What's new
Van's Air Force

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

Pondering if thermal expansion can break a canopy

chepburn

Well Known Member
There has been some discussion lately on cracked canopies and some speculation on the cold weather being the culprit.

I thought I would start a new thread based on Danny Kings comment:

http://www.vansairforce.com/community/showpost.php?p=1230209&postcount=13

Danny has postulated that the difference in thermal expansion between the steel frame and the canopy is the smoking gun. I think he has hit the rivet on the head here. Danny oversized his attachment holes.. and I think he is on the right track. I have attempted to put some numbers to the problem....

Assuming a canopy built to plans we have the following numbers:

A #40 drill through the steel rail. (0.098")
A #27 drill through the canopy. (0.144")

12 " between fasteners. (Lo in the equations)

EDIT: Thanks to Alex' comments below...things have changed A LOT.

The thermal expansion Ct for steel is approximately 0.000008 in/inR
The thermal expansion Ct for acrylic is approximately 0.0000417 in/inR

I'm going to take a lot of liberties with my assumptions from here on (feel free to shoot me down)

Lets say we go flying at -5 degF outside. Lets also say that our cockpit/sun heater is able to bring the interior of the cockpit up to 32 degF for a deltaT of 37 degF as we fly around.

The expansion of a material dl = Ct * deltaT * Lo

So...

The steel will expand by: 0.003552 inches between fasteners.

The acrylic will expand by: 0.01851 inches.

Yikes. But wait, we have some slop in the holes of 0.046 inches (#27 - #40)


So the difference in length the steel and the canopy is going to be 0.01852 - 0.003552 - 0.046 = -0.03104 in.

So, there is enough oversize in the hole to have no strain on the plexi.

If we assume we didnt drill the plexi hole in perfect alignment with the frame, (in other words, the rivet is right against one edge) there will be a maximum strain of

Well strain is deltaL/Lo = .046/12 = .00383

Stress = Youngs Modulus * Strain

Youngs Modulus for Acrylic is about 400000 PSI

So the stress rise in the acrylic is .015 * 400000 = 500 PSI

The tensile strength of acrylic is about 9000 PSI (this has a big range so I just picked one)....so we have some margin.

Now... if we increase our deltaT to 50 (meaning we managed to warm up to 45 degrees in the cockpit) the stress goes to 675 PSI.

But wait! We drilled a hole in this thing.... what did that do? Well, holes create a stress concentration.... and there's a lot of math we could go through to try to figure that out, but we will just take a shortcut, and set our hole stress rise factor to 3. (look it up)

So, NOW with a deltaT of 37 deg we have:

500*3 = 1500 PSI Still OK

Now, I have made a LOT of assumptions and taken a lot of liberty with geometry etc...

There have been a lot of comments wrt installation loading due to bending the frame to fit. But from this calc, maybe the cold differential is enough along with the preload to crack the canopy....

That being said, it looks like if you align those hole VERRY carefully during installation, you wont be introducing load due to thermal expansion.

The take away is Danny's approach looks like a good idea.

For all of us flying around with plans built canopies.... well..... I don't know the answer to that.

Except..... my canopy is still fine... and so are thousands of others..... So I have to conclude that there is some basic error in my calculations.. but, that being said, Danny's idea is still a good one.
 
Last edited:
I remember one of the first Earth Rounders writing about how firghtened he was somewhere around Greenland coming down the GS in his RV-4 in the extreme cold and watching a canopy crack begin to propagate and him wondering if it was all going to hold together.

I have had my canopy crack when moving it outside from a warm hangar into frigid temps.

vic
 
Every time my overhead radiant heaters kick on and off (lots last week in the -30 to -38C temps), I can hear the canopy "pop" multiple times as it expands and contracts on the rivets and frame. My 6A is parked nearly right underneath the heaters.

Same when I roll it outside like today to fly in the winter. Built to plans, flown down to -20C sometimes, nary a crack yet after 15 years and hundreds of flights and thousands of heat/cool cycles under the heaters.
 
I think it also has to do with the temp at which the canopy and steel frame were also drilled. If you drilled and set the canopy in a 65 degree hangar, and it gets down cold, it throws those numbers further out of whack. I believe you have to start your calculations at the point of theoretical zero stress on the canopy (The temp at which it was drilled and attached).

