What's new
Van's Air Force

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

In flight window loss

kkmarshall

Well Known Member
A bit of excitement yesterday as my wife & I were returning from visiting family in Houston. We were cruising at 6500,75%,170kts,when my wife said "my window looks funny" no sooner than she finished speaking her door window blew out completely! Throttled back to about 110 kts and started assessing the situation. Very controllable with little yaw,I was on flight following so told them what occurred. ATC advised nearest airport but I told them I was considering continuing to my home base. Amazingly there was very little wind coming in,virtually none in front and not blowing on my hand very hard in back. Couldn't see any damage to horizontal stabilizer and as I said,handling normally. We continued the remaining way home (about an hour) and landed uneventfully.
I am not the builder but he had built a 7a before and I believe was the first customer completed 14a to fly (or nearly so) I have been trying to contact him with no luck to verify if he used Weld-on or Lord or whatever. I have the usual cracks in the paint around the windows that you see on a lot of the earlier builds
('07 first flight on mine) but it was not raised from the original flush fit to the cabin top. It came cleanly out of the joggle on all but the top and about half the aft towards the top. But I was able to cleanly pull some more out of the joggle after we landed. Makes me think I should replace all the side windows! The residue of the adhesive is sort of tan in color if that is a clue to what type it was.
I had always heard the cracking was cosmetic only so I had never been concerned but this changes everything! This window's cracking was at the forward lower edge going up a 3rd of the way and probably back a 3rd of the way. Obviously it pulled off in that area first (what my wife saw for a split second) and then when the slipstream got under it a bit it was game over.
Bummed about the thought of replacing all side windows but very thankful nothing was struck on the tail and that it handled so well through it all.
Crazy prop blast thumping while taxiing though. Just a heads up for anyone with cracks of more than an inch or two.
 
Last edited:
In the past I?ve lost a door on a Pawnee and smashed an entire glider canopy during aerobatics. The one thing I took away from both of those occasions was how pleasant it was to enjoy the view and fresh breeze without Perspex getting in the way! In the long run I think it would work out cheaper to buy something open cockpit if I?m to enjoy that privilege regularly!
 
Wow great job flying the aircraft first. Kudos to your wife for being a good pax too! I’d like to see pictures too.
 
Looks like Lord adhesive and a bond failure between the adhesive and the fiberglass. The question is what caused the bond failure? Poor preparation, unsuitable adhesive, cyclic fatigue due to thermal expansion, bigfoot...?











 
That's what I want to figure out as well. I have small paint cracks on the corners of the other windows also though none are as long as were on the window that failed. Curious how you are able to identify it as Lord adhesive? Van's tech support wasn't sure from the pics. He said Weld-on is white and this looks cream colored. Is that the color of Lord? Data point,first flight was Jan '07 and it has 900 hrs with 400 of them by me over the last 5 years. Always in a hangar.
Thanks for posting the pics for me by the way.
 
Last edited:
That's what I want to figure out as well. I have small paint cracks on the corners of the other windows also though none are as long as were on the window that failed. Curious how you are able to identify it as Lord adhesive? Van's tech support wasn't sure from the pics. He said Weld-on is white and this looks cream colored. Is that the color of Lord?
Thanks for posting the pics for me by the way.

I could be wrong, but this looks like the Lord Adhesive I used... It is slightly off-white.
 
I could be wrong, but this looks like the Lord Adhesive I used... It is slightly off-white.

Thanks for posting that info.

I am not familiar with the Lord product so we couldn't make a call for certain based on the photo (colors don't always show in photos what they really are).
I would have been surprised if this failure was with Weldon considering the nearly clean de-lamination from the composite. When the structural testing was done, failure of the windows always occurred through the thickness of the adhesive, and generally left a bonded layer of adhesive on the composite that had to be ground off to bond in windows for another test.
If this had been Weldon my guess would have been a possible surface prep. error. I guess that could still a factor if it was Lord adhesive.
 
Thanks for posting that info.

I am not familiar with the Lord product so we couldn't make a call for certain based on the photo (colors don't always show in photos what they really are).
I would have been surprised if this failure was with Weldon considering the nearly clean de-lamination from the composite. When the structural testing was done, failure of the windows always occurred through the thickness of the adhesive, and generally left a bonded layer of adhesive on the composite that had to be ground off to bond in windows for another test.
If this had been Weldon my guess would have been a possible surface prep. error. I guess that could still a factor if it was Lord adhesive.

