focus them together, but there may be ways to do interesting things by separating them, for example a MOPA dye laser where one beam powers the oscillator and the other powers the amplifier.



Step One: Peakers

(06 December, 2008)

I am going to use the existing peaker caps, which are already in place from the rebuild that I wrote up on the previous page. There will be either 16 or 20 of them, depending on how much sparking I get near the ends of the laser with all of them in place. They are rated for 30 kV, and I hope to charge them to something on the order of 10 kV. (The “dumper” cap will be charged to 20 kV; but once the discharge starts it’s hard to get the voltage up much higher, and there is almost no chance that I’ll ever reach the theoretical maximum. If I get half of what’s on the main store, I’ll be satisfied.)

[Note, added 23 December, 2008: This has changed. See below for details.]

(from 13 September, 2004)

Here’s what the peakers looked like, installed, before I put the distribution rail back into the box:

(You can see the rail and its connector (the round brass object in the middle) lying against the interior insulation sheets, at the back of the photo. The peculiar object in front of the box, on the floor, is the vacuum regulator that I used last time; I do not expect to use it this time.)



Step Two: Main Store and Switch

(06 December, 2008)

As I have implied in the title, I will be using doorknob caps as the main store. These will be mounted on the lid of the enclosure, to either side of the opening for the switch. I have mounting holes for 8 of these on each side; as they are 2 nf apiece, the total capacitance will be 32 nf. The caps are SrTiO3, and are rated for 40 kV. My best guess is that they came out of decommissioned excimer lasers. (There’s a photo further down the page.)

For the switch, I am going to use an EG&G GP-70 spark gap. I’ve tested one of these, and it works fine with our old TM-11 trigger unit. It is rated for use at up to 20 kV in air, it will handle considerably more current than this capacitor array is likely to push through it, and it’s compact. Can’t ask much more than that. Here’s what it looks like:

The top and bottom plates are the same size, btw; they appear different in this photo because the camera was very close to the device. (“Big nose” effect of operating with the lens at a wide angle.)

=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=

Here’s my rationale on the current: at 20 kV, 32 nf stores 6.4 joules. If we guess that the effective system inductance is around 100 nh (which is probably better than reality), and if we ignore resistive contributions, the discharge should take something on the order of 170 nsec FWHM (“Full Width [at] Half Maximum”). In actual operation it will probably take longer, but we’re looking for an upper bound on the current, so I will go with this figure. This represents electrical power of a little less than 38 MW, which we can call 40 for convenience. If we guess that peak power occurs when the voltage on the dumper has fallen to about 2/3 of its initial value, we get something on the order of 3000 Amperes. Even if that’s low by a factor of 5, which seems extremely unlikely, we still win: the GP-70 is rated to handle up to 25,000 Amps.

=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=

Here’s what the spark gap looks like with its insulating hat on:

When the switch is actually installed, a trigger wire will emerge through a slot in the side of the hat.

Here is the cathode rail connector, to the back of which I have to mount the GP-70:

The side facing up in this picture is the side that screws into the rail. In the next photo I have just dropped the connector onto the hat of the switch, so you can see the general arrangement, though it is upside-down here.

(You can see that the size of the connector is not a good match to the size of the spark gap. I still need to make an adapter, and I’m thinking about appropriate design and construction.)

(18 December, 2008)

I may have some ideas about connecting the switch to the cathode; need to make a measurement or two.



Step Three: Assembly and Initial Testing

(18 December, 2008, early AM)

I have now wire-brushed the areas on the underside of the “lid” where the bolts hold it down to the box. I will have to wire-brush the corresponding areas on the box, but that only takes a few minutes. I intend to use silver-loaded conductive material betwen the lid and the box, and I’ve already applied it to the attachment points of the 16 large doorknob caps that are installed on the lid. Here is a view of half of them:

Once the spark gap is attached to the cathode rail, the only large step that remains is to connect the doorknobs to the spark gap, and that should be relatively easy.

(19 December, 2008)

Yesterday I acquired two solid steel switchplates and a hose clamp at the hardware store. The switchplates appear to be old stock; they are about 1/16" thick and relatively smooth, and the pricetags looked very old. (The newer ones are about half as thick, and are zinc plated or galvanized.) The existing mounting holes are almost the right distance apart, and I can just widen them to fit two mounting holes on the GP-70; I will have to drill the other two.

