MakerPlane Open Source LSA Looking for Guidance on Windscreen Design

[JNicol]

Hi Everyone,

Just a quick note to ask for a little bit of help with attachment of wind shield to frame on the LSA design we are in the middle of.....

Here is a render of the design. The doors are still being designed, so you can ignore the side windows that you can see directly below the wing. The wind shield includes the front clear panel, the two side angled panels and the top curved piece that conforms to the airfoil shape above the main spar in the rendering.

http://i47.tinypic.com/2db6q0z.jpg

I spoke to a couple of folks at Oshkosh and there were a couple of opinions about how the wind shield could be formed and attached. The reason we went with flat panels is ease of construction. The curved panel should be easy enough to bend on a form with a heat gun I think (I will try this soon). I have the EAA Sportplane Construction Techniques handbook as well as Ron Wanttaja's book Kitplane Construction and have delved deep into the windshield and canopy chapters.

A couple of questions for the experienced builders out there!

Now one person has indicated that it would be quite easy to form the front and two side panels as one piece and not to worry about building with the three flat panels. Might even be able to do the top curved panel incorporated and do the whole thing as one. Anyone have thoughts on that? I have no experience with forming wind shields.... Could we put filler material between the structural beams and form it on the cockpit itself? Anyone have any references to point to? Another idea we are looking at is having a one-piece front panel that is flat and then curves over the top surface. You can see the structure in the 3D model below without the cross bar that you see in the above rendering. (Note that the spars have been hidden in the image below.)
http://i48.tinypic.com/2e5nkid.jpg http://i45.tinypic.com/6sqoad.jpg

If we go with flat panels butting up against each other, what should I use to cover the seams/joints? My initial thoughts include an angled aluminium strip with silicon sealant. BTW, the structure will be composite. Perhaps we form a composite strip to cover the seams?
We have some update info on where we are at on the design in our forums here: http://makerplane.org/forum/viewtopic.php?f=22&t=84&p=428#p428

(http://makerplane.org/forum/viewtopic.php?f=22&t=84&p=428#p428)Thanks in advance for your help on this. :)

John

[Eric Witherspoon]

A couple of thoughts from my experience making the windshield for my Sonex:

1. If it's a sheet of polycarbonate, you can cold-form it (bend it in place). Down side to this is polycarbonate is pretty poor for fluid/fuel resistance. Spill any fuel on it, and it's quite likely to craze. Plus sides - you can beat on it with a hammer, don't have to worry drilling / countersinking fastener holes, and it is likely to have some impact resistance advantage. One disadvantage in the Sonex application is bending it puts stress on it, and the optics vary with stress. Then the Sonex design lays it down at a pretty low angle, so it's visibly wavy. For the more upright style of windhshield you are looking at, also with a rather large flat area, optics with polycarbonate should be ok - maybe a little wavy in the curves if you go with wrap-around, but still ok.

2. If it's a sheet of acrylic (typical aircraft windshield / canopy material), it's thermally formed. Two general ways are free-blown and molded. With molded, there's a big worry about inadvertently molding imperfections into the shape. Free-blown generally eliminates that risk, but the windshield shape you are considering here won't be made via that method.

Acrylic has very good / excellent fluid/fuel resistance. Spill some fuel on it, wipe it off, no big deal. Down side - drill it in any sort of stressed condition, it is likely to crack. And not just radiate little cracks out from around the fastener location, but crack all the way across the part. So it needs to be molded / formed to fit the supporting structure as closely as you can.

Though for what you are doing, I have seen single-piece windshields that wrap around a nice curve out front as well as taking a bend in the other direction to make a window up over the top of the cabin. So it's definitely possible.

How to attach? There's a bunch of ways. I've seen rivets. I've seen screws/nuts. I've seen screws into tapped holes. I've seen structural adhesives. Whatever you want. I used a structural adhesive on my Sonex's bubble in order to eliminate any stress risers at holes (the bubble on my airplane has exactly zero holes drilled through it).

Check out the Savor. This is a high-wing with the windshield shape I'm talking about.

From this article, he used polycarbonate, just accepting the fuel exposure risk. Pretty slick shape for a flat sheet.
http://www.eaa.org/experimenter/articles/2009-12_savor.asp

Another consideration is that by bending up a flat piece to create some curved corners, a gentle curve along the cowl line, as well as a curve that wraps up over the wing is that all these curves will be MUCH stiffer than a flat sheet. A big, flat sheet like what you are proposing might need stiffener beams behind it to minimize the bowing-in. The same thickness of material with a big, gentle curve to it is going to have much more resistance to "oil canning" (flopping to an inward-displaced direction) than if it was curved in the at-rest configuration. Look at more aircraft windshields - there aren't a lot of flat ones out there.