Some people have gone as far as heating the larger panels prior to drilling/Cleco/riveting to reduce the chance of oil canning. After the steel cools it is drum tight with little chance of oil canning. We might have to explore just the opposite for the RV8 canopies.
 
I think it also has to do with the temp at which the canopy and steel frame were also drilled. If you drilled and set the canopy in a 65 degree hangar, and it gets down cold, it throws those numbers further out of whack. I believe you have to start your calculations at the point of theoretical zero stress on the canopy (The temp at which it was drilled and attached).

Some people have gone as far as heating the larger panels prior to drilling/Cleco/riveting to reduce the chance of oil canning. After the steel cools it is drum tight with little chance of oil canning. We might have to explore just the opposite for the RV8 canopies.

IMO, the issue is the overall length of the -8's bubble. It substantially longer and maybe more restricted from movement than the other bubbles. So, when it grows at 5x the rate of the steel, the overall growth (or difference in growth between the plexi and frame) is larger than that of the other aircraft.
 
+1 For stresses due to the rivets coming up too tight on the holes, you need to compare dimensions at the initial fabrication temperature to those at the current temperature. For canopies cracking on sudden removal from a heated hangar to the cold, you may be looking at the internal shear stress - the outside of the plexi is suddenly much colder than the inside, the outside shrinks and the plexi fails internally in shear, starting at some stress riser.
 
I do believe this is on the right trail. I have a crack in my canopy that goes from one rivet head to another where it is riveted to the front slider bow, does not extend elsewhere, just between those two rivets. I do a lot of high altitude flight and have seen temps in the negative teens on a few occasions.
 
I think that another consideration is an"initiating event" while the canopy is under stress. My 8's canopy cracked on a cold winter's day in Big Bear. Overnight temps were in the tens, and I rolled out to go down the hill to Cable airport. The northerly winds that brought the cold air were howling through Cajon pass, and I was descending quickly because of the altitude I had to lose, when I was hit by thee or four bumps that would have knocked me into the canopy if my five-point wasn't really tight.

When I landed, there was a neat crack back behind the passenger seat. I don't know if it would have cracked without the bumps, but my thought is that it was under great thermal expansion stress, and all it needed was a bit of a nudge to go bang.
 
I think that another consideration is an"initiating event" while the canopy is under stress. My 8's canopy cracked on a cold winter's day in Big Bear. <snip>

That, and it was cold so the plexi was brittle. You rarely hear of a cracked canopy in hot weather when the plexi is <relatively> soft and pliable.
 
On the opposite side of things, I don't think having it out in a very hot day with the sun shining will cause cracks. My thinking is that with the high temperature the canopy is pliable and will deform before it cracks, unlike when it is cold out.
 
Canopy Cracking

My personal opinion is there's a fundamental problem with the RV-8 canopy mold that results in a sprung-out piece of plastic resulting in too much preload in the attachment to the frame. I believe this is the primary contributor (or certainly significant) to RV-8 canopies cracking in the back. The design of the canopy weldmet tooling should be very closely matched to the canopy mold and their resulting shapes. Not and easy problem to solve and expensive to get right.

After spending literally 1 month of tweaking my frame and canopy position, the best I could do was a 1" sprung-out condition in the back third area as others have reported. This preload really bothered me and meant more work on the canopy skirt. I put an extra thick Sikaflex bond line in that area to help minimize its effect. It will be a very sad day if I ever hear it pop back there and my heart goes out to those who have. Don't remember if I measured the force required to spring the canopy into position on the frame but this should be added to the stress analysis above.

Having designed composite and plastic parts for modern fighters for over 20 years now there are two approaches we take to getting the final shape correct - analytical and trial and error. Analytical doesn't work well regardless of the number of white papers on the topic. There's a computer program that claims it can take your part shape, cure cycle, part material and tooling materials and tell you the shape of the tool required to build a perfect part. Hopefully this program will work well some day. We plan on 3 sets of tooling for a molded part which is a very expensive approach. Two sets won't do it as the problem is non-linear. If the part is sprung-out 1" and you compensate the tool 1" in, your part will still be sprung out .5" or so. By the third attempt the part is pretty close and you've ran out of money to make it any better.