I am cutting a round hole in one of them, to accommodate the hat on the trigger electrode. If I have enough time and energy, I may also cut the edge off that one, to help minimize corona from it to the box, but it is more likely that I will just put an insulator around it. (A short length of large-diameter PVC pipe is a good bet for this.) The other one should be fine as is, because it will be up in the air at the top of the stack, and will be at ground potential except during a firing cycle.

(Later, that evening)

I have now cut a circular hole in the middle of one of the steel plates, extended the existing bolt holes, and drilled two others. The hole in the middle may be unnecessary — in order to use this plate on the trigger end of the gap, I would either have to widen the hole and put the brass shim through it, or cut the edge off the plate to turn it into a ring of appropriate size, and put the brass shim around it. Either way would be somewhat painful, and I am considering whether I need to bother with either of them. (Chances are pretty good that I need at least some sort of ring or pressure plate[s] to hold the shim in place; there are only 4 bolt holes in each end of the spark gap.)

(20 December, late evening)

I have successfully constructed the brass shim piece that connects the spark gap to the cathode connector, and I have used ultrafine sandpaper to remove the oxide from the edge of the connector itself so I will be able to achieve good contact to it. Need to cut a hole in the brass shim for the trigger wire. This isn’t fancy, but I think it will do.

(23 December, 2008)

I took the head off the shelf to get it ready for actual assembly (I want to wirebrush the contact areas for the lid, and also dust it out a bit), and discovered that the peaker caps were not present. This may be just as well; I now have some 1400-pf 20-kV doorknobs that appear to be intended for laser use, and have much wider terminations than the 780-pf 30-kV ones that were in there. True, the decreased voltage rating carries a bit of risk, but probably not very much. Unfortunately, the decreased voltage rating also means that these caps do not stand as tall as the ones I was previously using, and I have to add spacers to them so that the cathode rail will be at the proper height. (It has to press correctly against the phosphor-bronze spring fingers that carry the current to the cathode.) I have acquired some brass washers that should serve, and I will be using dabs of silver conductive paint on all of the “in-betweens”.

The bolt holes in the head are intended for #6 screws, so they don’t quite fit #8s; I put a #8 tap through the 10 holes I’m using, which I could almost do with my fingers; that enlarged them just enough. The capacitors are short enough that each of them will require at least 6 brass washers; looks like I only have half the number I’m going to need, so I will be obliged to get more.

I used the existing brass washers under the peakers, which are now attached to the inside of the box.

(24 December, 2008, early AM)

I was unable to find brass washers of an appropriate size at the local Home Depot, and acquired some steel fender washers instead, somewhat against my better judgement. I am thinking about going out on the Web and finding The Right Stuff, but will probably do a preliminary assembly with what’s on hand. That should tell me whether I’m on a viable path.

(25 December, 2008, very late evening)

It turned out that the holes in the centers of the steel washers were not quite large enough for #8 screws, and I had to drill all of them. It also turned out that 4 per cap was not enough, so I drilled 53 of the silly things, 2 of which were slightly defective. Then I mounted the cathode connector rail:

With all of the inner guts in place it was time to put the lid on, so I slid the two insulator sheets into the box and bolted the lid into position. (Some of the screws were not yet present when I took this photo.)

Time to fabricate the brass shim piece that connects the dumper caps to the switch. Here’s an initial look, with both trigger leads attached, and with an insulator in place to help minimize leakage from all those sharp points (which I will almost certainly remove, later). This was a preliminary assembly; the bolts that hold the shim onto the switch were not fully tightened yet.

I had hoped to see first light tonight, but I’m not quite there yet. Getting very close, though. A few of the boltholes in the brass shim piece are misplaced slightly, as you can see in the photo (the last two at the top; also one on the other side, not visible here), but otherwise this setup is essentially complete. Once I tweak the holes (trivial; should take about 3 minutes) and get the last 3 screws on, it will be time to check the vacuum. After that, things should move very quickly toward operation.

