Returning to the cowl outlet panel molded a few posts back. This is variable area cooling exit door. I'll detail the cooling results in other threads after test flight, so please, save those questions. This is about details of fiberglass fabrication.
The door needs a flange for sealing, actually two opposing flanges. Before cutting the door from the panel, the inside of the panel was sprayed with PVA and a four ply layup was done to net a glass section suitable for flange fabrication, as well as general reinforcement around the soon-to-be-cut hole. The section was popped loose, trimmed to overall size, and set aside.
The door was cut in the panel with a fine blade jigsaw. After the edges were sanded clean and straight the hinge line was determined and the previously made glass section was cut so as to form a flange on the big panel and another flange on the forward half of door itself. The cut flange/reinforcement parts were then bonded into the cowl panel and onto the door. Clecos are the easy way to locate and apply moderate bond pressure. The holes are easily filled during the finishing steps later.
The hinge points are phenolic rod drilled for -3 bolts. Phenolic bonds into glass assemblies very well. It's not suitable for hinges if there is a lot of movement, but will do fine here. The phenolic is imbedded in flox, then a few plies of glass is applied over the flox while wet. In the photo below, the phenolic hinges are not yet installed in the door itself.
The vertical ribs on the door started as a two-ply layup flat on the bench, merely to form a flat piece of glass which could be trimmed to shape, much as you might trim a piece of aluminum. Peel ply on both sides made the subsequent surface bond-ready without much fooling around. The ribs were cut from the cured glass, then jigged in place with a few tiny dabs of five-minute epoxy. When cured so they would stand in place, a generous flox fillet was wiped into the inboard corners, then a 6-ply layup was placed across the inside of the door and up the inboard side of the ribs.
Internal cowl pressure will apply a significant load to the full-time exit slot at the aft end of the panel, which now extends 4" aft of the firewall plane. To help with the point loads at the attach screws, 1' x 3" tabs of phosphate coated steel sheet were imbedded at the rearmost attach points. As it happens, steel and epoxy/glass are close on the coefficient of thermal expansion scale.
After everything is bonded in place, the usual filling and finishing methods make it slick.
BTW, none of this is West epoxy. There is a lot of radiant heat near the exhaust header, and exit air temperature can exceed 200F. The epoxy used here is HTR-212, chosen for its high heat distortion temperature. The door could not be expected to last if it turned into a limp noodle when hot.
The inside will get a fiberfrax layer in a selected locations, under a layer of reflective aluminum foil.
POSTSCRIPT Late December:
Test flight showed that the above didn't provide enough support for the hinge points; internal pressure bowed the panel outward and the door could not fully close.
The cure was the addition of a rib along each side of the door opening to support the hinge points. Wood is an excellent core material. These ribs are 1/4 thick red oak, but almost any straight grain non-oily hardwood would be fine here. For serious design efforts refer to ANC-18 and ANC-19.
The ribs were floxed into place. After cure the flox was sanded to a neat concave fillet and two plies of 8.9 oz 8 harness were added. After finish sanding a coat of neat epoxy sealed the surface, and later the ribs were covered in reflective self-stick aluminum as a heat shield.