It's a boat of sorts that will float.  It certainly won't carry anything beyond basic needs. Potato cannons are powerful, but this is a very different application of the explosive power. Making it float stably will be tricky. I doubt the sealed PVC tubes will be enough. I see my two best alternatives as being wood and pool noodles. The latter I can probably get from Pioneer's robotics team. The long barrel at an angle is the propulsion system. The large diameter pipe section at the top will be where the combustion (Hairspray and oxygen) takes place. The air will expand through the barrel and into the water. My bro and I tested his potato cannon (sans potatoes) in a tub of water this afternoon too. I think the recoil should be enough. Interesting to note, is that the barrel, if airtight, fills partly with water, so when we pulled the barrel out, a quart or so of water poured out! This problem will be taken care with already existing ideas.

A one-shot boat would be boring... Obviously it only has one barrel, so no multi barrel firing. The solution is indeed a self "loading" mechanism. The three pictures of the gearbox below might give an idea of how I'm doing this. More likely you'll still be clueless, and for that, there's my CAD image of the main combustion chamber as well.

The first gearbox and the original combustion chamber design:
Of what is shown in the picture on this page, It'll disassemble into two parts, the rectangular base and the cannon barrel itself. I'll probably use a screw through each of the top two red connectors to hold the barrel in position.

What's been completed so far
So what have I built so far? I've built the above gearbox, it works well with 3volts or more.  I've started building a control box for the contraption.  I've cut all the PVC pipe for the cannon as well.  What's left is wiring and machining parts for the combustion chamber.  I'll also need to do the mount for the can of hairspray.

The gearbox has a ratio of, if I recall correctly, about 32 to 1.  It uses a motor and two large gears from one of those motorized HotWheels tracks.  The other gears are K*Nex gears.  The shafts uses are K*Nex pieces as well, except for the last shaft.  That shaft is the one holding the load from the vacuum cleaner belt that I'm using; so it has to resist bending. (The K*Nex shafts will twist a bit, but I don't consider that a real problem...) The solution uses two arrow shafts; lots of those sitting around in my house.  I took one of my old wooden arrows and glued it to the inside of an aluminum shaft. Voila.

I've built the control box out of wood, in fact.  Currently I think I'll need three controls on it: ignition, gearbox motor switch , switch for a motor for a fan.  The ignition will be a piezoelctric igniter.  The big question about this igniter is whether it'll work through 50ft of wire.  The two motor switches will be wooden switches that slide side to side. I already made the first of the sliders. It conducts electricity using a thin strip of steel.  Know the best place to get those from? Cut open magnetic strips on various stuff you buy (generally electronics in their hard as hell to open plastic packaging).

The total length of 1.5" PVC used is just under 25'.  When I went to Stadium Hardware to buy the PVC, I had planned to buy 25', but bought 30' since the last 5' only cost an additional amount somewhere under $2.  I hadn't done any math beforehand, so that 25' estimate was amazingly close to what I needed.  Some of the connectors, I salvaged from an old archery target support system. The rest I bought.  The cutting of the PVC was all completed on a Saturday visit home from the dorm.  It fits together almost perfectly (i.e. a few imprecisions of >=1/16", but I've yet to glue stuff together (though I think it's safe for me to start doing so).

The CAD work that will be necessary is all complete.  I approached the CAD part of it by building the individual lengths of pipe and the connectors, and then assembling them from their separate files.  Initially, I designed the parts from memory, guesstimating various measurements.  When I returned home, the same weekend I did the cutting, I spent most of my time at home actually measuring pieces and updating the drawings.  That work payed off: every cut was right the first time.

Inspiration
I had some definite inspiration for starting this project.  The first of these was my brother's potato cannon.  He built it early during the summer with help from one of his fellow Boy Scouts.  Needless to say, firing chunks of potatoes a few hundred feet kicks ass.  In part, I didn't want to be a complete copycat, so I started thinking of other ways to apply that method.  I eventually ended up here.

The other big source of inspiration was money, or rather a gift certificate, and the need to waste it.  It's not every day you get a $50 gift certificate from school!  For that I thank my Drafting/CAD teacher, Mrs. Porter. The gift certificate (to Stadium Hardware specifically) was an award for CAD/Drafting student of the year. CAD was one of the funnest classes I had in high school...

-9-22-06