Having spent a fair amount of time on redesigning plastic parts that had cracking problems in the field due to loaded fasteners/holes I've concluded and preach this to young engineers, don't design holes in plastic parts - ever! The solution is always to remove the holes and design a frame with proper gaps to trap the plastic part. We've played extensively with hole tolerance, "super isolating grommets/inserts", silicons and the holes still crack out even when the stress analysis says they shouldn't. Some suppliers are starting to claim they have invented a "super plastic" that won't crack at the fastener holes under load, we'll see.

I believe Vans told me along time ago the -8 canopy mold/tool was built off the prototype canopy. I would be curious to know if they attempted to make any spring-back or shrinkage adjustments to the tool. I would enjoy helping Vans adjust their canopy frame tooling and canopy tooling to eliminate preload. This and making Sikaflex the standard attachment method could significantly eliminate cracks.

Dave
RV-8A
921E
 
Canopy

This discussion makes me glad I used Sikafllex not rivets or screws to install my windscreen. I wish I had also used it on the canopy.
Has anyone done cold weather tests on canopies glued on va screws?
After 700 hrs I have had no hints of cracks or stress on it.
 
+1 for ambrose.... comments.

Working with frame and canopy components from Vans is a disadvantage. I understand most working guys with families are not excited about turning the canopy into a 3 month project. But, if you reference the points made above , below is the path I have taken twice:
1-Form carbon fiber inner frame that Replaces the
metal. This should include a shelf for the canopy to sit on and 1/8? clearance to the canopy for Sikaflex.
2-form up your canopy in wooden/angle iron fixture and have Jeff ( airplane plastic) stress relieve it. That removes initial stresses just to fit in the frame.
3- laminate divinicell foam to the CF for a spacer to the outer canopy skirt skin.
4- form outer skin with glass bonded to foam, not plexi, so they expand differentially.
Finish the cosmetic work and put the canopy in the slot with Sika on the CF side only. I believe it is well known that long canopy tails leak air badly due to expansion; even my little -3 does, so a fast back has advantages. Cold shocking is apparently a hazard that only you can deal with.
 
Too wide

The canopy’s are too wide in the rear part. We have to preload them.
I think if the things fit better in the initial fitting, we would have fewer cracks
My $.20. Includes. .18 for inflatition -;
 
Last edited:
crack even with sika - wow

... It is clear, however, that Sikaflex is not the anti-crack panacea I bought into over a decade ago. ...
Hi Scott, really sorry to hear about that crack. Not that I can do anything about it on my canopy which is already complete, but it could be interesting for new builders to find out where it started, how thick your sikaflex layer was, temperatures, etc.
 
I agree. Without a doubt, I remember fitting the canopy over 11 years ago and anguishing over the pre-loads required to fit the canopy to the frame. Even the Todd's canopy I installed would not fit the frame without those 1" pre-loads you reference here. And after over 1400 airframe hours, plenty of g-loads, cold Wx, turbulence, and blasting sunlight, she gave way on a tarmac on December 1st while I slept in a pilots lounge. Not even a Sikaflex installation could save her.



On a side note, I have not yet heard anybody mention shrinkage. Unless the material is MIL-P-8184 or some other pre-shrunk acrylic, I think we can always expect the canopies to shrink over time. Whether the installation can compensate for that shrinkage is the key question. It is clear, however, that Sikaflex is not the anti-crack panacea I bought into over a decade ago.

The repair post is forthcoming later this week . . .

I'm in the same boat as Scott. 1500 hrs. +. Lots of G, 10 years and it finally let go at 11,000 ft and -1 F. My process for fitting the canopy has changed since I did this one.
After initial fitting and the "big cut", I know start clamping at the rear and work my way forward. The canopy stays in place at the top middle but grows forward at the bottom forward windscreen split. I drill it this way and cut off a wedge shaped sliver from each side of about 3/4 in to nothing and maybe 18 in long. Way less stress at the rear this way and the skins fit almost perfectly.

Alas though, mine cracked 8 inches aft of the windscreen split.
 
Working with frame and canopy components from Vans is a disadvantage. I understand most working guys with families are not excited about turning the canopy into a 3 month project. But, if you reference the points made above , below is the path I have taken twice:
1-Form carbon fiber inner frame that Replaces the metal. This should include a shelf for the canopy to sit on and 1/8” clearance to the canopy for Sikaflex.

The CTE of carbon/epoxy laminate is near zero, which makes it a very poor choice for this application, certainly far worse than steel. The problem is compounded by its high modulus. It is working as above only because (1) the design floats the canopy on highly deformable foam and sikaflex, and (2) apparently does not bend the plexi to make it conform to the frame.