It would still be interesting to get a piece of Keith's window and compare the adhesive on it with a samples of weld-on and lord. Matching colors over a computer screen probably isn't good enough to be absolutely certain.

My fear (?) with all of these adhesives is that if we're seeing cracking at the window edges, maybe the heat/shrink cycles are enough to shear the adhesive over time, particularly if the joint wasn't perfect from the beginning.
 
Poor bond prep

A good bond failure is a cohesive failure where the bonding agent fails leaving material on both components. This would also be the case with a poorly mixed adhesive resulting in poor strength. This applies to either a single overload event or a cyclical load.

A bad bond failure is an adhesive failure like what is shown in the photo where the adhesive has separated fairly cleanly from the substrate. This generally is caused by poor surface prep. In this case it appears that the separation is almost 100% of the area including imprinting from pinholes. There also appears to be no scuffing of the surface to prep for bonding.

A simple way to test for surface contaminates is to mist some water onto the surface. If it forms a continuous film then the surface is likely clean. If it beads up, the surface is contaminated.

Judging by the photos, poor bond prep is a likely cause.
 
Of course it doesn't show in the photos,but the composite of the door frame does have sanding marks on it that appear to have been made by approximately the correct grit sandpaper called out in the plans. Contamination of the surface by something before bonding certainly could be the cause though. I will check with Van's and see if they would like a shard sent in to inspect. I would like to use Lord based on comments posted here in the past on how everyone hates working with Weld-on. But not if Lord gives an inferior bond!
 
Thank you for sharing this information

Really great information. It might help if a layer of fiberglass was laid over the top of the window edge to capture the plexiglass. This might be a way to prevent this. I also have window edge cracks. I will be inspecting very closely today!
 
The bonding failure, as others said, could be from poor prepping to squeezing out to much material when it was installed, leaving only a skin coat to adhere the window. Going forward, I would recommend the Cee Bailey’s windows. Their windows will require only minimal trimming. Also, google “Glasair Sportsman 2x2 window installation. There is a 5 part series how to do it. My windows were installed that way. There is no exterior fiberglass work when installing windows with this method.
 
Very interested to see what comes from this. Not sure if anyone else has reported a window loss event like this or not, but with lots of folks discussing alternatives to the Weldon product its good to know that this seems to be a very rare occurrence.


Also, am I the only one that is going to comment on the drapes installed in your -10. Not only that they are installed, but they look to have survived the window loss. Awesome! Looks like a very comfortable interior.
 
I bonded my plexi on my showplanes fastback RV8 with a product called PRO-SET. I called the Gougeon Brothers in Bay City, MI (West System) and asked them what I could use for this application. This application is very similar to what is done on the RV10 windows with plexiglass laid up on a fiberglass substrate. Pro-set is what they recommended. It?s a thickened epoxy structural adhesive. Spaceship I was put together with this adhesive and it made to outer space. It is compatible with all West System epoxies and will even wet out cloth. I don?t remember whether I used 175/275, or 176/276, but you can click on the epoxy type in the link and it will show you the properties. Working time was over an hour, so I didn?t have to hurry. I roughed up the inside and outside of the plexi and the fiberglass joggle where they overlap and cleaned it well with isopropyl alcohol before applying the adhesive. The adhesive had the consistency of creamy Jiff peanut butter. I drilled a 1/4? hole in the plexiglass every 4? on the overlap area and countersunk the hole on the outside (holes were deburred/sanded to make them very smooth). This made the adhesive that oozed out through the hole act like an epoxy rivet with the flush ?head? on the outside. You could mix some flox in the mixture in those spots to give it more strength if you wanted to, but I don?t think that is necessary. I then laid 2 or 3 layers of cloth on the outside overlapping onto the fiberglass canopy structure before the PRO-SET cured. This layer wetted into the epoxy rivets to form a very strong bond. I had 1/8? holes drilled every so often to hold the assembly in place with clecoes until it cured. These were easily filled and sanded later. The plexiglass is held firmly in place between fiberglass on the outside, and the fiberglass frame on the inside.

https://www.prosetepoxy.com/standard-products/assembly-adhesives/
 
The bonding failure, as others said, could be from poor prepping to squeezing out to much material when it was installed, leaving only a skin coat to adhere the window. Going forward, I would recommend the Cee Bailey?s windows. Their windows will require only minimal trimming. Also, google ?Glasair Sportsman 2x2 window installation. There is a 5 part series how to do it. My windows were installed that way. There is no exterior fiberglass work when installing windows with this method.