(Late afternoon, 26 December, 2008)

The laser is now on the bench, and I have acquired the polyethylene tubing I need in order to attach it to the vacuum pump and the nitrogen supply. I checked the vacuum manifold and pump, and although it would never pass muster on a real vacuum system, it is more than good enough for what I’m doing here: goes down smoothly and swiftly to well under 1 Torr, and takes several minutes to come back up over 10 Torr when I turn the pump off. (A real vacuum system, even with just a roughing pump, would go down to a few microns, and would take days to come back up.)

First Light

(about 6:30 PM, 26 December, 2008)

I have now operated the laser. It is quite well behaved. There are needle valves on the inlet and outlet, which give me some control over the pressure, and I have a gauge on it. I ran it with indicated pressure as low as about 12.5 Torr, and as high as about 42 to 45 Torr, above which there is sparking at the top of the bleeder resistor that keeps the peaker caps discharged between shots. (It is very close to one of the support legs, and I may paint insulating varnish on it later, if I decide I don’t need it as a safety to prevent overvolting the peakers.)

The early appearance is that something around 30 Torr will be optimal. The fact that there are small air leaks in the gas/vacuum manifold is not a problem: up to about 0.5% oxygen is not harmful, and in fact about 0.3% is actually slightly helpful. I do not have any really good way of calibrating the gauge, and it is connected to tubing at some distance from the head itself, but both the head and the gauge are between the needle valves, so it is probably reasonably close. It is also moderately repeatable and settable, which is important both for characterization and for optimization.

The photo on the left shows the output at 40 Torr or a bit higher; the odd structure of the beam is evident in the photo. Unless I am misremembering, the photo on the right shows the beam with the pressure between 25 and 30 Torr. The camera is better at distinguishing between these than the eye: in the photos you can clearly see that the laser is much brighter at the lower pressure, but when I was taking the pictures I couldn’t see much difference. Both of these, btw, are actually pictures of the fluorescence of the paper target, not the beam itself.

           

Before I proceed to the next section I will probably try running a small dye laser, to see how well I can focus the beam.

[Note, added the next day: It occurs to me that I failed to provide a sense of scale, so here is a photo with a ruler. Notice that the central bright part of the beam is at least 62 mm long. The spacing between the cathode and anode is almost certainly about 60 mm, which is remarkably large for this sort of laser.

As you can see, the laser is not quite level on the bench yet. I shimmed it, but apparently not quite enough, and I will be adding more shims.]

(A little later, Friday evening...)

For the next three photos, I dissolved 2 drops of Dharma Trading Company “Optic Whitener” in a cc or so of 95% ethanol, in a small cuvette that I built a few days ago. Optic Whitener is probably the best DIY laser dye I have found so far, and it lased at all pressures I tried. I didn’t seem to see much dependence of the output on pressure, though there was clearly less at the highest pressure I tried (just under 38 Torr, photo on the right), and slightly less at the lowest pressure (about 10.6 Torr, photo on the left)...

                       

Some of the color you see on the target is probably just the way the camera “sees” output of the dye laser, which is actually a beautiful violet; some of it may be fluorescence from the target; and some of it may be the brightness of the dye output overloading the camera’s sensor a little. The difference in color between the cuvette and the wall, however, is real.

This particular cuvette, btw, is somewhat asymmetric, and at some point I will try to provide a picture that shows the differences between the two outputs. When you effectively have only one mirror on a high-gain medium, you can get some odd little effects, and when you have two low-reflectance mirrors on a short-pulse laser you can get even odder effects. The nitrogen laser itself, of course, at least in most low-pressure designs, is one example, as it is typically operated that way; but the cuvette is different because the “mirrors” reflect only about 6% (assuming that they are fused silica, as the ones on this cuvette happen to be), and you get output at both ends of the medium. I am considering constructing a more complex cuvette, with adjustable windows so I can tweak the angle of the reflection. If I do that I will probably use sapphire windows on it, to get about 14% reflection, and if it works it will become a “TJIIRRS” entry. Designing such a device, however, is not exactly trivial, and I haven’t had time to give it careful thought yet. Meanwhile, here is a photo with one mirror in place and a bit less room light, taken with the pressure gauge reading 16.9 Torr:

Here is some mediocre (I think it’s only about 85% pure, definitely not laser grade) Rhodamine 6G dissolved in 99+% isopropanol, with the nitrogen at 24.3 Torr indicated pressure. In the first photo, the dye is using just the reflections from the walls of the cuvette (no external mirrors). For the second photo I have added (and more or less aligned) one mirror, which is not visible here:

           

With both dyes, the nitrogen laser’s output was approximately focused on the front of the cuvette, but I did not make much effort to optimize it, and I noticed that the dye was happy to lase with the cuvette even roughly positioned. With the Rhodamine in place I tried turning off the supply of nitrogen, and the dye continued to lase as the indicated pressure went down to 5.3 Torr, probably with a certain amount of air in it from leaks. At that point I shut down the system.

This laser is clearly performing quite well, and I am extremely pleased with it. I suspect that a generous application of silver-conductive material to the lid attachment points and the stacks of washers on the peaker capacitors would improve it even more, and when I have both time and enough silver goop, I will probably try that. In the meanwhile, however, I certainly can’t complain.

(27 December, 2008)

I think that if I can get the old Tek 7104 to behave, it is about time to check the pulsewidth. See below...



Step Four: Measurement and Characterization

(Noon, Saturday, 27 December, 2008)

My first attempt at taking a pulsewidth measurement has failed. Although the EMP from the laser does not seem to bother the Canon G3 camera (which was sitting on a tiny tripod right on top of the TM-11 trigger unit when I took all of the photos above!), it certainly bothers the oscilloscope. I am getting only hash on the screen, regardless of whether I have the detector turned on or off, and even regardless of whether it is connected. When I get a chance I will try powering the scope from a different source, and moving it farther away from the laser.

(That evening)

My second attempt was slightly improved, but something still isn’t right. Here are two traces that I took with the vacuum gauge at 14.4 Torr, but I’m not sure how much of that was nitrogen, so take it with a grain or two of salt.

           

The scope faithfully assures me that these were showing 2 nsec/division, which would mean that the nitrogen laser pulse is about 1.5 nsec long, FWHM. That’s highly unlikely; at 2 nsec/div, the pulse should fill about half of the screen. Notice all the noise, as well. I do not trust these traces. OTOH, it is clear that the peaks were generated by the laser: if I change the gas pressure, the peak height changes.

(afternoon, 28 December, 2008)

I am attempting to shield the bench a bit, by connecting lengths of hardware cloth to it and running them up toward the ceiling. I don’t think I want to try building anything resembling a Faraday cage, as it would be impossible to get power and control in or out, but I hope that I can reduce the electrical noise at least a little.

(later that evening)

I had to retreat to the 600-MHz vertical amplifier plugin, and there is still rather a lot of noise, but at least I was able to capture some traces. Here are two of the least nasty ones:

           

Despite the noise and wobbles, I tend to trust these more than I trust the ones I took yesterday. The FWHM pulsewidth is 8 or 9 nsec, not just on the two images here, but on several others as well. Next, I get to measure the pulse energy.

(30 December, 2008, early AM)

I set a sensor up with out little homebrew instrumentation amplifier, and looked at the output of the amp with a DMM. The sensor head was able to “see” my hand, and it turned out that reading the output of the amp with the DMM, which I hadn’t really tried before because I thought the reading would be too squirrelly, was actually okay. When I set things up on the bench and ran the laser into the sensor, however, I got figures that seemed too small. The best number was perhaps 12.7 millivolts, and that was after X100 amplification. (I’m using half of an INA 2141.)

...So I swapped out the batteries in the amplifier, and immediately started getting readings of -1.6 V. Worse, the sensor didn’t seem to pick up the IR from my hand. I think I may have to build a new amplifier board.



Interlude: The Oscilloscope Camera

(I may move this, later, to a different page.)

(30 December, 2008, early AM)

It may not be easy to tell, because I worked rotation and shear and perspective magic on the ’scope photos with the Gimp, but I have been having a rather difficult time trying to capture traces on the screen of the scope. It occurred to me, a day or two ago, that I could probably take an old scope camera and modify it. I thought I remembered getting one with the scope, and that turned out to be correct, so I took a look at it.

