Now that I'm sure you're clear on the concept, let me take you through the process of my creation. Obviously the first thing I would need was the gun itself...
Realizing that I wanted reasonable speeds while still keeping chamber pressures low, I decided that a "large" bore gun (used with a sabot) would be in order. The local hardware store had a 4 ft long piece of 2 inch plumbing pipe that seemed pretty smooth on the inside (read: I couldn't see the seam) so that's what I went with. Add an end cap with a hole drilled in it and you've got yourself a gun suitable to the task at hand.
With the gun bought and paid for, it was time to build a rocket...
It was immediately clear that any rocket subjected to the extremely high g forces associated with a gun launch would have to be a bit different than rockets normally seen in the rocket section at your local hobby shop. Such a rocket would have the following properties:
The "standard" high strength body tubes used in hobby rocketry that I am aware of are fiberglass. If commercially purchased, they are too expensive (for my budget). If home manufactured, they are too much work for a rocket that is most likely only going to be flown once.
Thinking about alternative materials to use for body tubes, I remembered a piece of very thin walled PVC pipe that I had once seen. After a short search of the local building supply industry, I came up with the perfect body tube: 3/4" (nominal) thin walled PVC pipe. It turns out that such pipe is nowhere near 3/4" ID. Rather, it has the same OD as 3/4" schedule 40 pipe but with the thin walls, ends up with an ID of about 23 mm. Now, the Estes D engine that I had selected (for pretty much no reason at all) measures, as I'm sure you all know, at 24 mm OD.
Thus, it would be impossible to put the Estes engine in the PVC body tube, right? Wrong. One characteristic of PVC is the fact that it has a very high coefficient of thermal expansion. It turns out that if the tubing in question is heated to somewhere around 140 F...(well, I'll be honest, I don't know the temperature, but I was still holding the pipe with bare hands)...that it's ID increases to well over 24 mm. Thus the engine slides right in. Cooling the body tube back down after motor insertion results in a very tight fit. As in, so tight that I could not budge a motor so installed with every ounce of strength I had. A thrust ring or other motor mount was deamed unnecessary. Cool.
But how do you attach fins to a PVC pipe? Well, if you have PVC fins, you just use the appropriate cement. But I know of no suitable flat pieces of PVC! What to do? What to do?
I decided to go with a finless design. It made sense such a design has very little drag, and best of all is very easy to manufacture. The only downside to a finless design is that they often require a lot of noseweight to maintain stability.
Given that the rocket would fly only once, I decided that the best way to get around having to add excessive amounts of nose weight would be to move the engine all the way forward and go ahead and burn the inside of the body tube (see sketch below). Yes, this would result in an decrease in thrust, but I hoped that this decrease would be in part be negated by the fact that the design was finless.
Given that the PVC body tube was not of a standard size, I also realized that I would need an alternate nosecone. I tried a number of materials and techniques, but what I finally decided upon was plaster of paris poured into a disposable mold (That actually formed the external surface) made out of old plastic milk jugs.
The sabot turned out to be easy. A 1.25" PVC endcap, 3 inches of 1.25" schedule 40 PVC, and a 1.25" coupler made a rather nice, easy to manufacture sabot (Albeit a bit more massive that I would have liked). If you want more details, read the construction page.
The first test flights involved nothing more than tube (not gun) launching my rocket design from a two foot piece of 1.25" pipe. The only objective was to prove stability.
The rockets flew straight as an arrow. Unfortunately, despite only weighing in at about 85 grams, they only went about 100 ft up....the thrust loss was much greater than I had anticipated. Still, the tests were considered a sucess as the rockets had flown straight and true.
The objective of the second set of flights was to prove the gun, sabot, and motor ignition.
The gun performed flawlessly. Frame advancing through video tape of it's operation showed sabot/rocket separation within about 20 feet of the muzzle and a velocity guesstimated to be somewhere in the neighborhood of 550 ft/s (mach 0.5).