One could argue that the CF frame doesn't change dimension with temperature change, so it always fits the fuselage, but that would be true only if the fuselage was also a carbon/epoxy structure. It would have a big CTE delta with an aluminum fuselage.

If a composite canopy frame is desired, glass fiber/epoxy is the superior choice, as its CTE is about half that of acrylic plastic. Compare with aluminum at roughly 3x less, steel at 5-6x less, or carbon/epoxy at about 70x less.

BTW, if you really want a metal frame, plutonium would result in the least CTE delta with acrylic. ;)

Assembled glass/epoxy frame from a Showplanes fastback kit. Doing fine so far. Something similar could be fabricated for a standard turtledeck, although there is an advantage to the shorter acrylic bubble.

10qx5jr.jpg


qqzb5s.jpg
 
Last edited:
Good points Dan

Yes Dan, I am a little casual about CTE. Assuming fuse clearances fore and aft are generous enough. I may be mistaken but it feels to me like CF achieves dimensional stability better than glass for the structure. That said glass may be a better choice for the whole task.Just trying to complete the dialogue that may help someone.:)
 
My personal opinion is there's a fundamental problem with the RV-8 canopy mold that results in a sprung-out piece of plastic resulting in too much preload in the attachment to the frame.

The canopy?s are too wide in the rear part. We have to preload them.
I think if the things fit better in the initial fitting, we would have fewer cracks
My $.20. Includes. .18 for inflatition -;

There is nothing wrong with the RV-8 canopy mold, because there is no mold.
The canopies for the RV-3, 4 and 8 are free blown bubbles.

When making the canopy, the heated material is restrained within a frame that establishes what the perimeter shape will be around the base. Differential air pressure is used to push the bubble to the specified height. The overall shape is then whatever the bubble naturally settles at (all of the bubbles are snowflakes in a sense.... no two are exactly the same)
The molding frame shape is what establishes the free blown shape.
The profile of the molding frame could be changed to produce canopies that more closely fit the canopy frame but it would greatly influence the canopy shape so that head room would be reduced an undesirable amount (I.E., the molded width has a direct correlation to what the free blown height will be).

While I don't disagree that having a pre-load on the canopy as installed is a less than ideal situation, I have seen no indication that RV-8's have a higher tenancy towards cracking in service than any of the other RV models.
 
As a counterpoint I'll comment on the Glastar / Sportsman windshield installation technique, which is quite similar to the Sikaflex method of bonding the windshield with an adhesive much akin to RTV silicone.

While clearly not as demanding an installation as the RV8 (not a full, tight bubble) the Glastar and Sportsman windshields span a large area with compound curves and are exposed to significant air pressure loading in flight. Yet breakage of windshields is very rare indeed, and almost always as a result of an "initiating event" as Paul Dye suggested.

What's the difference? The windshields are bonded with what effectively is an elastomeric, are bonded to a composite frame, and are free of pre-load when bonded in place.

Ultimately I believe it's the combination of pre-load, through-hole fasteners and age/thermal cycles that are the nemesis of RV canopies.
 
I'm in the same boat as Scott. 1500 hrs. +. Lots of G, 10 years and it finally let go at 11,000 ft and -1 F.

Uh oh... sounds like my canopy will be cracking any day now; I?m at 11 years and 1475 hours, no cracks yet. Yet...

Last month I was droning across northern Arizona and New Mexico at 9500? and 18F OAT; I kept waiting for the crack to happen but it didn?t. Lucky me.
 
Plan B

The problem discussed here is very real, and steps to avoid the problem (plexi vs metal expansion rates) are nothing new.

I inserted a few steps in the F1 assy plans to help reduce the cracking problem: drill the w/s holes at ~3/8" before final assy and use a 'washer' cut from auto vacuum line as a buffer between the plexi and the fasteners. The plexi is retained by a tinnerman washer/screw, and it is not to be tightened down firmly - the idea is to allow the plexi to move around as necessary in changing temperatures - AND fuselage flexing. The fiberglass fairing is then laid up over the washer/screw setup. No fasteners are evident in a proper installation.

The slider canopy is sandwiched in the same manner as the RV-6 slider, but the holes are again oversized, and the canopy is sealed with paintable caulk - allowing some relative movement.

