Pneumatic Cannon

I want to install a functioning air-powered cannon in my 1/5 Scale Sherman Tank. While this is pretty big for a model tank, it is pretty small for a "spud gun", so I had to develop my own design that met some tough criteria:

A lot of different designs came and went before I settled on this one.


As mentioned above, the Sherman had a 75mm gun. At 1/5 scale, this comes out to 0.59". I actually used a piece of ¾" copper pipe. This works out to 95 scale mm, but what the hell.

Just mounting a piece of ¾" copper in the tank makes for a very skinny looking gun, however - the wall thickness of the Sherman's 75mm gun was much thicker. Since I already had a piece of 1" PVC painted green from the mockup gun, I just installed the copper pipe inside the PVC, with cardboard spacers to center the copper, and epoxied the whole thing in place.

Air Gun Barrel Detail


This is the other reason I used ¾" copper pipe - I had a 3/8" brass ball valve that was just over ¾" O.D. Using a file, I rounded and trimmed the end of this valve so it would just slide inside the copper pipe. I cut a pair of slots in the end of the pipe, and inserted some screws into the valve body to engage the slots, making a kind of "bayonette" fitting.

Air Gun Breech Detail

Firing the gun is a simple matter of loading it, inserting the breech, and opening the valve.

Experiments on the breech revealed that even when the breech was only half inserted into the barrel and the barrel blocked with a rubber cork, the breech did not pop out, even at 50 psi. The force put on the breech by pulling the firing lever is sufficient to keep it in place.

Air System

The air system is simple, too. A 60 cubic inch air tank (from Surplus Center, rated to 225 psi) supplies air through a regulator set to about 50 psi (I plan on making some modifications so I can use a lower pressure). The regulator supplies a 5/16" I.D. hose, which carries the air to the breech. At some point I will install a needle valve between the regulator and the hose to limit the flow rate.

Air Gun Schematic

The hose serves two purposes; not only to convey the air to the breech, but also to accumulate a charge of air of several cubic inches of air to fire the gun. The larger the volume of air stored in this hose, the lower the pressure required. An 18" long 5/16" ID hose provides 1.38 cubic inches, and seems to require 50 psi to fire with authority; at 25 psi, it is positively pathetic. If the pressure and volume is too low, the air won't completely fill the barrel upon expanding to atmospheric pressure, resulting in a partial vacuum "sucking" the bullet back! Increasing volume will allow me to use a lower pressure without sacrificing performance, and I suspect that the speed with which one opens the valve will matter less, too.


My original plan was to fire small styrofoam balls, but since I had a short piece of ¾" copper pipe left over from the barrel, I just sharpened one end of it and made a styrofoam plug cutter. This provides me with as many ¾" foam "slugs" as I can stand.

Unfortunately, they have virtually no range. The fly out of the barrel at very high speed, but then instantly stop and flutter to the ground. At lower volumes and pressures, they sometimes do not even leave the barrel! Not quite the effect what I am looking for.

One interesting experiment used a foam slug pushing a small aluminum foil ball - when the pressure and volume were low, the foam slug stopped in the barrel, but the aluminum foil ball kept right on going, and actually reached a pretty fair velocity. However, the system was a pain in the butt, as it required not only loading two separate projectiles into the barrel, but basically required a second shot to expell the foam slug.

I am currently experimenting with, of all things, pennies. Gluing a penny to the front end of one of the foam slugs with silicone cement produces a light, cheap, and fairly aerodynamically stable bullet.

Air Gun Shot Detail

The center of gravity is far enough ahead of the center of drag to keep the bullet from tumbling as it flies, and the flat surface of the penny distributes the impact energy over a relatively large area, so it won't penetrate the skin. Pennies only weigh 2.5 grams, and the foam doesn't weigh more than a few milligrams - even at 45 feet per second (an estimate which proved to be quite low) such a bullet would carry only 0.6 Joules of energy, at the most, and this would be spread out over almost 3 square centimeters on impact.

