What's new
Van's Air Force

Don't miss anything! Register now for full access to the definitive RV support community.

Windshield Installation (Sika Process)

Jwputnam

Well Known Member
I have completed my canopy using the Sika process and I am moving on to the windshield. I have viewed several builders web sites and I have searched this forum and the web for instruction, but I still have several important questions that I could sure use some help with.

My plan is to attach the windshield with spacers to the roll bar and the forward deck with primer and Sika, then allow to cure before attempting to draw the sides in and gluing them.

The attachment of the windshield to the roll bar is identical to the process used on the canopy to its frame. The use of spacers and back filling with Sika is easy enough on the roll bar, but I am unsure of what is necessary for the forward section and eventually, the sides.

The Sika process seems to demand that there be a gap between the two primered surfaces for proper adhesion. I suppose that a small gap can be set for the forward section where the windshield attaches to the front deck. The forward section of the windshield will be primed on both sides for perhaps a 1/2? strip and Sika then applied with a continuous bead to both sides. (It may be that no gap is necessary in this case. I am unsure. For that matter, the forward section could simply be held in place by the fiberglass laminate.)

Considering that the windshield is fast at the roll bar and the forward section, the bulging sides need to be addressed. If the sides are to lay flush and be faired seamlessly without an obvious bulge, spacers cannot be used. How to apply the Sika? Do I prime the inside of the windshield sides and the fuselage and then simply inject Sika or smear it into the gap before clamping down? For that matter, if the roll bar and forward section is fast, can I eliminate the side bulge at all?

As you can tell, I am a little lost here and I need some direction from someone who has successfully accomplished this. It may be that I am making very incorrect assumptions and not attacking this properly at all.

Thanks for any help.
 
I did as you are doing. I used temporary clips held with clecos to hold the front and sides. I primed and applied sika and skipped the clips. I ground the windscreen to a taper with a belt sander first to allow a larger surface. The point of the screen touches the metal. This method removes the bulge the first pass. Next day remove clips and apply sikaflex to the small areas left.

When dry I trimmed with a sharp knife the sticky outy bits. Then fibreglass over the joint. Mine has not flown yet but appears to be solid.
 
I gather that you used the belt sander on the entire forward edge, including the sides. I am afraid to death of the belt sander. What grit paper did you use? How wide was the final finished edge? Did you polish the edge to prevent stress cracking?

I am surprised that the clips would hold in the sides. I have not tried pulling the sides all the way in yet, but my first attempt would indicate that the amount of pressure required would be significant. I see that others have used a 2 x 4 and big clamps.

I assume that you primed the roll bar, inside of windshield and the fuselage in advance. Did you place a bead of Sika onto the entire inside (primed) windshield edge (except for the roll bar area) and then place the windshield on the plane and clamp down? (This would probably have to be accomplished very carefully by two people.) Otherwise, you could only add the Sika by injecting a bead from the inside of the cabin, which would be a little difficult.

I am also assuming that you did not use spacers along the windshield forward and side edges since you mention that the point of the windshield touched the metal. I would guess that you relied on the gap behind the ?point? to supply that requirement.

In truth, it might be debatable that the forward edge and sides need to be glued at all since you are holding them in place with a considerable epoxy/glass laminate, correct? The Sika gives you additional assurance and perhaps some flexibility.

Am I understanding correctly? (Thanks so much for your help.)
 
If the epoxy/fiberglass fairing is actually bonded to the plexiglass, the Sika adhesive does little or nothing structural along the bottom of the windscreen.

The edges of the plexiglass should be rounded, and sanded/polished to a fine finish, i.e. as near to polished as possible.
 
Dan,

I agree with you. The edges need to be polished smooth so as to avoid any future stress cracking, but I still don?t know how to adhere the sides. I can clip the forward edge without much difficulty and I can Sika Flex the windshield at the roll bar easily enough. The sides, however, would seem to be a problem. I am not sure that they can be held in with clips and a 2 x 4 bracing system would interfere with the laminating process.

