Mark,
Your problem is in the choice of rivets. The "Q" rivets are self sealing in that the mandrel is pulled into the shart of the rivet and breaks off in the shaft of the rivet sealing the rivet from water entry. You will have pieces of the mandrel sticking out almost every time. To avoid this, and assuming you don't want a water tight seal is to use the "N" type rivets. You won't have a water tight seal, but the mandrel will break off at its base rather than in the middle
Graham,
I beg to differ. the frame will have some flex or movement in it, as it is not a solid structure. The true issue with rivets is the grip lenght being correct(the shaft is longer than the two pieces being joined) and the purpose of the rivet. There are an endless number of types of rivets out there. Ones for irregular holes or surfaces, plastic to metal, plastic to plastic etc. With metal to metal, the rivet should be the same material as the pieces joined if they are similar. Otherwise you could use aluminum rivets on stainless etc. and it would not hold very long, and there are also oxidizing issues. The shear strenghts of the rivets are a lot higher than most would think. The 1/8" stainless size in the 2000 lb area. The real problem is the constant changing of the forces on the rivet over time that cause them to give up their hold and start rattling. Remember, airplanes use solid rivets, we are using hollow. The other problem is the rivet isn't set correctly and the grip is less than what is expected. With the number of rivets that go into todays cars, you can bet that one or more rivets won't set correctly and will work loose and start to rattle. One trick is to add ordinary window caulking(dries clear) under the head. this way the rivet won't rattle when and if it does work loose. There are tricks to getting the load correct especially if the materials are dissimilar, such as fiberglass to metal. While installing my paneling to the rear of the engine bay window, it was impossible to pull the rivet from the fiberglass side as the roll cage blocked all access to the fiberglass at the top, which meant I had to pull from the stainless side. The mandrel would have just pushed the fiberglass aside and would not have held for 10 minutes. The solution was to add a stainless washer to the back side of the rivet and let the mandrel set against the washer, spreading the load and giving a rivet head on both sides.
The only reason I know these things is I just finished researching an article on panelling for the GTD club mag. I got a real education in the process.
Bill