To prove their safety, I proceeded to (intentionally) shoot myself in the hand, at a range of about two feet and a pressure of 50 psi. Think of it as "extreme safety testing". Hey, my son's safety is at stake here! Well, actually, he'll be in the tank - I guess it's my own safety I need to be concerned about!


Ouch! It stung pretty good and left a nice numb spot (don't ask me how it could sting and be numb at the same time, but it did) but no real damage, not even a bruise. In this image, taken less than a minute after the shot, I've fiddled with the color to bring out the redness at the impact site - it really didn't look anywhere near this bad. You can just barely make out the shape of the penny. After an hour, the redness and soreness went away. (I was aiming at the middle of my palm, but I missed.)

Of course, if it hit someone in the eye, it would Really Hurt™, but I've long ago decided that this is definitely a parental-supervision, saftey-glasses-and-cup type toy. With a little care, I don't think the use of penny-tipped bullets is particularly dangerous. Of course, there are lots of 6-year-olds who I wouldn't trust with penny-tipped bullets, or anything else, but my son isn't one of them.


To measure muzzle velocity, I fired the gun straight up in the air and timed how long it took for the shot to reach apex, and then how long it took to fall back to the ground.


Number Time to Apex
Time to Ground
1 1.88 60 Styrofoam
2 1.88 60 Styrofoam
3 1.92 60 Styrofoam
4 1.88 60 Styrofoam
5 1.78 60 Styrofoam
6 1.73 60 Styrofoam
7 2.00 60 Styrofoam
8 1.79 60 Styrofoam
9 3.97 60 Styrofoam
Avg 1-9 1.86 3.97 60 Styrofoam
10 1.92 4.16 65 Styrofoam
11 1.68 4.07 65 Styrofoam
12 1.66 4.13 65 Styrofoam
13 1.65 4.09 65 Styrofoam
14 1.73 4.22 65 Styrofoam
15 1.84 4.14 65 Styrofoam
16 1.77 4.09 65 Styrofoam
17 1.67 4.30 65 Styrofoam
18 1.71 4.13 65 Styrofoam
19 1.86 4.09 65 Styrofoam
20 1.67 4.23 65 Styrofoam
Avg 10-20 1.74 4.15 65 Styrofoam
21 1.52 3.68 65 open-cell foam
22 1.87 65 open-cell foam
23 2.00 4.42 65 open-cell foam
24 2.05 5.14 65 open-cell foam
25 1.95 65 open-cell foam
26 1.82 5.19 65 open-cell foam
27 1.88 4.18 65 open-cell foam
28 1.88 4.29 65 open-cell foam
29 2.00 4.39 65 open-cell foam
30 2.01 65 open-cell foam
31 1.09 (?) 3.41 65 open-cell foam
32 1.67 4.43 65 open-cell foam
33 1.93 3.99 65 open-cell foam
34 1.77 4.40 65 open-cell foam
Avg 21-34 1.82 4.31 65 open-cell foam

Doing the math turned out to require calculus, and calculus causes my brain to melt down and squirt from my ears. Instead, I developed a spreadsheet based on the Feynman Algorithm that calculated velocity and height every 0.01 second, and fiddled with the muzzle velocity and aerodynamic drag of the shot.

As drag increases, the bullet reaches apex sooner, and comes down later - as muzzle velocity increases, it reaches apex later, and comes down later. There is only one combination of muzzle velocity and drag wherein the shot reaches apex and returnes to the ground at the observed times.


Muzzle Velocity
Drag Max. Altitude
Styrofoam, 60 psi 79.6 0.0000645 63.3
Styrofoam, 65 psi 121 0.000149 68.4
Open-cell foam 123 0.000133 73.8

These results are very interesting. First is the change in drag between 60 psi and 65 psi; about twice. Note that the drag numbers are only good for comparison, and don't mean anything by themselves.