So far, no cracks from temp extremes in the fleet.

I don't see why such a setup could not be used on almost any aircraft that we are discussing here, but the -8 install would likely change to screws, not rivets. The screws could be hidden - overlaid with fiberglass, as is done with the w/s plexi.

YMMV, as usual.
 
I agree. Without a doubt, I remember fitting the canopy over 11 years ago and anguishing over the pre-loads required to fit the canopy to the frame. Even the Todd's canopy I installed would not fit the frame without those 1" pre-loads you reference here. And after over 1400 airframe hours, plenty of g-loads, cold Wx, turbulence, and blasting sunlight, she gave way on a tarmac on December 1st while I slept in a pilots lounge. Not even a Sikaflex installation could save her.



On a side note, I have not yet heard anybody mention shrinkage. Unless the material is MIL-P-8184 or some other pre-shrunk acrylic, I think we can always expect the canopies to shrink over time. Whether the installation can compensate for that shrinkage is the key question. It is clear, however, that Sikaflex is not the anti-crack panacea I bought into over a decade ago.

The repair post is forthcoming later this week . . .

The reflective discontinuity across the crack seems to indicate (as you note, Scott) that more than thermal stresses are at work. Looks like either built-in curvature problems, or curvature which doesn't match the canopy frame.
 
There is nothing wrong with the RV-8 canopy mold, because there is no mold.
The canopies for the RV-3, 4 and 8 are free blown bubbles.

When making the canopy, the heated material is restrained within a frame that establishes what the perimeter shape will be around the base. Differential air pressure is used to push the bubble to the specified height. The overall shape is then whatever the bubble naturally settles at (all of the bubbles are snowflakes in a sense.... no two are exactly the same)
The molding frame shape is what establishes the free blown shape.
The profile of the molding frame could be changed to produce canopies that more closely fit the canopy frame but it would greatly influence the canopy shape so that head room would be reduced an undesirable amount (I.E., the molded width has a direct correlation to what the free blown height will be).

While I don't disagree that having a pre-load on the canopy as installed is a less than ideal situation, I have seen no indication that RV-8's have a higher tenancy towards cracking in service than any of the other RV models.


Well, then there must be a lot of them cracking because mine cracked yesterday as well. That makes 5 ,that I’ve herd of, in one day. I’m beginning to think the only fix is to let it crack and live with it.
 
Last edited:
A lot of insightful comments so far..now...

I'm glad (although that really isn't the correct word here) there has been a lot of comments on possible causes for cracking of a canopy. I'm sure if we perused some soaring web boards, we would find quite a few as well... I know my old 1-34 had a lot of "stitching" on the canopy because we were too cheap to replace it.

Now the next obvious question:

For those of us with canopies that are just fine, what does the collective brain think should be done to mitigate cracking risk on installed canopies?



Based on the assumption that there IS preexisting stress on the canopy, either from installation loading or from thermal stress, what can we do ?

Obvious statement #1: Try to minimize environmental thermal shock so if winter ops, and you are in a heated hangar, cover the canopy with an insulating blanket when moving it outside.

Obvious statement #2: If you need to preheat your electronics in a COLD hangar, just preheat the panel.... NOT the whole cockpit.

There are other possibilities... like don't fly when OAT is below XXX Degrees..... but that's probably not too useful....


Any other ideas? There has to be a few....

Of course, the obvious one might be

Don't worry about it. Just fly the airplane and enjoy it. If it cracks at some point , deal with it..
 
Last edited:
There has been some discussion lately on cracked canopies and some speculation on the cold weather being the culprit....

...So, NOW with a deltaT of 37 deg we have:

3454*3 = 10360 PSI === SNAP

Chris,

Interesting calculation!

Any idea what the shear strength of the aluminum rivets might be? Perhaps they would fail before the canopy? I did an experiment on this long ago while I was building, by riveting together two ~3 cm squares of acrylic with their edges offset by ~5 mm (I think I used two rivets). I then squeezed the assembly edgewise in a vise, which caused shear failure of the rivets but no cracking of the acrylic. Repeated the experiment with an assembly that had been chilled in the freezer and got the same result. I'm not sure how relevant this is to the problem at hand but it did make me wonder about how much deformation the rivets experience in actual service.

Like some of the other posters I've been flying for over 1000 hrs in temps down to -20C and no cracks - yet.
 