Any suggestions?

I wonder if it would be possible to scuff the windshield sides on the interior and scuff the fuselage as well, then epoxy them together? That might be messy and dangerous. Maybe Sika would work using primer on both surfaces, but not using spacers?

Still thinking, but I believe that Sika on the roll bar with spacers and epoxy/glass laminate on the forward area is the way to go. I just don?t know how to attach the sides.
 
.... but I still don’t know how to adhere the sides. I can clip the forward edge without much difficulty and I can Sika Flex the windshield at the roll bar easily enough. The sides, however, would seem to be a problem. I am not sure that they can be held in with clips and a 2 x 4 bracing system would interfere with the laminating process. Any suggestions?

Sika vs fiberglass...there are obvious advantages to Sika. My concern is with mixed attachment methods. With Sika, I'd suggest a separate fairing layup which is not epoxied to the the aluminum boot cowl or the plexiglass. After the windshield is installed, cover the area with clear tape, do a fairing layup, pop it off, finish it as desired, and stick it in place later with a bead of sealant or Sika. Now the windshield can expand and contract without constraint, one of the reasons to use Sika.

As for pulling in the sides, some of that is normal. However, be sure you have the front edge trimmed properly. When you pull in the lower rear corners of the windshield, as it would be bonded, is the edge parallel to the roll bar? Is the sliding canopy edge parallel to the roll bar?
 
Dan,
The sides, however, would seem to be a problem. I am not sure that they can be held in with clips and a 2 x 4 bracing system would interfere with the laminating process.

Any suggestions?

While different than yours, I just used Sika to install all of the plexi on my 10. I cleco'ed the plexi in place. I used a standard #10 washer under the plexi at each cleco location, along with other strategically placed spacers ( 1/4" squares of aluminum attached with super glue). It works well. If you are cover the area, no need to address the 3/32 holes remaining. If not, you can used the sika flush in the hole and is not really visibly distracting. You wont need many clecos. Just be sure to push out the excess sika to make the plexi where you want it between clecos. The sika isn't very viscous and will cause the plexi to bulge between clecos if you don't push on the plexi to force it out.

As Dan mentioned, the sika will allow the plexi to expand and contact and any f/g fairings over the edges could be problematic. I have already seen the sika bulge to accomodate expansion.

Larry
 
Last edited:
I would fiberglass the front, drill the sides and sand the edges. In my case I Sika Flex the front, drilled the sides and Sika the roll bar. I have the tip up canopy. I believe if I only would have used fiberglass on the front it would have been easier to prepare for paint and finish. Just my two cents.
 
I did the fillet around the windscreen on my tipup using Sika, with no fiberglass at all. Very easy to sand, using sandpaper wrapped on a large dia PVC pipe. Unfortunately, I can't tell you how it's surviving, because it's back in the crate, waiting on me to finish the rest of the plane...
 
Sika vs fiberglass...there are obvious advantages to Sika. My concern is with mixed attachment methods. With Sika, I'd suggest a separate fairing layup which is not epoxied to the the aluminum boot cowl or the plexiglass. After the windshield is installed, cover the area with clear tape, do a fairing layup, pop it off, finish it as desired, and stick it in place later with a bead of sealant or Sika. Now the windshield can expand and contract without constraint, one of the reasons to use Sika.

As for pulling in the sides, some of that is normal. However, be sure you have the front edge trimmed properly. When you pull in the lower rear corners of the windshield, as it would be bonded, is the edge parallel to the roll bar? Is the sliding canopy edge parallel to the roll bar?

This is how I did mine.
 
(snip) If the epoxy/fiberglass fairing is actually bonded to the plexiglass...