A) Why would the Bailey windows make a difference?

B) With ~5,000 installed and flying RV-10 windows and one failed window, does the data support changing the process?
 
The aircraft appears to have received its AW cert back in 2007. It?s certainly possible that Lord adhesive was used but it?s relative widespread use in RV-10s didn?t start until several years later when it was recommended as an alternative by a well known supplier of aftermarket products. With that in mind, it MIGHT be safe to say that the adhesive isn?t Weld-On because of the color (Weld-On is a bright white color), but beyond that it?s a LOT of speculation. I?ve personally done test samples and with proper surface prep I know for a fact that the Lord adhesive will adhere well with the plexi breaking before the bond fails.
 
A) Why would the Bailey windows make a difference?

B) With ~5,000 installed and flying RV-10 windows and one failed window, does the data support changing the process?

A: Less trimming, but certainly the same adhesion issue.

B: This is a very good point. Statistically speaking, 1 failure is an anomaly. Speaking as nuclear engineer - it's an issue that needs further review. I recommend obtaining a several samples of remaining adhesive, possibly one of the other windows if you're going to redo those too, and contacting the main vendors of the potential products, Lord and/or Weldon, and see if they are willing to identify the product. If it's theirs, I'm guessing they would like to understand the failure as well and you might be able to share for the group anything you learn.

P.S. Glad you landed safely with a good outcome; good job man.
 
I didn't care for the window installation method Van's recommends in the instructions when I built my RV-10. I have read way too many stories of the transparencies crazing from the Weldon and opted for the Rutan composites method that I used building the Cozy MKIV. The Rutan planes have used this method since the 1970's and I still follow their groups. Never heard of a window falling out using this method or crazing from a chemical reaction using epoxy vs Weldon. My building buddy has over 700 hours and 10 years on the Cozy MKIV we built and not a sign of any transparency issue. They fly at speeds and altitudes comparable to the RV flyers. I am not familiar with the Lord adhesive method but wouldn't get near Weldon with a 10' pole due to crazing. I replaced all the transparencies on my C-152 years ago and used Weldon to install the compass mount following the mfg instructions. Within a couple days the area around and under the compass mount had crazing. :mad:
This is how I installed my transparencies on my RV-10:
Prep the mating surfaces of the transparencies 3/4" wide on both sides and edge using 220 grit. Use a light to inspect after you blow off the dust. There can be no shiny areas. I use several layers of 3-M electrical tape to tape off the joint line then cover that with a layer of Gorilla Tape. The Gorilla Tape will stand up to any abuse and even light sanding while the electrical tape under it will allow EZ removal when your done leaving a nice clean edge that will require minimal work to finish. I left my tape on right up through paint.
Prep the fiberglass mating surfaces with 40-60 grit. There MUST not be any shiny areas in the prepped surfaces and they must be dust and contaminate free. Paint the mating surfaces with a very thin coat of pure epoxy. Lay a bead of epoxy/flox mixed to a peanut butter consistency and install the transparency. Make sure there are no voids of flox under the transparency. I used the per plans method to clamp and hold the transparencies in place. Clean up the squeeze out and let cure.
Make a two layer 2" wide BID tape wet out on clear 4-6 mill plastic long enough to go around the joint. Lay the plastic down on a flat table and mark it using a permanent medium point Sharpie. I use a 2" wide 6' straight edge and just trace around it. Depending on how many tapes you need you can just stack them up. Cut your BID a couple inches wider and longer that what you need. Roll the cut bid up carefully to maintain size and shape and transfer it to your marked plastic and carefully unroll it over the marked plastic. When its properly positioned wet it out. Don't over wet it, just make sure its wet out. Now do it again with the second layer. You will be able to see the cut line you drew right through the wet out fiberglass. Cut it out to the lines with a good pair of scissors. Again make sure the area is properly prepped and paint the surface with a thin coat of pure epoxy. Fill any low areas between the joints with epoxy/flox mix level. Lay the 2" wide cut fiberglass tape butt up against your taped edge over the transparency and fiberglass peeling the clear off as you lay the tape down over the joint. The plastic will hold the tape to size and you will be able to make curves as you peel and apply the tape. Stipple it down with a 1" wide paint brush. Use a light to make sure there is no trapped air. Now lay peel ply over the joint and stipple that down. When cured, strip off the peel ply and you have a very secure window installation that will require minimal finishing to get ready for paint.
 