Your rivet shearing experiment used the acrylic plastic primarily in compression,I believe the cracks result from tension occurring in or to the canopy from its relative quick reaction to shrinkage from exposure rapid cooling. Also repeated "G" loading can account for the warm weather cracked canopies such as Ron Shreck's. As stated earlier in this thread accurate alinememt of mounting holes
( canopy to frame) and impeccable workmanship of the holes in the acrylic, and an isolating rubber/ silicone bushing with washers will most likely give the best service.
Sikaflex maybe better but I don't trust the bond of the powder coat to the frame.
 
Good Morning All,

I have a couple of ideas concerning the canopies cracking when sikaflex is used. I wonder how thick the filet is on the canopies that failed. If my guess is correct, it's as thin as possible. Personally, I always try to keep weight to a minimum, and I expect most others do too. I don't think sikaflex has enough flex in it if the gap it's filling is too tight. That and it gets quite a bit stiffer when it's cold. I did the refrigerator/freezer test back when I used it on the RV8A. After seeing how stiff it became I opted to use a fairly thick spacers to set the gap.

Can those of you with cracked canopies post the thickness of the filet used. Maybe we can figure out if there's a minimum thickness that will keep this from happening, or if I have an over active imagination.

Thanks, Lance
 
Your rivet shearing experiment used the acrylic plastic primarily in compression,I believe the cracks result from tension occurring in or to the canopy from its relative quick reaction to shrinkage from exposure rapid cooling.

Interesting point. It doesn't seem like this should make a difference in terms of material strength though? It probably would affect the direction of crack propagation.

I do know there has been some deformation/movement of the aluminum rivets on my canopy, based on paint cracks that have formed around the rivet heads. This is most noticeable at the aft end of the canopy skirt.
 
SNIP

The thermal expansion Ct for steel is approximately 0.00008 in/inR
The thermal expansion Ct for acrylic is approximately 0.000417 in/inR

SNIP

I believe there is one zero missing from each coefficient.

4130 Steel: about 7 x 10E-6
Acrylic: about 40 x 10E-6

Both with in/in/R units.
 
Judoka question

Sika thickness is referenced frequently here to be desired at 1/8". I don't know how that is validated nor the Temp range corresponding. Perhaps someone has talked to Sika ????
 
I have three cracks in my RV-8 canopy after 315 hrs. I used Sika Flex with a 1/8" bond gap and sanded all the edges with 1000 grit sand paper to remove any stress risers. I did all this in a 100 deg garage in during a sweltering Alabama summer. I used a stock Vans canopy, which had about 1" gap between the plex and the aft portion of the canopy frame "hoop", starting at the canopy frame aft cross over brace and going equally around the aft end of the canopy frame hoop to the other side of the brace. Despite all sorts of machinations of trying to reduce the gap, I ended up clamping the plex to the frame with 1/8 spacers and then sika flexing it together, which forced the plex to fit despite having such a large gap. It wasn't a month later until the first crack showed up after an accidental ding in the plex, it cracked instantaneously. It was obvious that where the plex was under stress, it was going to be prone to cracking. I put a fiberglass Targa strip over the first crack and forgot about it. One day while flying faster than I should have in moderate turbulence with the A-P, I experience an a bit of wild yawing rolling event and added a second 4" crack, 1" in front of the cross over brace on the RH side....which got stop drilled and covered with a NASA meatball. About a year later, another 4" crack showed up 1" aft of the cross of over brace on the LH side....which got covered up with a US Flag decal. From what I have seen on other RV-8s and my own experience, most cracks happen around the aft cross over brace of slightly forward of it. A possible solution is to trim and re-fit the canopy frame hoop aft the cross over brace so that there are no gaps between the plex and the frame hoop. It seem that making the perimeter of the hoop about 1" wider around ends would help. Not sure if anyone has done this, but when I replace my plex after it cracks to the point of not being airworthy, I plan to do just that. I have heard all pros and cons of riveting vs Sika Flex, but in the end I think most RV-8 canopy cracks are due to the poor fit / pre-load of the plex...regardless of riveting or Sika Flex, which with a little enticement, will crack.
 
A related suggestion

Hopefully this is not too far off topic, but 5 or 6 years ago, after hearing of a rash of cracked canopies that occurred in cold weather with the aircraft parked in the hangar, I decided that when my RV-8 is parked I would not completely close the canopy, i.e. I do not latch the canopy when parked in the hangar.