Therein lies a potential problem. Is it in fact bonded? When I was building I saw numerous local RVs in the blazing-hot-in-the-summer/cold-in-the-winter Sacramento Valley with delaminating windshields. When I got to that point, having used Sika for the sliding canopy, I conducted some experiments that convinced me to use the Sika primer not only for Sika 295UV but for improving adhesion between West Systems epoxy and aluminum/plexi. Somewhere on this forum I reported on that back in the day. Unfortunately all of the photos I took of my sample coupons went away :mad: In any case no way could I make as strong of a bond between epoxy-laced fiberglass and bare acrylic, regardless of treatment, as I could with primer-treated acrylic. No comparison. Former all adhesive failures, latter all cohesive failures.

When windshield fairing time rolled around I first bonded the windscreen with 295UV using the usual process, then primed the areas getting epoxy with the same (forget the part number, have had a couple of beers) Sika primer. My huge sample of 2 (ha) is so far holding up. One with 9 years and numerous thermal cycles, the other not so much, but hey it hasn't cracked.

As for initial retention prior to applying any adhesives, I used the Van's suggested method of fabricating some tiny aluminum clips that I pop-riveted to the skin. A couple up front, a couple along the bottom back by the roll bar. Made some crude spacers over the top out of some urethane tubing I cut up, to space the windscreen off the roll bar, then a bazillion spring clamps to secure it. Sika, let cure, epoxy and fiberglass on to primed aluminum and plastic. Used carbon over the top in the hopes of adding a little strength. Humans can hang from the windscreen fairing at the top of the roll bar, it won't delaminate or crack. Has been tested a bunch.
 
Sometimes it helps to quantify things. The engineers all know, but other builders may not realize how much plexiglass expands and contracts with temperature change, as compared to the materials with which we may attempt to bond it. Quick concept tutorial...

Coefficients of thermal expansion, 10^-6in/(inR):

polycarbonate 65-70
aluminum 21-24
steel 11-12.5

So, plexiglass has about three times the expansion rate of aluminum and six times that of steel.

Assume a flat sheet of plexiglass 30" x 30", bonded to a steel frame at 70F. Raise its temperature to 120F, and the plexi grows 0.000065" x 30 x (120-70) = 0.0975". The steel grows 0.000012 x 30 x (120-70) = 0.018". The difference is 0.0795", more than 1/16 of an inch.

Sika is low modulus, a label meaning it will easily stretch. In the above example, the Sika would allow the plexiglass to (more or less) freely expand, resulting in little stress on the plexiglass or the steel. The desired stretch behavior is why a Sika installation requires spacing the plexi away from the frame prior to placing the adhesive. Without thickness, shear stress on the Sika would be very high. The thicker the bead, the lower the stress on the Sika and its adhesive bond.

Same example, but this time bond the plexi to the frame with a structural adhesive which cures hard. Or if desired, fasten with rivets or screws. When fastened with an inflexible (high modulus) joint, the difference in thermal expansion must stretch the steel and compress the plexiglass, placing stress on both. The frame, plexiglass, and the joint must all be strong enough to withstand this stress.

In the fully bonded case, the practical point of failure is usually the adhesive bond, as Lars described above. Screws and rivets add a twist. A hole in a sheet typical raises local stress by a factor of three. Given expansion stress, the practical point of failure is usually a crack starting at a hole, a product of that tripled local stress and the tiny nicks and sharp edges left by the drill bit.

A mixed fastener approach quickly gets complicated. Assume a standard RV windshield install, with the plexiglass bedded into flox at the boot cowl junction. The subsequent epoxy/glass fairing layup overlaps the plexiglass. Given thermal expansion, the hard flox foundation is generally in compression, relieving much of the shear stress at the fairing overlap joint.

Now do one in which the windshield is bonded into place with Sika, followed by an epoxy/glass fairing placed in the usual bonded manner. Given thermal expansion, the low modulus Sika yields to the expansion stress. It provides no significant support to the epoxy/glass overlap joint, which must bear all the stress. As Lars noted, if the bond isn't really good, it may fail. And if the bond is good, but the fairing is thin, it may buckle and/or crack paint.

So, go one way or the other, but go all the way when you do...no mixed method. Either hard bond all the way, or float the plexiglass on Sika and then float the fairing on top of it as a separate part.
 