It consisted of a flange that attaches to the front of the scope, a large box of electronics and optics, and a Polaroid film back. Even if we could get film for the back, we would just have to photograph or scan it, so I had no qualms about putting the camera part on the shelf. It turned out to be trivial to remove the box from the flange, so I did. Then I constructed a very simple box from scrap plywood, to which I have attached an adapter that fits the lens shield tube on our Canon G3. I have a new tube on order, and I will attach it more or less permanently to the new adapter; the G3 mounts very quickly and easily to the tube, so that is probably the preferred way to handle this. It will now be much easier to take pictures of scope traces...

(31 December, 2008, early AM)

Here’s what the completed camera looks like:

Although I get some pincushioning with the G3, the new setup works. End of that annoyance.

Also end of interlude; back to measurements.



(31 December, morning)

Not only did I build a new instrumentation amplifier board, I built two of them, using what I learned from the first to do a better job on the second. Let’s call them 1 and 2, ignoring the original one that I replaced. There is something very strange going on with the amp; when I power it up (with either new board), it shows a surprisingly large output voltage. #1 tends to be around 15-20 mV negative, and #2 tends to be around 40-50 mV negative. This voltage takes several minutes to decrease to something on the order of 2.3 to 6.6 mV negative, but can increase or decrease for no obvious reason. I have seen it as small as 0.8 mV and as large as 7.8 or 8.1, for fairly short periods.

I tried shorting the inputs of board #1 together; the output went to 0.1 mV or so, and stayed there. This rather strongly suggests that the weirdness is coming from the sensor head, not from inside the amp box. I’m not sure whether there is much of anything I can do about it, but I will be checking with our other Scientech head to be sure that it isn’t just the one device.

In addition, and this is no real surprise considering the fact that I am putting pulses into the sensor, the output reading varies quite rapidly. I think I will put about 10K ohms in series with the output, inside the amplifier box, and then put a largish capacitor across the input of the multimeter. If I get a time-constant that is perhaps 5 or 10 seconds, I should be able to take readings much more easily. OTOH, it will take much longer for the reading to go back to zero when I stop pulsing the laser, so I may include some provision for momentarily shorting out the capacitor.

I don’t have a diagram yet, so here is a list of the connections to the INA2141:

PIN  FUNCTION  CONNECTED TO
___  ________  __________________________________________

 1    - in      1 M to Gnd, 100 pf to 2 as input bypass
 2    + in      1 M to Gnd (100 pf to 1)
 3   select     shorted to 4 for X100 amplification
 4   select      (shorted to 3)
 5    - out     Gnd
 6    + out     shorted to 7
 7    Ref        (shorted to 6)

 8     V-       negative supply; .047 μf bypass to Gnd
 9     V+       positive supply; .047 μf bypass to Gnd

10               NC
11               NC
12               NC
13               NC
14    - in       GND
15               NC
16    + in       GND
_________________________________________________________

Notes:

=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=

The 2nd new board may be of mild interest because the version of the INA2141 that I’m using is a surface-mount device, which means that the pins are at half the spacing of the pads and holes they must connect to (I built each amp on part of a small Radio Shack prototyping board), and I had to finesse things a bit. You will have to forgive my technique; I do not have a small enough tip for my soldering iron, and some of the work is a bit gloppy. Here is a photo of it, with the 10K resistor (left side, about halfway up) in place:

(later that afternoon)

Because it is so difficult to see what’s going on there, I have made a pseudostereo macro shot, intended for cross-eyed viewing. (If you try the large version and decide that you really need more pixels, change “17c” to “35c”.)

Note the two resistors that sit on the back of the chip package. If you look very carefully just to the left of the upper bulge of the one on the left, you can see pins 3 and 4 shorted to each other; they hang out in the air in a little wishbone shape. I lucked out a bit here: pin 1 is bent a little, and goes to the second pad down from the top; pin 2 goes to the third pad; pins 3 and 4 are up in the air, so pin 5 goes to the fourth pad; pins 6 and 7 are shorted together, and go to the fifth pad; and pin 8 goes to the sixth pad.

(02 January, 2009 [it feels weird to write that number!])

Last night I checked again with the new board, and found that when I shorted out the input, I saw -1.6 mV or so on the output. It was quite stable. When I connected the sensor head, the output again went to >30 mV, and changed fairly rapidly. This suggests to me that the grounding resistances I’m using, which are 1 Megohm, could be too large. I am going to try 1/10 of that value, to see whether it makes a significant difference.