I figure there's no need to put stress on the canopy when it is just sitting especially with the rapid changes in temperature we can experience in Texas. Perhaps it has helped, I have no way of knowing for sure, but so far no cracks. If you currently are latching your canopy when securely parked in a hangar, consider leaving the canopy just slightly open. It may make a difference.

Chris
 
floating canopy

Scott - how did you get your canopy to float in the air like that? Are you an illusionist on the side?
 
....have Jeff ( airplane plastic) stress relieve it. That removes initial stresses just to fit in the frame....

Larry, can you elaborate a bit about this?

If I understand what you did, you put the canopy into a fixture as it would be when it's complete, with the same forces and stresses and deflections, and then shipped this off to Jeff, who did something (?) to it and returned it, still in the fixture.

Thanks,
Dave
 
I believe there is one zero missing from each coefficient.

4130 Steel: about 7 x 10E-6
Acrylic: about 40 x 10E-6

Both with in/in/R units.

I have fixed the results in the first post. It changes things.

A LOT.

Thanks for the correction Alex.
 
Thread drift to answer David, hopefully useful !

Using a wooden form that defined the bubble perimeter and top surface at panel, roll bar and fastback. Jeff blew the canopy to conform to these objectives. As stated here by Scott, this method ( not a female mold ) will usually over low the side constraints attempting to achieve depth dimensions. I wanted straight sides conforming to the top longeron profile AND some straight vertical wall at the fastback to match the natural evolution of the fastback cone. Of course, all this can be achieved with clamps and static stress levels when you Sika as noted previously. So, I screwed steel angle to the wooded form to constrain the plexi canopy where I want it. Took it to Dayton. While it was in the oven for 1hr at 290F, went to lunch with Jeff( name dropping here), and when cool, the overblown curves were as straight as Angle iron. Jeff has the fixture now for RV4. E mail me for picks and I will forward. [email protected]
 
Except..... my canopy is still fine... and so are thousands of others..... So I have to conclude that there is some basic error in my calculations.. but, that being said, Danny's idea is still a good one.

Error or not, one simple piece of actionable advice I learned from your analysis is to not fit the canopy when it's hot!

The ideal temperature of all parts when drilling or gluing would be closer to mid-40s, given the temperature range they'll be exposed to when flying.
 
Error or not, one simple piece of actionable advice I learned from your analysis is to not fit the canopy when it's hot!

The ideal temperature of all parts when drilling or gluing would be closer to mid-40s, given the temperature range they'll be exposed to when flying.

Exactly. Factory guidance has us heating the plexi so it is less brittle and doesn't crack during installation. This causes problems when the canopy is exposed to much cooler temperatures during use.
 
Exactly. Factory guidance has us heating the plexi so it is less brittle and doesn't crack during installation. This causes problems when the canopy is exposed to much cooler temperatures during use.

I don't think any of the construction manuals recommend purposely making the canopy hot during installation. It does caution against doing it with it cold. In that situation use of artificial heat to attain normal room temp reduces risk of cracking while cutting and drilling.
 
Hindsight is always 20-20!! Dang...

Reading about the preload problem and stress relieving got me to thinking. I HAD an RV8 canopy frame, and gave it away when I decided to go to a Fastback. For FREE, what a big dummy I am. I could have clamped new canopies to that and stress relieved them.

However, there is a plan "B". If someone wants to measure their frame width at 2" stations and send me the numbers, I can weld up a stress relief jig.
 
Error or not, one simple piece of actionable advice I learned from your analysis is to not fit the canopy when it's hot!

The ideal temperature of all parts when drilling or gluing would be closer to mid-40s, given the temperature range they'll be exposed to when flying.

I think that's good advce. I'll add another bit. Grab a few pieces of scrap plexiglass, drill a few holes, and deburr them using your favorite tool. Now pull out a 10x or 15x magnifier and look at those holes. Viewed at the micro level, you'll be shocked at the nest of stress risers you've created.

Now experiment with using small, fine grit rotary rocks, shaped as needed, to deburr the holes by grinding the edges. Seeing is believing.
 
Now experiment with using small, fine grit rotary rocks, shaped as needed, to deburr the holes by grinding the edges. Seeing is believing.