[JNicol]

Hi Eric,

Thanks for the info and the link to the Savor! It is definitely a cool design and I can see how the polycarbonate sheet was cleverly bent in two directions from a flat sheet to create the wind shield. If we decide to stick with flat panels, any hints on covering the butt joints? I was thinking of aluminium or composite strips glued down over the joints?

Regards,
John

[FlyingRon]

For a nice straight bend like that, you can just just make a strip heater and bend the acrylic.
If you insist on butt joints, the best way would be to solvent weld them together. Again pretty standard acrylic technique.

If you're going to make a seperate frame as depcited, why not flush mount the panels to the interior?

Be sure whatever you're using as a sealent is compatible with the aluminum as well as the plastic.

[Eric Witherspoon]

Ron has a good point - if you are going with straight-line joints, and if this part is intended to be made available from a central source (rather than each person developing their own heating/bending tooling), just get the acrylic part bent - that way there is no joint to seal or cover.

I'm thinking about "coverings" - on the Sonex there are metal strips down each side at the bottom edge of the canopy bubble. These are simply aluminum riveted on. For sealant, on my airplane, I actually ran the structural adhesive on the inside of these metal trim strips and bonded the acrylic bubble material to the inside of the metal trim, then riveted the metal trim to the canopy frame. There is no external trim at all on the forward and aft sides of the bubble. The structural adhesive can be seen through the canopy material from the outside, and I masked that to form a neat, constant-width strip. Kind of like what you see on car windshields and back windows - they are just bonded on from the outside.

There's so many ways to get things done I think you need to know more about the specific geometry you are trying to work with - will it be a flat piece "wrapped" into a couple of curves, or will it be flat panes stuck onto a frame. Either way, there's lots of different shapes that could be needed to trim it out / finish it off.

For another idea, find the BK-one website and how they hold their windshield to the frame. It's an aluminum strip (to start with), but part of it is hammer-formed into a 3-d curve, and another part is aluminum that is flat-wrapped in 2 directions. I think a good bet would be to get your windshield supporting structure in place and your clear material at least cut to fit, then mock up covering/trim with paper patterns before transferring to another material.

[JNicol]

Hi Eric, Ron,

Thanks for the info. I am actually going to be starting a full scale mockup of the cockpit within the next week or so and will start playing around with different techniques to figure it out as suggested. I will put up progress in my forum as I do it. The information provided here has given me some good starting points, which is what I need!

John

[JNicol]

One more.... what thickness material should I use.... I am looking at 3mm, or would I need to go thicker?

[Eric Witherspoon]

what thickness material should I use.... I am looking at 3mm, or would I need to go thicker?

That's where engineering comes in. Need to know what material (to know the properties). Need to know the speed of the airplane (or whatever factor is used in windshield design - to know the load). Need to know the supporting geometry, attachment method (adhesive, rivets, screws, hole spacing, hole size, edge distance, etc.). Need to know windshield geometry - wrap it into a curve will be a LOT stiffer than a flat plate... So this is where the answer gets complicated.:D

[Kiwi ZK-CKE]

