Here is a Coil Gun (also called a Gauss Cannon) made in three days, using all parts from the junkbox! OK, that isn't entirely true. I did buy the SCRs (salvage) and I did buy the firing switch. Believe it or not I couldn't find a momentary normally open switch in my stuff! Still, this GC was close to free and relatively simple to make. I wanted to have it working to show some students who were engaged in building a CG, so this is how I spent Memorial Day weekend when I wasn't playing volleyball.

Remember all pics can be clicked for a larger image.


A Coil Gun shoots ferrous metal projectiles using intense electromagnetic fields. That's why the projectile must be a ferrous metal. The Rev. 1.0 CG is a three stage model. The second and third stages are triggered by an infrared detector via some hefty silicon controlled rectifiers (SCRs) to make them fire at the correct time as the projectile travels down the barrel. It is vital that the coils fire at the right time, or the projectile will be slowed or even stopped.

Note: All Imperial measurements on this page are approximate. I use metric to build things, and frankly can't comprehend why folks still use fractional measurements. Lowest common denominators indeed! I only do fractions when I have to, I make too many mistakes.

Another Note: I don't go into theory and design on this page because there are some really good sites with more information than I could ever hope to add here. Therefore, there are links the the bottom of this page for those who want to know more about theory and design. This page focuses on the building of a CG. I will add another page of testing results as soon as I finish my velocity timer and can get some concrete results to report. (UPDATE: Got a velocity timer working. You can see it HERE.)

Safety: Coil guns are inherently dangerous devices. They shoot projectiles using enormous stored charges of electricity. As such, they aren't a beginners project. One mistake with a big cap bank and you are dead, no second chance. (Click HERE to see why I am so wary) *Never* "short out" a cap with a metal object. This is hard on the cap, and may blow bits of molten metal everywhere. Always use a high value resistor instead. Check with a VOM to be sure the caps is discharged. Placing the projectile too far into the coil when firing can cause a "backfire", meaning the projectile is propelled backwards, so look out! These devices generate intense magnetic fields that will be happy to eat your watch, computer, credit cards, anything that is even vaguely susceptible to magnetic damage.

The Capacitors: The caps I used are 22,000uf/75V computer grade caps. Comp. grade caps aren't really that well suited to CG pulse work. They have a relatively slow discharge and are damaged by pulse operation. However, they are cheap and easy to scrounge. I got these from a 50VDC/200A (yes, 200A!) power supply. I have fired my CG approaching 500+ times, and the caps still seem fine. I do however expect them to die at any moment. These caps exhibit a strong dielectric memory, and so bleeder resistors are even more essential than usual. After firing the caps actually build up a few volts in spite of the bleeder resistors! I hear that photo flash caps are the way to go, and am on the lookout for some to build the CG Rev.2.0. Since this pic was taken, I have added a LED with a 10k resistor across each cap. That way, I can see if there is any significant voltage left in the caps. This is a safety feature I strongly recommend.

The SCRs: I was very fortunate to run across these puppies. SCRs can be quite spendy, but I got these 3/10$US!! They came from the power supply of an IBM "Big Blue". Downside of scrounging these is that I have no way to get the specs on them. The numbers on them aren't GE numbers, and the guy at IBM started laughing when I asked. However, from researching similar looking SCRs of the same era, they must be rated around 600V/250A. If anyone knows anything about these, please LET ME KNOW. However, I suspect I'll discover their limits sooner or later...

The Barrel and Coils: The barrel is constructed of acrylic tubing. I got mine from a Home-brew store, they sell it as a "racking cane". It's intended purpose is to siphon beer from one container to another. The outside measurement is 9mm (3/8") outside diameter and 6mm (1/4"). I made my barrel 245mm (9 3/4") long. The coils are 15mm (5/8") long, with 6 layers of 20 GA double coated magnet wire. The coils are spaced 50mm (2") apart, and 50mm overhangs each end. The extra length at the back end is to allow loading of long projectiles, and on the front to allow the installation of the velocity meter that is in the works.

I found that fiber washers from the hardware store fit onto the tube snugly. I read that using a metal washer decreases performance, but one of these days I intend to try a metal one with a slit cut in it to reduce eddy effects. I think this would give the benefit of containing the magnetic field without the dreaded eddy effect. I drilled two holes into 3 of the washers to allow the wire to enter and exit the coil. After scribing marks on the tube I glued the washers in place, then wrapped the coils. One downside of this type of tube for a barrel is that the projectile scratches it during firing, and I figure that someday it will get too scratched for the IR emitters to be able to "see" through it. So far so good though..