Unfortunately, none of the motors ignited. We (My uncle, who came along for the show and provided the shuttered camcorder) theorized that the air in the body tube was getting compressed up around the nozzle opening and forming a barrier to the really hot gases associated with the gun charge.
More bad news arrived when we found some of the post test rockets. The engines had been blown forward through the nosecone despite the seemingly tight fit of the engines during construction. We agreed that the most likely explanation was that the body tube was stretching as it was filled with high pressure gases. This would release the tube's hold on the engine and it would be free to blow it's way through the nosecone. Clearly, thrust rings would be required in the future.
For the third test series, I went back to tube launching to address the issue of thrust reduction due to the Kushnik (or however it's supposed to be spelled) Effect. The Kushnik Effect is nothing more than the formation of a vacuum via venturi effects inside a body tube with a recessed motor. The result is a large pressure imbalance on the rocket body that, technically speaking, causes a huge (speaking from experience there!) drag increase that is more commonly viewed as a thrust decrease.
In order to beat Mr. Kushnik, you must break the vacuum. I reasoned that some holes drilled into the body tube just behind the nozzle exit plane should allow external air into the body tube to replace air "sucked out" by the exhaust plume's venturi effects (see sketch).
The holes worked! I launched two rockets of this design. Both flew like a bat out of hell, never to be seen again.
As an added bonus, it was felt that the holes would solve the motor ignition problem, but just to be sure, I applied a FFFG BP and dextrin goop to the nozzle throat of all rockets to be gun launched from this point on (Note: The BP/dextrin, it was hoped, would also provide a pyrotechnic delay so that the motor wouldn't ignite until after the rocket had cleared the barrel.).
I had high hopes for this series. I felt that if only the engines would ignite, then my quest would be over. Silly me.
The engines did ignite. Flawlessly. Every time. Unfortunately, the rockets displayed a stability problem. This is not to say that they were unstable, but to say that they corkscrewed badly. Further, I noticed that there was a direct corrolation between how well I felt I'd drilled the holes (read: how evenly they were spaced) and how well the rockets flew. I realized that the boundary effects of the incoming unevenly spaced air currents were creating a crude thrust vectoring effect. If the rockets were to ever fly straight they would require a better level of quality control than I felt I could achieve in my poorly equipped garage. I thought long and hard about it before deciding that I would need a new scheme for stabilization (Note: I still feel this was a good design, it just requires a better equipped garage than I happen to have.).
After several weeks of attempting to affix fins to the basic rocket design I had a revelation. Why not use the very thrust vectoring effects that were hurting me to my own advantage? I decided that I would purposely build a rocket that would use a crude thrust vectoring system to spin up the rocket and thus, spin stablize it.
I ended up cutting the body tube off about one inch behind the nozzle exit and cutting four longitudinal slits around this section. Then, using a pair of pliers (How's that for precision!) I bent the corners of the "tabs" in toward the expected path of the exhaust. The result was four very crude thrust vector vanes (see sketch) that I hoped would last just long enough to spin the rocket up and then burn away so that they wouldn't sap any more energy than was absolutely required out of the motor.
Sadly, I can not report on whether or not the vanes burned away as I had hoped as no rockets were ever recovered (Going...Going...Gone!). They went up with measured muzzle velocities ranging from 650-800 ft/s (If you really must know how it was measured, feel free to email me and I'll let you know...I just don't feel like discussing it right now as it has little to do with the rocket itself.). This may sound like a lot of variation, but considering that I was merely eyeballing the BP charges, I think it's actually quite good!
In any event, as I have implied, the rockets flew straight. Yes, they corkscrewed, but it was a very tight, very fast corkscrew.
I am pleased with the design: It flies reasonably true, it costs less than $0.50 each (neglecting motor), and it takes only 15-20 minutes to build. It provides excellent bang for the buck. No, it isn't perfect, but it's close enough that I believe further efforts will require much work with little benefit.