Hey Kyle, the Bailey’s glass almost fits perfect out of the box. I just had to sand down an edge a little for the final fit. Not saying the method I used looks better or anything else. It does not require doing the fiberglassing over the plexiglass. Once the windows is glued in, the joggle gets filled in with the same adhesive/sealant, and your done. Also, if a window does need to be removed, it’s simpler to remove and does not require cutting out the glass work and redoing it. Either installation methods look nice.
 
Also, am I the only one that is going to comment on the drapes installed in your -10. Not only that they are installed, but they look to have survived the window loss. Awesome! Looks like a very comfortable interior.
Well, since you brought it up... I was going to comment that the window may have jumped out on its own accord just to get away from the curtains... :)
 
Builder confirmed that Weld-on was used. Sent some shards to Van's for their inspection. Leaning towards using Lord per the experience of many here and glassing over the seam on the new window as well as the remaining windows.
I like my curtains,thank you very much!😄
 
A good bond failure is a cohesive failure where the bonding agent fails leaving material on both components. This would also be the case with a poorly mixed adhesive resulting in poor strength. This applies to either a single overload event or a cyclical load.

A bad bond failure is an adhesive failure like what is shown in the photo where the adhesive has separated fairly cleanly from the substrate. This generally is caused by poor surface prep. In this case it appears that the separation is almost 100% of the area including imprinting from pinholes. There also appears to be no scuffing of the surface to prep for bonding.

A simple way to test for surface contaminates is to mist some water onto the surface. If it forms a continuous film then the surface is likely clean. If it beads up, the surface is contaminated.

Judging by the photos, poor bond prep is a likely cause.

I second the notion that poor bond prep as the root cause. There was virtually no adhesive left on the fiberglass of the cabin top, which tells me that there was minimal sanding to scuff the area to give the adhesive more "tooth" to bond to.

On another note, I wish Van's would come out with an RV-10 B version of cabin top with the joggle for the windows on the INSIDE. This would eliminate the cracking of paint that happens on the current version. Also while he's at it, redesign the cabin doors to pivot at the front, like a Cirrus, so that failure of the latch (or the pilot to latch it) would not result in the loss of the door. This would be a huge safety improvement.
 
On another note, I wish Van's would come out with an RV-10 B version of cabin top with the joggle for the windows on the INSIDE. This would eliminate the cracking of paint that happens on the current version. Also while he's at it, redesign the cabin doors to pivot at the front, like a Cirrus, so that failure of the latch (or the pilot to latch it) would not result in the loss of the door. This would be a huge safety improvement.

Sounds like an aftermarket opportunity for someone. Seriously.
 
cabin top

I like the idea of the inner joggle...the door pivot, not so much.

My biggest beef with the cabin top is the fit and finish...if they only did ONE thing, the cabin top door opening should be 3/8" narrower...then the fitting could be done per the plans instead of grinding off ALL of the flange and 25% of the door channel corner radius...
 
I know it's impossible to see in photos and hard for me to tell what grit sandpaper was used,but there are sanding "scratches" in the fiberglass. I agree a good bond would not pull away cleanly like that,only the forward upper corner is well bonded and I tried chiseling it out with a gasket scraper. It's gonna take some work on that part. Contamination of some kind is my guess,whether wet or oily hands or whatever. Since it is confirmed that Weld-on was used,I am leaning towards using Lord from all the comments on the difficulty working with Weld-on. It is stuck really well to the plexi on the shards I have.
 
If you don't mind a Lancair guy chiming in, that fiberglass (Looks like Vinyl Ester resin) looks shiny. That's not good. All I know is that when I bonded my acrylic canopy to the fiberglass frame, I was religious about prepping the bond area. First I cleaned the fiberglass with Prep-Sol, then sanded until dull with 80 grit, then cleaned with prep-sol again. I used Hysol structural adhesive to make the bond. Then I applied two bid over the joint on the outside.
 