Another interesting thing is the great increase in muzzle velocity, over 40 fps, between 60 and 65 psi.

There are three possible explanations for both of these observations:

I believe the correct explanation is the first: inadequate data. If, for example, the single datum for time-to-ground from the first series should actually be closer to 4.28 seconds (only a quarter of a second longer than measured) then the muzzle velocity would be 108 feet per second. The drag comes out a lot closer to the drag measured in the later series of tests, too.

The 60 psi tests revealed a tendency for the Styrofoam plugs to break about 1/4" below the pennies if they should hit anything, even soft things like leaves. This allows the penny to tumble, rather than flying face-on, and if it happened to retain significant kinetic energy and struck someone edge first, it could do quite a bit of damage. The obvious solution was to use a stronger foam, and to that end a second batch of bullets was made using gray open-cell packing foam. These bullets proved to be much, much tougher than the Styrofoam bullets.

The grey open-cell foam bullets, however, proved to be very difficult to see against the ground, and two were lost during the testing simply because they landed among leaves and magically disappeared. None of the highly visible white Styrofoam bullets were lost, but many were destroyed.

All in all, the grey open-cell foam proved to make a better bullet. If I could find white open-cell foam, I'd be all set! I was afraid that the pourous nature of the open-celled foam would allow the air to pass through the bullet and decrease muzzle velocity, but this doesn't seem to have happened. Perhaps because the foam, being more pliable than Styrofoam, formed a better seal against the sides of the barrel - the Styrofoam actually had to be a bit undersized to allow it to pass easily through the barrel.

The open-cell foam bullets also made a much more satisfying "poot!" when they were fired, as opposed to the Styrofoam bullet's "ffffft!". At higher pressures, I even noticed a bit of water vapor from the muzzle, the result of the temperature drop as the propellant air expanded adiabatically, an effect which did not appear when firing Styrofoam bullets, though I am not sure why.

The Next Step

The next step is to aquire all the necessary fittings, assemble the system, and install it in the tank.

Update ??/??/2003 - Gun Installed

The whole system is complete and installed in the tank. Pictures will be forthcoming as I get time.

The only differences from the above diagram are:

The reason for the first was to limit how fast air would flow from the tank to the gun while the breech valve was open. Without it, I could only get two or three shots from a tank of air before the tank had to be refilled. With it, I can get ten or twelve shots. It takes about two seconds for the hose to come up to pressure after the breech valve is closed - not a problem, as it takes considerably more than two seconds to reload the gun.

The reason for the second change was to make the thing physically fit in the tank.

Update 5/11/2003 - Gun Installation Pictures

This first picture shows the gun system sitting on a table - it can be easily removed intact from the tank to "disarm" it.

Gun System

On the right end is the epoxy-filled nipple (not visible) and the pressure regulator. The breech is at the end of the hose. On the left end of the tank is the fill valve, the tank pressure gauge, and the overfill protection valve. It's actually easier to see these in the next picture:

Gun Mechanism Installed

This shows how the whole contraption fits into the rear of the left sponson.

Firing the gun has proven to be a bit problematic - pulling the breech valve lever fast enough to get a good shot results in the barrel being yanked upward into the anti-aircraft position. I am considering adding a coil spring to the valve - the spring would open the valve very fast, without moving the gun. It would also allow a more normal trigger to be used.

Update 10/31/2003 - Happy Halloween!

A friend gave me a handful of shotgun wads to experiment with as projectiles. They look like they'll work pretty good, but we'll see. They're not as cheap as pennies and foam!

Update 2/21/2004 - Paintball Gun

I am looking at paintball guns for the purpose of arming the tank. Of course, in keeping with the spirit of the project, I'd have to build one myself... but I've got a plan in mind. I've designed a blow-forward, open-chamber, semi-automatic paintball gun with one moving part. Check it out.


© 2003 W. E. Johns