The Infrared Detection System: I used the schematic found HERE at the Magnetic Gun Club. This site has a TON of excellent CG information. I never would have got this CG going in three days without it. The only thing I changed was the resistors feeding the IR emitting diodes. I found that I had to lower the value in order to get the IR to penetrate the barrel reliably. However, I am sure that this resulted from me using diodes and phototransitors scrounged from old floppy drives. Note that my power supply is completely different from the one shown on the schematic.

If you compare the schematic to my circuit, you will see I used many more resistors than called for, because I had to make up the correct values from what I had at hand. Almost any NPN transistor will work in this circuit. The capacitor values can be changed substantially as well. The connectors are salvaged from old computer equipment. The circuit is powered by a small regulated PSU from an external CD-ROM drive. The leads going to the detectors/diodes is phone wire, twisted using a drill to prevent induced currents triggering the SCRs.

The Power Supply: I happened to have a "industrial control transformer" rated at 150VA with 120VAC input and 220/440 output. I ran the xfrmr backwards to drop the voltage, then rectified it with a 400V/10A bridge (not visible in the pic, it's under the ballast resistors). Then I played around with charging and bleeder resistor values to prevent over-volting the caps and allow a reasonable charging rate and safety discharge rate. I ended up with a 25ohm/100 watt ballast/charging resistor and 1.2k/5watt bleeders. The bleeders are underrated and get hot. I have ordered some 1.5k/10watt to use here. Oops, I'm spending money on this dang thing already! (Note: The new 1.5k/10w work great)

The small PSU to the left of the transformer is the detector circuit power source. A fuse for the AC supply is visible on the upper left. The charging switch is placed so as to protect the xfrmr from any kickbacks during operation. Note that the mains ground is tied to the DC grounds. I'm not sure if this is a good idea or not, but it seemed wise at the time. With separate charging and detector circuit power supplies like this the grounds of both must be tied together in order for the SCRs to fire.

Mounting and Assembly: The SCRs are mounted on a 3 sided box of 12mm (1/2") thick Plexiglas glued together. The mounting feet are positioned to allow tightening of the SCRs and coils contacts. All connections are kept as short as possible and soldered. The coil is mounted using two "L" brackets drilled out to accommodate the barrel. The IR detectors are mounted on tracks from an old terminal strips, and the diodes/phototransistors are glued into sections of the contact mounts. This allows easy and secure adjustment of the photogates. The entire montage is affixed to a redwood board.


Projectiles: The barrel fits 6mm (1/4") all-thread rod stock nicely. The all-thread does scratch the barrel as mentioned above. The CG seems to prefer skinny projectiles however. The only trouble is that skinny projectiles don't always trigger the photogates, so performance is erratic. I need to make a new set of photogates that are aimed lower on the barrel. Theory says that a projectile the same length as the coils is best, but my CG shows a definite preference for longer projectiles. This means that my pulse length must be too long. More on all of this when I get some real testing done. Starting position is critical, I find a 2mm range for best performance. One thing I will include in the next CG is a way to position the projectile consistently in the breech.

Operation: OK, you have your CG assembled and ready for testing. First verify photgate operation by sticking something through the barrel and testing the gates of the SCRs. Don't load the caps up for this. Once the photogate operation is verified, you can begin firing tests. Start by disconnecting the second and third stages, put a VOM on the first cap. Tape a piece of lined paper on a box and use this as a target. If placed in the same spot every time, the lined paper will give you a crude indicator of muzzle velocity. Fire off a few shots to get a feel for the average speed and to make sure your first stage fires consistently. Play with projectile starting position. Hook up stage two and adjust the photogate for maximum speed. Repeat with stage three. Check all caps after each shot to make sure all stages are firing until reliability is proven. If you are wise you have the LED/resistor setup mentioned above, but you should still check with a VOM. Verify how long it takes for the bleeders to drain the caps.

Misc. Notes: It is important to note that this is not an optimally designed CG. It is simply what I came up with out of what I had laying around. The primary flaw of this model is that the pulse length is obviously too long, and a good deal of "suck back" is occurring. I believe this to be because I have too many turns on my coils, and I know now that each stage should have less turns than the previous. As mentioned previously I am building a velocity timer and will make a page with full details as soon as I can.

More Safety Notes: This CG is NOT very safe. The power supply is exposed, there isn't any containment for the caps, the wiring for the coil supply is exposed, the electronics are out in the open and unshielded, the charging supply can be left on during firing, there should be a shield/trap to catch any backfires, the list goes on. This was made in a hurry to give some students a CG to look at and experiment with (under supervision of course!) and to give me a testing platform to learn about CGs before designing the Rev. 2.0. Consider the bleeder resistors NON-OPTIONAL!! Always assume every cap in the entire world is charged. Treat the CG like you would an ordinary firearm. Don't omit any of the diodes shown in the schematic. They are vital protection for your caps and you.

Links to Coil Gun sites:

This page created 06/08/02
Last modified: 07/06/02