=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=-_._-=

I made several sets of measurements and calibrations with the first new board in place, after which I built the second board, put that in place, and made another set. I then did some averaging to help remove the variations caused by the input weirdness and resultant wobbliness of the readings.

Procedural note: the way I usually do this is to record the initial [low] voltage reading, pulse the laser once a second for a minute or so, and record the highest or second-highest reading I get near the end of that period. (These are typically only 0.1 V apart, and the top few readings are generally similar, though there is some variation, which may relate to the variation in the low end readings.) Then I let the reading go down again, and record the [new] low point, as well as the gas pressure.

For calibration I generally take the sensor, the amp, and the meter back to the workbench without turning off the amp. I then use the bench power supply, plugged into the calibration resistance, to take readings, using the same protocol: note the initial low, apply a small steady voltage to the calibration resistance, note the high, measure the calibration voltage if I haven’t already done so, turn it off, and note the new low. For most of the later readings, rather than attempt to match the output numbers I got from the laser, I have just left the power supply at 232-233 mV and noted the resulting numbers, but because I got to that input voltage by matching the laser numbers in the first place, they were still fairly close. The calibration resistance appears to be between 39.1 and 39.2 ohms with the lab at its current temperature, so I used 39.15 for the last calculation. (I used 39.0 for the first few runs, because that was the value I got yesterday. Not that it makes a huge difference — 0.15 ohms out of 39 is less than half a percent.)

P (watts) = E2/R; with E at 0.232 and R at 39.15, the bench supply was putting ~1.37 mW into the calibration resistance. This gave me, on average, about 12 mV change in the output of the amplifier, which is roughly 114 microwatts in per millivolt out. 12 mV was pretty close to what I was getting with the laser, so it was convenient to run the bench supply to that level. Earlier runs gave me higher numbers, on the order of 125 to 137 μW/mV, and I will have to do at least one more calibration run after I put some smoothing on the input of the multimeter, to see whether that changes anything.

There is, btw, no real need to match the laser numbers and the supply numbers; all you need is the number of watts that produces 1 mV output. I matched the numbers because it was easy to do so.

Given the above results, it appears that the sensor is receiving about 1.4 millijoules per pulse. I seem to get best operation around 15-17 Torr, though it is really too early to make any firm claims about pressure dependence. Considering the behavior of the dye cuvettes and the fact that the pulse seems to be relatively long, 1.4 mJ seems rather low. (At 9 nsec FWHM and with a slightly long tail, the peak power is really not much more than 150 kW.) Still, I am only charging the main store up to 20 kV, so I probably shouldn’t complain too loudly. Also, I am losing a little bit of energy in reflections from the lens surfaces, and I may or may not be hitting the front of the sensor cleanly. (I am thinking about adapting the back end of the sensor to give it a mounting plate, so that I can remove the tube from in front of it. That will let me be quite certain that the beam is going where it needs to. It also lets the sensor detect IR from somewhat off-angle, which is not so desirable; I may add a shorter tube to restrict that a little, while still letting me check the beam delivery.)

(early afternoon, still New Year’s Eve)

With 10 K in series with the output and 680 μf across the meter, I performed yet another set of measurements. It was definitely easier to read the meter, and the time-constant was short enough that I didn’t have to worry about shorting the cap to restore the value after running it up with the laser.

This time, I calibrated with the bench supply putting out as close to a quarter of a volt as I could set it: 250.4 to 250.5 mV, according to the multimeter. (I unplugged the 680 μf cap for the purpose of measuring the supply.) Averaging things out, I appear to get very close to 120 μW/mV, not far off from the two previous calibration sets. (I’m still not sure about the run where I got 137, though it may have something to do with the variability of the readings at the low end.)

It is hard to calculate the laser’s output, because the low-end numbers were so variable, but it looks like I am getting just over 1.5 mJ/pulse at an indicated pressure of 16.9 Torr. If I can figure out a better way to measure this, I will report it. In the meanwhile, I think it’s reasonable to conclude that the laser is putting out something on the order of 150 kW peak, in a pulse that is about 8.5 nsec long, FWHM.

