I spent a lot of time yesterday sealing (for the second time in a month) a leak at the top of the window of the TB20.
It made me wonder how the hell is the system supposed to work, given differential expansion.
This is the joint at the top of the front window. There is just one screw along the entire distance, and that screw is shown below. The screw is in a “bushing” with a lot of clearance to the window, so the window can “float” around it.
The roof on the GT is composite. The black filled-in bit is the sealant; Socata use an expensive 2-pack stuff but it is a bastard to get, and to dispense unless you buy it in a cartridge which auto-mixes. It is also not very flexible. I am using Sikaflex 295 as described here
The original seal broke at point X.
The Sikaflex seal also broke at point X.
Due to the high pressure driving any water at that point, it takes only a 0.01mm gap there (basically any loss of adhesion) and you gets loads of water dripping into the cockpit.
The butt seal at X is clearly unsatisfactory because butt seals don’t work if there is any movement unless the sealant can stretch, and there is bound to be movement, not least due to differential expansion. The original factory seal also had the sealant applied on a “shear” surface here
but this turns out to be a fallacy because the window has a painted band along the top, so any sealant would only stick to the paint, and that will obviously just come off – which is exactly what happened, a few years ago. It was a visual thing only and I just left it. It obviously didn’t matter until the new break occurred at X.
If one used a sealant at X which supports a lot of expansion, that might work, but there could be maybe 2mm of expansion which is 30-100% of the material thickness (the joint width varies along the top edge, with about 5mm at the top and about 2mm at the sides).
Can anybody see what these people were thinking? Is there some highly sophisticated rationale for having just one screw at the top (and two more on each side, at the bottom)?
The other approach would have been lots of screws so the window can’t move at all, but with differential expansion “something” has to give… How is this dealt with in pressurised aircraft? I know they have relatively tiny windows but maybe not in the front – but then all the serious ones (that can go to -60C or so) are glass, not plastic. They do use lots of screws.
FWIW, this time I applied the sealant into the shear joint too, plus built it up as shown below so as to avoid a total reliance on the butt joint
To get it into the shear joint is obviously tricky without removing the whole window. It can be done by removing the one screw at the top and (obviously after digging out the old adhesive) gently warming up the upper window edge with a hair dryer. This causes the window to expand and it takes only a tiny expansion to lift the window off the composite roof by say 2mm. Then, from inside the cockpit, pop in pieces of 2mm diameter wire, spaced about 10cm apart, to support the gap, squeeze in the sealant (lots of it), and when the sealant is all in place, pull the bits of wire out. As the window shrinks back, the sealant gets squeezed into the shear joint. Not pretty but the best one can do without removing the whole window and a hangar where one can work.
Haven’t done these on TB, but on some others, normally it is the shear joint which is completely sealed as well, and I would think that makes the main difference. I think the only good way to seal it, is to remove the window completely, remove all old material and do it completely over again. Sealing only the edge is just a work around, that never last.
Remove one screw at a time, won’t give you a good one part seal, it seems to me to be a waste of efforts.
Removing can be tricky.
Citation windscreens are plastic.
There is sealer and quite a lot of screws
I just wonder how the whole “system” is supposed to work with differential expansion.
Something has to give!
One solution is to make the (flat) window so small that it can expand within the elastic limit of itself and the surrounding material. That’s probably done on big jets.
Another is to make it curved so as it expands it moves in/out. Maybe that is how this one works? But the curve on this one is only 2D – the little red dots are almost straight
Another is to have the window floating and expanding within the elastic limit of the sealant – can’t be the case here, with that 20mm wide shear joint!
The stress in that plastic, with differential expansion, must be pretty significant.
To me it looks like you need to seal it with something that doesn’t cure completely but remains liquid. The question is about the right combination of viscosity, adhesion and surface tension, though.
To get it into the shear joint is obviously tricky without removing the whole window.
That is the only way to do it. Several RV builders has simply glued the canopies in place with some RTV. It seems to hold year after year.
A Falco builder used Sikaflex to glue his screen in and now, a few years later, he’s getting problems with distortion due to the lack of expansion capability.
I wonder if butyl tape would be the answer? It never hardens and is completely waterproof.
Maybe a tape would work.
But the stuff which Socata used is very hard. You can only just poke your fingernail into it. The window cannot move at all, and the leak started only when the window shrunk away from it. Why it took 12 years to shrink away from it sufficiently, I don’t know.
BTW Sikaflex make dozens or hundreds of different sealants.
I have heard about some boaters use this stuff, (with unknown long term effect).
I´d have a look at car glass sealants. They are some black rubberlike stuff, definitely more robust than silicone and stick VERY strong. Car glass is only glued with this and some thermal effects must be there as well but no problem.
Vic