This isn?t exactly a one off. I know of 3 that have come out and one was actually on the ground. The belief is the hot sun reflecting off the unpainted wing popped the window out of the joggle as the window expanded.

So this isn?t a one-off as was previously mentioned. There is some history there and certainly more than the 3 I?m aware of.

That being said, there are several great adhesive choices out there that are more than adequate to hold a window in. But if the prep-work is sloppy they are all equally inadequate too. Regardless of which adhesive you choose, the quality of the bond is only as good as the prep-work performed on the mating surfaces and adequate mixing of the product.
 
Interesting,Van's was not aware of any having come out in flight. I agree with you though, prep work is the key factor.
 
When flying our -10, always nervous about the doors

being "sucked" open as there was a definite venturi effect. I am sure the constant cycling of the pressure difference would eventually uncover weak bonding on the windows. (or poor door latching!)

Ron
 
This might shed some light on adhesive. The aircraft was finished in January 2007. It shows airworthy certificate January 6 2007.
Lord adhesive for the window install was not being used till 2010 at the earliest.
JUST FYI
It looks like there was no sanding or prep done on the fiberglass.
 
There are sanding scratches on the fiberglass but they are light,hard to tell what grit. I agree that prep seems the problem. I ordered some Lord from you,by the way.
 
This might shed some light on adhesive. The aircraft was finished in January 2007. It shows airworthy certificate January 6 2007.
Lord adhesive for the window install was not being used till 2010 at the earliest.
JUST FYI
It looks like there was no sanding or prep done on the fiberglass.

I think the original poster has said more than once that there are sanding scratches on the glass but they don't show up in the low res photos. I agree that it looks shiny, which is bad, but he says that is not the case. I agree that an internal joggle would be a better design as it would natural resist the pressure loads. That is an area of accelerated air so low pressure. Glad nothing valuable got sucked out!
 
As for something getting sucked out,I wouldn't have believed it either,but I had a ball cap sitting on the back seat & it just sat there. Less wind in the cabin than in my pickup with the window down at 70 mph.
 
Last edited:
I think the original poster has said more than once that there are sanding scratches on the glass but they don't show up in the low res photos. I agree that it looks shiny, which is bad, but he says that is not the case. I agree that an internal joggle would be a better design as it would natural resist the pressure loads. That is an area of accelerated air so low pressure. Glad nothing valuable got sucked out!

Visible sanding scratches, and 100% sanded, are very different things.

I have seen many instances were someone thought their surface prep was ready for fiberglass and it wasn't. Even after they sanded some more, it still wasn't.

You are not sanding to make (just) scratches. You are sanding to abrade 100 % of the surface area. If you can look between the scratches and see surface area that isn't abraded, it is not sanded enough.
Best practice is to sand in multiple directions (length wise, cross wise, diagonal, etc.).
 
I don't understand why Van's does not discuss adding a couple of bids over all the glass to cabin top joints. This would eliminate the classic "window framing" paint crack from people filling the joint with just micro before paint. It also provides a good second window retention element.

And - I know of no downside doing this.

What I did:
- Glue in the glass per plans, let set. Of interest I used the window adhesive prescribed for Lancairs - Hysol 9394
- Fill in the joint with micro and sand so the window is flush with the cabin top.
- Two bids of medium glass over the now sanded joint. The glass does not extend beyond the glue joint so you do not loose any window area. Peal Ply over the bids. Let set.
- Skim coat of micro over the glass and final sand to prep for paint.

The only thing I would change for the next RV-10 is shifting from fiberglass to carbon - I find it easier to work with.

Carl
 
Last edited:
Could someone identify the specific Weld-On and Lord adhesives discussed here? I'd like to look up the TDS for each.
 
As for something getting sucked out,I wouldn't have believed it either,but I had a ball cap sitting on the back seat & it just sat there. Less wind in the cabin than in my pickup with the window down at 70 mph.

In that case just close the curtains and press on :D
 
The procedure Carl Froehlich described is what I did, only I extended the 2 bid maybe 1/2" onto the plexiglass. As others have noted, due to Bernouilli's Theorem, the window wants to pop out at speed.
 
Back
Top