Hey, that is MY secret and how I did mine years ago. It "still" hasn't cracked and looks as beautiful as the day I finished it...
 
First time builder here, and getting close to starting on the canopy (-7 slider)....so the consensus now is to cut and drill with colder temps? I was gonna wait til the warmer AL spring temps, but now I'm conflicted.
 
Another pass -

The OP took one approach, lets look at another.

Per Alex above:

Thermal expansion of steel = 7x10E-6, plexiglass is 40 X 10E-6, all per deg R

Lets say the steel is not deflected by the stress of the canopy, then the result

is the plexiglass strain is merely the difference of the two or 33 X 10E-6

Stress is E X Strain = 0.4 X 10E6 * 33 X 10E-6 = 13.5 psi per degree [R] delta.

Now, pick your stress at fracture, and a stress concentration factor to get the temperature delta.

Lets say assembly at 80F and operation at -20F or 100 deg = 1350 psi.

OK stress concentration at 3, then we have 3*1350= 4050 psi - still short of 9000 but getting closer.

Conclusion: the temperature alone is not the cause. Additional factors must add stress.

Factors:
1. external loads - latching, aero,
2. the fact that the canopy is only restrained by the edges, so the loading is not spread over a broader area of the canopy.

Since they can break in the hangar, not likely aero forces.
If edges are polished, not likely initiated due to poor surface finish at the edge.
If sika canopies fail in the same place, it is not stress concentration from fastener holes.

Jay may be onto something re: latching. Is there something that pushes up on the aft end of that frame?

BTW: I fastened a strip of plexi to a steel angle at 80F and then left it outside down to -20F and it survived.
 
Yes, temp changes can cause cracks. A couple winters ago my wife and I needed a large work space. I put the plane outside in -30 weather so we could use the heated hanger, and moved it back inside several hours later. A few minutes later I heard a loud suspicious "tick", and discovered a four inch crack that started at a side rail screw.

The hole in the canopy was oversized, deburred, and sides were Sika Flexed. Screw was torqued just enough to prevent it from falling out.

Stop drilled the crack with plexi bit, no trouble since.

Jay
 
Last edited:
Good discussion. One thing that may have been overlooked in the original calculations - you need to also look at the overall change in length of the plex vs steel, not just the short runs between fasteners. If you oversize the holes and avoid contact and stress between two holes - i.e. the plex has shrunk so that the outsides of those two holes now just touch the rivets or fasteners, the next two holes in each direction have already had that slack taken up and those two holes will be under stress + the subsequent holes even more stress. To never have any loads imposed on the plex, you would have to support it on floating fasteners or something that has the exact same coefficient of expansion. I'm thinking you'd be better off with "tighter" fit between fastener and plastic so stresses are all taken up hopefully evenly and at the same time?

I like Dan's tip - finer polishing of the plastic if you're making holes. I'm still leaning toward Sikaflex, but now will plan on doing the cutting with the plastic very warm, but gluing up the parts fairly cool.
 
The OP took one approach, lets look at another.

Per Alex above:

Thermal expansion of steel = 7x10E-6, plexiglass is 40 X 10E-6, all per deg R

Lets say the steel is not deflected by the stress of the canopy, then the result

is the plexiglass strain is merely the difference of the two or 33 X 10E-6

Stress is E X Strain = 0.4 X 10E6 * 33 X 10E-6 = 13.5 psi per degree [R] delta.

Now, pick your stress at fracture, and a stress concentration factor to get the temperature delta.

Lets say assembly at 80F and operation at -20F or 100 deg = 1350 psi.

OK stress concentration at 3, then we have 3*1350= 4050 psi - still short of 9000 but getting closer.

But of course, brittle fracture can occur at a flaw at stresses WAY less than normal tensile strength. Its like the stress concentration were in the thousands. I'm having a hard time finding data for the surface energy density for PMMA. If I can find a believable number, we can do a little bit of fracture mechanics. The one source I did find had dimensionally inconsistent units, so I don't think it applies.


BTW: I fastened a strip of plexi to a steel angle at 80F and then left it outside down to -20F and it survived.

If your strip has flaws along the edges that are representative of typical canopy edges, this is a useful test. I wonder if there may be an effect of non-uniform stress distribution. Try bonding a big square of plexi along one edge to the steel bar. This replicates the large portion of the canopy that is not strained by the frame.

It sure gets cold where you live!
 
Back
Top