Last edited:
I completely get what Dan is is telling us and I intend to produce the windshield with the Sika process in its entirety, but I am not exactly sure how to accomplish this.

As i stated before, the roll bar area with spacers is simple enough, but how about the forward windshield, the sides and ESPECIALLY the final fairing?

I intend to have the windshield edges polished so as to avoid any chance of stress cracking and I am NOT going to drill any holes in it for any purpose at all for the same reason.

I talked with the Sika lab and Sika provides NO adhesion if a space of 3 to 5 mm (call it 1/8") is not provided.

Soooo.....here are the immediate issues:

1. Can I sand the edge of the forward windshield where it will contact the deck so that the primer will adhere? I will then use 1/8 spacers for the Sika introduction. (Aluminum deck will be sanded and primed as well.) If sanding that edge is not possible, will a polished edge provide proper adhesion?

2. Using 1/8" spacers on the sides after drawing them in will create a significant bulge that may be difficult to cover with the fairing, but i don't see that there is any way to avoid this and still keep with Dan's formula.

3. I do not see how the fairing can be fixed to the entire windshield once it has been produced without again creating an eight inch gap for the Sika, unless I simply epoxy on, but then I have defeated the whole concept of floating the plexiglass, haven't I?

I truly appreciate (!) all of this help, but I am a little confused as to how I should proceed. My apologies if I am stretching your patience!

PS I suppose that the fairing can be epoxied to the aluminum deck BELOW the windshield and then the Sika applied between the fairing and the plexiglass.
 
Last edited:
Sika

I get the 1/8" but does it have to occur at the exact edge of the plexi?
Seems like that 1/8" bead could and may well fit better, just inside the edge.
Not being judgemental here. I am right behind you so I am following closely.
I plan to Sika the windshield, layup the glass, pop it off and sika the final skirt in place.
 
Dan makes some important points about thermal expansion. In this case the 3-D curves of the canopy and windscreen help us. Consider a windshield fixed at both ends of the (curved) roll bar. The windscreen is growing away from the rollbar since the plastic is getting longer faster than the steel, but (assuming a constant radius curve) the gap is more like 1/3 (as in pi, not pie) of the linear growth.

More recollections about my installation... I left something like a 1/4" gap all around the base of my windshield. I used more crude vinyl (or was it urethane) spacers made from little squares cut out of plastic tubing to maintain that gap. I did the Sika application in a couple of steps, removing the spacers and filling the subsequent holes. I attempted to create enough of a fillet at the acrylic/aluminum intersection that the fairing glass wouldn't have a sharp-ish edge adjacent to the acrylic/aluminum interface that could lead to a crack propagating from the inside out. Probably not that important but it was easy to do with the Sika and a gloved finger.

It's 90?F out on the ramp here in the Sacramento Valley today (thankfully the office has a/c); just went out and did a close look at mine for the umpteenth time. Still looks ok other than the less-than-perfect radius blend in a few spots due to the horse-is-getting-near-the-barn syndrome I was experiencing around that point in the build :rolleyes: Again, mine is a sample of one, thus of no statistical validity. But it is holding up.
 
I found a couple of photos. Not great (digital cameras have come a long way since I did this) but maybe slightly helpful.

This one shows aluminum clips I referred to in a previous post. Along with the plastic spacers (not shown here) they were removed prior to glassing. Drilled out the rivets. Unfortunately the gap (done for the requisite compliance required by the Sikaflex) at the bottom is impossible to see here.

i-rZ8Zvdx-XL.jpg


First layer of Sika added, clips still visible, can't see vinyl spacers but they were almost certainly there still.

i-stBk6MS-XL.jpg


Not sure how helpful this is, but the clips are gone and I've done the first layer of glass. The short vertical strips are peel ply.

i-xsd9Sgd-XL.jpg
 
On my not-yet-flying tipup (all aluminum frame), I used the 1/8" shims around the top of the rollover structure and matching rear bow of the canopy, but allowed the 'glass to touch the aluminum as it approached the top longeron, and the side rails of the canopy frame. The back glass is bonded directly to the window cutouts, but also has 1/4"+ fillets around the edges. The base of the windscreen sits on the frame, but is secured by a large external fillet of Sika, which is intended to also replace the fiberglass fairing, and an internal fillet. The areas along the bottom of the plexi touching the frame side rails do have a very thin layer of Sika, but there's also a thick fillet at the top edge of each horizontal member on the sides, and in the 'step' in those horizontal members.

