A friend of mine, Dennis Pratt, had a similar door = problem=20 apx. 3 years ago. His problem did not originate in the window area = (bolt=20 passing thru window) but was a result of a bad hysol bond between the = door skin=20 and the door stiffener in the forward lower corner of the door. = When the=20 bond separated, the door flexed and subsequently caused the window to = crack with=20 similar results to what you described on your IV-P. Dennis' window = punctured the top layer of carbon and honeycomb in a couple of places on = the top=20 of the wing - no damage to his flap. When we started dissecting = the door=20 to get access to the hardware components we found several places that = did not=20 have complete hysol contact. This was an incident that was just = "waiting=20 to happen".

I don't think you will be able to salvage the door = stiffener=20 and door skin because everything should be completely bonded together = and I feel=20 like you would spend more time trying to separate the parts. My=20 recommendation is to order the new skin and stiffener along with the=20 window. This is what I did with Dennis' repair. The = biggest=20 problem we had with his repair was making the new stiffener fit the door = opening. In order to maintain the 1/2 inch gap required between = the=20 stiffener and the door frame (gap required for the inflatable seal), we = had to=20 cut the stiffener and push out in several places to obtain the 1/2 inch=20 gap. I am not familiar with the recommendation to oversize the = bolt holes=20 passing thru the window and fill them with silicone. We = counterbored the=20 bolt holes apx. 3/4 inch in diameter and filled them with hysol as = outlined in=20 chapter 28 of the manual. In my "opinion", the silicone doesn't = offer any=20 support for the fasteners. This in itself may have placed undue = stress on=20 the glass.

Dennis has over 800 hours on his IV-P (he travels a = lot) and=20 I'm certain he has over 300 hours since the repair. For what it's = worth, I=20 just finished assembling the door on my IV-P about two months ago and it = took=20 about 3 weeks of work. We still have to fit the latch hardware = which will=20 be completed after gluing the top half on the plane and then, of course, = the=20 finish work. All in all, I estimate my door will take about half = the time=20 it took to build Dennis'. This certainly adds validity to "the = second one=20 is always easier".