A question to be investigated is bond strength to aluminum and to the fiberglass section.
Bill, the urethane rubber is itself quite adhesive, but when cured, bonding it to anything else can be a problem. Smooth-On makes a urethane adhesive for the job (Ure-Bond II), but they don't make strong claims for it.
I reasoned that bonding a fabric reinforced urethane to an inflexible metal ring was less likely to be successful (long term) than bonding it to a flexible rubber ring...a matter of stress concentration.
So, first I mounted the ducts on the engine and trimmed them a little short of the cowl.
Then I machined aluminum reinforcing rings, 1/16" thick with a "L" cross section. They got a few dabs of Ure-Bond on the outside, and then they were slipped into the ducts.
The cowl was added immediately, and the rings were pulled forward so they were against the inner surface of the cowl's glass inlet.
When cured, the aluminum rings were adhered to the urethane duct...not strongly, but well enough to precisely set the position of the duct ends, so I could next cast the rubber rings. The ducts were set nose down in a machined nylon mold. I painted some Ure-Bond on the exposed urethane in the casting area, then poured in liquid urethane. I reasoned that the co-cured Ure-Bond would improve the interface between the cured and uncured rubbers.
Note that the glass reinforced urethane is on the
inside of the cast rubber ring; inlet pressure cannot exert a peel stress. The joint is pretty much just shear, something the rubber does well, in particular when the cast rubber keys into the fabric weave pattern on the duct surface.
After cure the aluminum rings were easy to remove. There is no need for them to be fastened to anything, as they cannot escape. In service, they eliminate any possibility of the rubber ring collapsing inward, and they look nice as polished trim in the inlets.
Overall, it's an entirely repeatable procedure. I actually made two sets of ducts, the first being multiple plies of 8-harness, and the second being a single ply of 9 oz plain weave. The second set incorporated a spigot for the prop governor. Here the first four photos were taken while doing the original set, thus no spigot. I still have all the molds, and can make another set when these wear out or fail. Right now (five years, 550 hours) they're doing fine.
The cowl itself got matching flanges cast from micro and flox, but that's another story.