Next: I want to put a prism into the beam. There are several reports in the literature that indicate output at 357.6 nm as well as the usual 337.1, and I should be able to see two fluorescent spots on a paper target if this laser is putting out both of those wavelengths.

(01 January, 2009, morning)

I did not see two spots, but I wonder whether the dispersion of the prism can separate those wavelengths enough for them to be visibly distinct in a few inches, which was the distance to the target. I may have to try this again with the target farther away, and perhaps with a slit to narrow the part of the beam that hits the prism. I did, though, use the opportunity to realign the laser’s mirror, and now I need to re-check the output energy. Speaking of which, I have been triggering the laser by hand while doing that, which makes for some irregularity even though I use a clock as a metronome, so I just ordered a 60-rpm synchronous motor on eBay. One cam, one microswitch, and a way to mount everything (scrap plywood is probably my friend), and I will have a nice even timing signal for the trigger unit.

[Note, added 07 January, 2009, evening:

The motor arrived and I went to put a knob on its shaft, to serve as a cam. The shaft, however, was too large to fit into the knob I had on hand. I went to Radio Shack, and found that all of their knobs were built for ¼-inch shafts, so I bought a pack of two knobs that looked like they might work well, drilled out the shaft hole in one of them so it fit on the shaft of the motor, and epoxied a washer to the edge. I had to shim up the motor a bit so the washer would catch the bar of the microswitch, as you can see in the photos. Here is the working unit. Sorry about the peculiar angle of the first photo.

           

(I will eventually put feet on this device so it stands up on its own.)

Having the laser pulsing once a second turns out to be really handy, not just for taking energy measurements, but also for making adjustments on (for example) dye lasers I’m driving with the C5000.

End of note...]

(Evening, 04 January, 2009)

In an effort to stabilize the output of the instrumentation amplifier, I added a 100 Kohm resistor in parallel with each of the 1 Megohm resistors you can see in the photos above. I would have thought that 1 M would be enough, as the input impedance of the amp chip is something like 1012 ohms, but perhaps not...

With the new resistors in place, I set about trying to make another set of measurements. When I turned on the amplifier, its output went up and down about 8 times over a period of perhaps 5 or 10 minutes, and finally appeared to be settling near -2 mV, so I carried everything over to the bench, whereupon it went nuts again for a while. It eventually seemed to settle down, this time closer to -2.4 mV (though still considerably more variable than I would have liked), and I took data points at several pressures. By the end of this, perhaps as much as an hour later, it was mostly coming back to -2.3 to -2.5 mV... mostly.

The readings were essentially consistent with what I was getting the other night, differences as large as ~11.5 mV at pressures around 17 Torr. I continue to think about alternative sensors, about optimizing the head by adding silver-conductive coatings where I had originally planned them, and also about building a new version of my largest previous “doorknob” head.

(Early AM, 09 January, 2009)

Another way to get a sense of the output is to see whether the camera can detect some difference as a variable is changed. I ran a Rhodamine 6G dye laser with one mirror, and definitely saw some differences as I changed the pressure. At 3.9 Torr, I occasionally saw weak lasing. Brightest output seemed, oddly, to be around 14.7 Torr, though I didn’t test at closely spaced pressures when I was taking this set of photos, because I couldn’t see the differences as clearly by eye. Here are some of the photos; 6 Torr, 10.4 Torr, 14.7 Torr, 23.4 Torr, and 29.8 Torr...

                                               

The peak appears to be fairly broad; I don’t see all that much difference between 14.7 Torr and 23.4 Torr in these photos, though a look at the entire set does suggest that the 14.7 Torr image is slightly brighter than the others.



Back to Page 3 of this set

Back to Page 2 of this set

Back to the beginning of this set

To a page about another commercial nitrogen laser we acquired surplus

To a page about a commercial TEA nitrogen laser we acquired surplus

To a pageset about some high-power nitrogen lasers that I’ve been designing and building

To the top of the LASERs section

Home



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



Contact Information:

Email: it’s the usual “a@b.com”, where you can replace A by my first name (jon, only 3 letters, no “h”), and B by joss.

Phone: +1 240 604 4495.

Last modified: Fri Jun 14 23:08:37 EDT 2013