Celebrating the independent kiwi spirit of invention.

AIR CANNON Page 3:
Construction

Back to
Air Cannon page 1:
Pipe and Safety Considerations

Back to
Air Cannon page 2:
Design Considerations

On to
Air Cannon page 4:
Cost, Specifications, Results

On to
Air Cannon page 5:
Gallery

Attempt One

As a trial cannon I decided to use pressure pipe offcuts of 65mm for the pressure chamber and 40mm for the barrel (they were the only two sizes lying around). A ball valve was donated by a friend for the exhaust valve on the back of the cannon to trigger the firing, and I had a car tyre valve lying around to use for the air inlet. A tyre wholesaler friend donated an old tractor tyre inner tube as a source of sheet rubber.

A trip to a local plumbing shop provided the extra components required, namely, an end cap, adapters to reduce the size at the front of the pressure chamber, pressure pipe glue, and all the nipples and stuff to mount the ball valve on the end cap.

I made a diaphragm for the back of the barrel from a sheet of 2mm polypropylene (or maybe it's polystyrene) with a thin sheet of rubber on each side, trying to cut it as circular as I could. The back end of the barrel I sanded completely flat to make a good seal with the diaphragm.

With the diaphragm loose inside the cannon, I put a couple of 20mm thick layers of open cell polyurethane foam (found in the back of a friend's truck) behind the diaphragm with the idea that it would act as a spring, and push the diaphragm back against the barrel after firing, ready for the next pressurisation.

Using a handy air compressor, the pressure chamber (the 65mm pipe) pressurised wonderfully, with the diaphragm sealing the back of the barrel perfectly. However, the diaphagm would not blow backwards when the ball valve was opened, and all the air in the pressure chamber was vented out the ball valve. Sawdust on the ground (which the barrel was aimed at) was not disturbed at all by the triggering of the cannon.

The diaphragm was probably not a good enough fit on the inside of the 65mm pipe considering the ratio of the cross-sectional areas of the pipes. There just wasn't enough kick to move the diaphragm in a hurry, but I think the foam was also to blame. Just using gravity to put the diaphagm back in place would have been more sensible, as there really wasn't very much room at all behind the diaphagm, certainly not enough for two layers of foam.

While the cannon pressurised, an interesting popping noise was produced by ramming a broom handle down the barrel. This could be done 4 or 5 times before all the pressure was released.

The failure to fire meant I had to saw the end off the pipe several centimeters down from the end cap – enough to put back together later with a 65mm pipe socket (a double ended joint). The foam was removed and I returned to the plumbing shop to buy a socket.

Attempt Two

Being a long-term project, about a month later I was going over the design with a friend (the one who suggested the exhaust pipe lagging around the pressure chamber), trying to come up with something better fitting than my hand-made diaphagm. After trying just about every plastic container and lid in his house he found a peanut butter jar lid (!) that was exactly the right size to fit inside the 65mm pipe, and with a piece of rubber sheet glued to each side, made a perfect diaphragm/piston for the end of the barrel. The lid is slightly tapered, and allows air to slowly escape around its edges for pressurising, but when the moment comes for firing, the lid is forced back in a great hurry, and with the rubber on its rear surface seals the ball valve outlet quite nicely.

Not wanting to glue the cannon together, wait a day for the glue to dry, then have the possibility of it not working, we bound the end cap on with duct tape then pressurised the chamber with my friend's one-size-fits-all bike pump. I pointed the business end of the barrel into a bucket of water and pulled the ball valve open.

What happened next was not quite expected considering my previous construction, as the cannon actually worked this time! We suddenly found ourselves covered by half the contents of the bucket of water. Success!

Unfortunately, all the pumping had weakened the seal around the tyre valve, which had developed a slow leak, so I couldn't glue the cannon together then, needing to get at the inside of the end cap.

Another test, though, actually firing something. A small water bomb (to fit the small barrel) poked into the end of the barrel would be ejected a couple of meters (intact) when the ball valve was opened. How far they would have flown if they had accelerated down the barrel I don't know, or if a decent pressure had been built up before firing. (As it was I blew air into it until the friction fit of the end cap gave up and the cap popped out into the slightly loose duct tape – still holding pressure but nowhere near maximum.)

Attempt Three

I decided that the rubber tyre valve was not going to last long term because of the movement it allowed – doesn't work too well with a hand pump – even if I resealed it. So I pulled it out to replace it with a compressed air snap-fitting. Unfortunately, I overtightened the adapter nipple while screwing it in and cracked the end cap. Now I have a 65mm socket that hasn't been used and an end cap that needs replacing.

I think I'll use a metal tyre valve with nuts in the next end cap.

For my next cannon I think I'll have a pressure gauge – always a good idea for those times the air source pressure is unknown.

Attempt Four

A metal tyre valve from my tyre wholesaler friend means a hand pump is a real possibility again. This is convenient since deserted shooting areas often don't have air compressors handy. Also new is the endcap, since I managed to crack the last one (see Attempt Three above).

I didn't have my thread tape with me when I put the nipple on, but I figure epoxy resin should do the job, and the thing is done up pretty darn tight anyway – in the photo above you can see the scuff marks from the vice.

On the inside, I've abandoned the plastic sealing washer and, yes, put the tightening ring on backwards. It looked like too good a surface for the diaphragm to seal against not to do it. Actually (as you can see in the photo below) it wouldn't fit the other way because of the metal tyre valve.

Putting it all together and heading off to Motoihe Island in the Hauraki Gulf for a relaxed day testing was the easy part. Getting any pressure in the cannon was the hard part. I think the bike pump I was using lets the air out on the back stroke. Occasionally it worked and this I think was due to the valve centre pin being bent sideways, missing the thing in the end of the pump hose that for some reason depresses the valve pin. Certainly depressed me.

However, when it did pump properly the diaphragm seemed to be leaking, resulting in the potato popping out of the end of the barrel without any great pressure behind it. I couldn't find any good reason for the diaphragm leaking. Maybe it was the way I was holding my tongue.

Attempt Five

A bit of modification of my pump and a little bit of work on the barrel and suddenly all the problems are fixed. Specifically, I plugged lumps of folded rubber from the inner tube around the inner end of the barrel to ensure the thing stayed in the middle of the pressure chamber. This means the diaphagm always sits neatly centred on it. Not a problem getting it to fire reliably, but I think a better pump is in order.

There is remarkably little recoil of the cannon considering that the potatoes we were using to test the thing were traveling so far. I wonder if this is inherent to the back-venting design.

So with the cannon working at last, this chapter in the saga is closed. Head on to page 4 for costs, specifications, shooting range, etc.

If clarification is wanted on any of the above, please do .