I realize that hope isn't a plan, but my working 'hope' is that the fillets will more than provide the 1/8" of flexible adhesion needed, even if the thin layer of Sika does shear. It's there more as temporary bond/stabilizer until the fillets were added. If this is a false hope, I'd love to hear about it now. :)

I'm certainly not advising anyone else to follow my lead; it's not yet tested in flight.

Charlie
 
Last edited:
Very interesting Lars.

Did you apply Sika on the cockpit side of the windshield as well?

Was the bump along the sides difficult to cover with the fairing?

Was the fairing glassed directly to the windshield as it appears, or did you make a fairing and later attach? If glassed directly, doesn?t this defeat the floating that we are attempting to achieve, or are you saying that the curve of the windshield will compensate? Not sure that I followed that.
 
I talked with the Sika lab and Sika provides NO adhesion if a space of 3 to 5 mm (call it 1/8") is not provided.

I spoke with two different Sika techs. First, there are two gaps to deal with. First is the bond gap and the other is the side gap. First tech wanted 1//16 to 1/8 bond gap and 3/16 edge gap. Second tech said 1/32 - 1/16 bond gap is fine and 1/8 to 3/16 edge gap. The recommendation was for the smaller windows of the RV-10. As Dan mentioned, I suppose larger area may require thicker adhesive to handle it.

I used a 3/4" wide by 1/16" deep bond gap with just shy of 3/16" edge gap and I can assure you there IS adhesion. I haven't gone so far as to try and kick my windows out, but they are quite secure.

Just another data point here.

Larry
 
Last edited:
As I see it now from all of the comments above, the Sika is going to work to seal the entire windshield and the sides will only require a small gap for adhesion that will probably be faired out enough to not be too significant aesthetically, but the issue remains:

How to produce the fairing? As Don has stated, epoxy laminate bonding the fairing to the windshield and fuselage defeats the use of the elastic Sika and creates a struggle between aluminum and plexiglass that could be catastrophic. Of course, the pre-produced fairing could be completely bonded with the Sika, but again, the additional 1/8? gap will be rather unsightly I would think, AND the fairing height along the roll bar would be an 1/8? above flush with the canopy.

From photos that I have seen, I am guessing that builders have ignored the issue and gone forward with a direct epoxy laminate for the fairing. They seem to report great success, but I am still unsure of what is really best.

One possible solution that I see is to pre-produce the fairing, then Sika bond it to the entire windshield and finally fill the large edge gap that is left with additional Sika and sand it smooth. I have heard that it can be sanded and painted, though I have not tried it, and if it isn?t....well, then we have another horrible issue.

Those of you who have used Sika on your windshield section, please weigh in for me. How did you handle the fairing production? Surely I am not the only builder who has confronted this. IT IS DRIVING ME NUTS!
 
Larry, what a you referring to as “edge gap”?

A standard window is installed in a setback channel or frame. The bond gap is the clearance between the flat window section and the parallel piece in the frame. The edge gap is the clearance between the windows edge (perpendicular to the flat window area) and the corresponding side or edge of the frame's channel. The window expands along it's flat axis and therefore a reasonable edge gap is needed to allow enough sealant to be able to compress and bulge out as the window expands.

I suspect there are no edge gap concerns for canopies, due to the design. However, I thought people should understand that there are two distinct recommendations on final adhesive depth.

Larry
 
Last edited:
Very interesting Lars.

Did you apply Sika on the cockpit side of the windshield as well?

Was the bump along the sides difficult to cover with the fairing?