I've had a bit of experience with Polycarb windscreens - including how not to do it!
My Own design Adventurer has a very large polycarbonate windscreen - cutting it out flat takes a good percentage of an 8 x 12 sheet. It is also quite tricky in that the fuselage side cant outwards, so the bottom edge tucks in tighter than the top!
2368
This was made out of 3mm Polcarb and was cold bent into place - retained by a flange on the door pillars and secured by the trim fairing around the base and angle strips inside. the problem is, I didnt do it right. The tight radius bends at the wing roots started crazing after a few months, and not long into the test programme the whole lot split open when I was taxiing in after a flight. a crack had propogated from the radius down to a small nick on the bottom edge. The tension held in a cold - bent screen is huge (it took two people to wrestle it in when I first fitted it) and this doesnt dissapate - it will be trying to pull apart throughout its life. When I replaced the screen I did things differently, and it appears as this is working better. With the replacement screen, I was meticulous about the surface finish of the edges - they were cut with a rotary cutter rather than a jig saw, and the edges were chamferred and sanded smooth with 800 wet and dry sandpaper. Before fitting, I had to hot bend it to remove the latent tension. This is really tricky as Polycarb will cloud if heated too quickly (It is hydroscopic and the water boils). According to the material specsheets, the minimum cold bend radius of polycarb is 100 x the material thickness - in other words, for 3mm it will be nearly a foot! for the top radius (Where it curves back into the fuselage top) I fount a large diameter steel pipe and clamped it in line with the start of the bend. At the free end of the bent over section I clamped a piece of timber to each side. The plastic covering was removed from areas to be heated (very important!). A friend and I then SLOWLY played two heat guns along the bend line - heating both sides at once, and constantly moving - never hold the gun in one place. eventually the weight of the timber started pulling the bend down and at the desired point heating was stopped and the panel allowed to SLOWLY cool. The side radiuses where the panel curves under the wing root were done using a similar technigue, but a steel bar coulnt be used since the radius varied from very tight at the top, to a large smooth radius at the bootom. I still used two heat guns, but concentrated on the tight bend area, and sort of "feathered" the heat application. When ready I Bent it carefully by hand, pressing the bend line wearing leather gloves. It wasnt perfect, but the latent stresses were essentially completely removed.
As far as geometry goes. Mark up the curve profiles you need and divide the curves into equal divisions. Measure the distances between key points using a piece of string, and triangulate the screen from there - I got mine almost exactly right first time using this method.
Another useful pointer is if you are transitioning between flat and angled, you can bend 3mm Polcarb in a sheetmetal folder and it will come out very neatly - this also stiffens the panel and reduces drumming. The rear cabin window on my aircraft was done that way:
2369

I'm not an expert, but I hope experience will be of use!

[Kiwi ZK-CKE]

Oh, and in response to your original query, heres what I would do...
Make it one piece - the top is a big enough radius to be cold bent, the side pieces can be bent on a sheetmetal folder. No joints to worry about sealing, nice and rigid and no need to try the tricky heat method unless you are that way inclined....
3mm Polcarb will be fine - My screen is huge and only starts to shake at about 135kts - It is the limiting factor on my VNE!

[JNicol]

Hi Kiwi,

Thanks very much for a comprehensive and awesome answer....! Nicely explained. I will certainly give this technique a shot and report back here. It has probably saved me a lot of time, effort and band-aids. A couple of questions on the sheet metal folding technique used on the rear windscreen, just confirming that you did that cold? Also, did you see any fracturing or was it a clean bend? It looks like it will work on our front screen for sure.

(BTW, found a link to a home-made brake that looks like it would do the job... http://www.ablett.jp/workshop_archive/give_me_a_brake.htm)

[Eric Witherspoon]

A couple of questions on the sheet metal folding technique used on the rear windscreen, just confirming that you did that cold?

He's doing that cold. One other point I didn't include in the earlier post about the two common materials - polycarbonate, though tough, attracts scratches like nobody's business. Just cleaning off bugs - scratches. Look at it funny - scratches. Acrylic is more scratch resistant.

One other thing you might want to notice about Kiwi's windshield - the big front section is curved along the bottom edge. Meaning most of what you look through is in fact a curve, and not flat. Just the slight curve is much stiffer than flat. Mock it up with poster board and poke it with your finger.

You might want to consider a curve shape to the top of your firewall in order to get some stiffness into your windshield configuration. Or before you build it on the airplane, at least mock up this windshield on a frame, mount it on the roof of a car, and get it up near 100mph. With some video on it, you can see whether you really want a big flat panel or not.

[JNicol]

It might be that Kiwi's method of construction as described is actually going to be easier and more robust than three flat separate panels, particularly if the brake method works for us. If we allow for a small bow in it when we attach the whole thing it might be enough to stiffen it. The intention is to prototype this on a non-flying full scale model. Blasting it with some wind would certainly help to see if it is wobbling at all.

[Eric Witherspoon]

Good luck - sounds like you're on a path to success. Might be a nice wrap-up to this thread if you could post some photos of what you end up with - and any other comments with windshield development you find relevant.

[Kiwi ZK-CKE]

yes the rear window was bent cold. Its a technique used quite a lot for making clear shelving and display stands and creates a very good result - no crazing or cracking as the the fold is supported during the process. Leave the protective wrap on when doing this - it may also be beneficial to put some masking tape on as well to prevent scratches, because yes, polcarb is really scratch prone!

post pics of how you get on!

[JNicol]

Thanks Eric and Kiwi!

I will certainly post pix of progress, testing and the final installation to wrap things up!

Cheers!