Was the fairing glassed directly to the windshield as it appears, or did you make a fairing and later attach? If glassed directly, doesn?t this defeat the floating that we are attempting to achieve, or are you saying that the curve of the windshield will compensate? Not sure that I followed that.

I did apply Sika on the inside. Helps to have long skinny arms and fingers (I have neither). I masked the deck before plopping the windshield down, used low-adhesion 3M masking tape. It was still fun to remove after the Sika cured.

Here are a couple of photos I took yesterday showing the bump I think you are describing:

i-qhxLbqL-XL.jpg


i-nmWxWnN-XL.jpg


The above areas are 7 years old now, pardon the grime and scratches.

I bonded the glass directly to the fuselage and windshield. I used Sika primer on all surfaces that came in contact with epoxy (West Systems). I used carbon fiber over the rollbar with a final layer of glass for sanding purposes. Curved and radiused surfaces handle stresses better than abrupt transitions, but beyond that there was nothing scientific about my method other than doing tests with glass/epoxy on acrylic coupons with various substrate treatments including testing Sika primer.
 
As I see it now from all of the comments above, the Sika is going to work to seal the entire windshield and the sides will only require a small gap for adhesion that will probably be faired out enough to not be too significant aesthetically, but the issue remains:

How to produce the fairing? As Don has stated, epoxy laminate bonding the fairing to the windshield and fuselage defeats the use of the elastic Sika and creates a struggle between aluminum and plexiglass that could be catastrophic. Of course, the pre-produced fairing could be completely bonded with the Sika, but again, the additional 1/8? gap will be rather unsightly I would think, AND the fairing height along the roll bar would be an 1/8? above flush with the canopy.

From photos that I have seen, I am guessing that builders have ignored the issue and gone forward with a direct epoxy laminate for the fairing. They seem to report great success, but I am still unsure of what is really best.

One possible solution that I see is to pre-produce the fairing, then Sika bond it to the entire windshield and finally fill the large edge gap that is left with additional Sika and sand it smooth. I have heard that it can be sanded and painted, though I have not tried it, and if it isn?t....well, then we have another horrible issue.

Those of you who have used Sika on your windshield section, please weigh in for me. How did you handle the fairing production? Surely I am not the only builder who has confronted this. IT IS DRIVING ME NUTS!

My favorite time of day to confront conundrums like this is about 2AM. Ha... For what it's worth I also laid up the canopy skirt rather than using aluminum. I used packing tape as mold release (3M brand works best by far, in my experience, mostly because the adhesive releases instead of being left behind) then laid up the whole thing. Peeled it off after and finished it off the airframe. Used blind rivets into the side of the canopy frame only as per plans, and Sika everywhere to bond it. Belt and suspenders. No holes in acrylic. I've done a couple of those now. Only real advantage to a lot of extra work is that they conform almost perfectly to the fuselage.
 
i-nmWxWnN-XL.jpg


I bonded the glass directly to the fuselage and windshield. I used Sika primer on all surfaces that came in contact with epoxy (West Systems). I used carbon fiber over the rollbar with a final layer of glass for sanding purposes. Curved and radiused surfaces handle stresses better than abrupt transitions, but beyond that there was nothing scientific about my method other than doing tests with glass/epoxy on acrylic coupons with various substrate treatments including testing Sika primer.

Hi Lars. Looking at the thickness in the photo, I'd guess the boot cowl section of the fairing is strong enough to handle the compression stress by itself. Sika primer use to boost epoxy/plexiglass adhesion, per your experiments, is a stroke of genius I will not forget. The combination probably makes the low modulus Sika largely redundant.

Carbon over the roll bar section is interesting. As I'm sure you know, the graphite fabrics have a negative coefficient of thermal expansion. I seem to recall a CTE near zero when incorporated with a matrix material (i.e epoxy). If negative, it would tend to shrink more tightly with temperature rise, like a belt over the roll bar.

Just goes to show there is more than one way to skin a cat ;)
 
Last edited:
Back
Top