Sikaflex debate continues.

Danger Will Robinson! Danger!

I masked the adhesive line, used 60 grit to scuff to metal, applied primer. Since I have no interest in debating the naysayers further, I will make no further posts on this thread, it's pointless.

My original post (on the other thread about "who can you trust") was to point out the fact that the usefulness of this forum was to allow everyone to state their position. The negative side of that is that oftentimes, where there's a difference of opinion, people start using descriptors like "ridiculous, stupid, misinformed, dangerous, crazy" and the like. Once a conversation reaches that dialog, basically all value (at least for me) is lost. This thread has reached that point.

Another angle

Not advocating either side. Just to look at things a bit differently, has anyone actually done a rough calculation for maximum aerodynamic load the canopy will be subjected to?

One would think that it would be less than a -4 g loading. That load distributed over the contact/rivet area would not seem to add up to very much.

A discussion vs an argument

A discussion is an interchange of intellect, an argument an interchange of ignorance.

Nothing is perfect; make your decision as informed as possible and proceed with care. 300 years ago, the cry in the New world was "Head west, young man." Some stayed in the east. I have flown my RV IFR in icing conditions and VFR over the Ice Cap of Greenland at -12C' so can attest to the shrinkage of the Plexi, as the canopy bow opened a gap and allowed the ingress of water. Now my RV resides in a more tropical environment and I have yet to find any indication of the Sika tearing or delaminating.

Van is not perfect neither am I; all designers have to make compromises. Van recommends the trailing edge wedge of the RV rudders & elevators be 'glued' with PRC. This in my mind is a bad compromise, as PRC is too flexible, making the trailing edge bulge between rivits. I used a two part structural adhesive the plastic planes use; do we start a thread on that now?

To me the interesting part of the Sika survey is that nobody has submitted a reply to say 'Done that, but am unhappy.'

Chalkie said:

A discussion is an interchange of intellect, an argument an interchange of ignorance.

Click to expand...

I love that, may I quote you???

Chalkie said:

To me the interesting part of the Sika survey is that nobody has submitted a reply to say 'Done that, but am unhappy.'

Click to expand...

I had been wondering if I was the only person to notice that.

Chalkie said:

To me the interesting part of the Sika survey is that nobody has submitted a reply to say 'Done that, but am unhappy.'

Click to expand...

I noticed that, too. I'm done after this post, though...when "engineers" have to make their point by calling people "zealots" and "guinnea pigs", I'm done. It's clear that nothing will ever change their minds, even though there's no evidence to support their claims.

lostpilot28 said:

I noticed that, too. I'm done after this post, though...when "engineers" have to make their point by calling people "zealots" and "guinnea pigs", I'm done. It's clear that nothing will ever change their minds, even though there's no evidence to support their claims.

Click to expand...

I think there may be something here. I'm working on a -6 today with a canopy that has been replaced once (Slider). It has about 12 rivets in the whole thing, the rest is being held on with RTV. How much pressure are really on these things?

This has been flying this way for 200+ hours. I just started working on it BTW.

Sorry,

Wrong Bob; I meant Bob "Captain Avgas" Barrow. What did you use to replace the powdercoat?

Paige
RV-8A

Bob Brown Danger Will Robinson! Danger!

--------------------------------------------------------------------------------
I masked the adhesive line, used 60 grit to scuff to metal, applied primer. Since I have no interest in debating the naysayers further, I will make no further posts on this thread, it's pointless.


Today 05:20 AM
PaigeHoffart Bob,

What sort of prep and coating did you use to replace the powder coat?

Paige
RV-8A

Click to expand...

OK, I lied, I'm back. But only because I have a technical question. Since it seems that the root cause of failure will be the "breaking" of the Sikaflex bond because of thermal expansion, I want to know if anyone knows actual numbers for our application.

I came up with a "rough estimate" that for a 100 degree (F) change, the steel canopy frame (roughly 50 inches long? I'm not home to measure it) will expand .0325 of an inch. The plexi is a little less than 7 times that much, so it would expand about .22 of an inch (maybe a little less). The delta is less than 1/5 of an inch. Keep in mind that this is for a 100 degree (F) swing.

Is my math correct on this? I'm looking at the tech specs for Sikaflex 295 UV and they list the "Elongation at Break = 500%". Does that mean it'll stretch over 500% before pulling apart? If so, that 1/8 inch bead would have to stretch quite a bit to break. Of course, this is assuming that the entire (or most) of the surface of the powder coating doesn't come off at the same time...or that the powder coating was removed down to the metal before using primer (although if the powder coating was secure, it's not required to do so...just an observation).

Sika Product Testing

gasman said:

Show us the SIKA product that you are talking about..............

Click to expand...

This gets technical. All Sika products including adhesive caulk and sealants say "meets"ASTM section number or "test to" ASTM section number They never say "Exceeds" nor do the chose the higher ASTM test code section It is always the minimum. If you look at Sonneborn or GE the say "meets or exceeds" and in their technical data sheets the use broader testing. They will often use more than one ASTM classification section.

All ASTM testing is done at a fixed temperature 75 degrees F. So elongation, tensile stress, lap shear, and recovery( which is the most important word) are the qualifiers to judge the product material by. Recovery is always presented in a percentage and it affects all other qualifiers (example Recovery 90% which means it has reduced all quantifiers by 10%)

Our cockpits very in temperature and pressure. ASTM test do not very temperature or pressure unless stated in the test. The sealant and adhesive standard ASTM testing classifications are at a fixed temp,75F and standard pressure. This includes all Sika Flex Sealants, Sikadur, Sika Bond, and all other Sika construction adhesives (not Sika structural adhesives which are mostly epoxies of some sort).

Plus when installing any caulk material the joint must be at it's mid point not it's fixed point or it's expanded point. I assume (not sure because I have never installed a canopy) a canopy is installed at a fixed point. This fixed point installation reduces the lap shear point, and elongation by over half and eliminates most of the recovery. Add varying temperature and pressure you no longer have known ASTM numbers.

You say your canopy doesn't move. This is not true all materials expand and contract due to temperature, pressure, and stress fluctuations.

I'm not saying it will not work. What I am saying is the there is no ASTM testing to quantify success or failure. The great unknown. And when not if(because there will be some type of failure) there is a failure it will be impossible to find out why. Because there is no base line testing method. Unless we know: Who is flying when, where, how long,what they were doing, temperatures, installation, pressure, type of preparation, and the list goes on and on and on we never find how to succeed with the product.

I don't want anyone to think I'm down on the Sika Corporation products I'm not. I use numerous products from them every day. Their concrete products are some of the most innovative in the industry . I just don't use any of their caulking products when I have to warranty my structural and waterproofing work past ten years ( or on my own home and personal buildings).

Alan

Sika Product Testing 295

lostpilot28 said:

OK, I lied, I'm back. But only because I have a technical question. Since it seems that the root cause of failure will be the "breaking" of the Sikaflex bond because of thermal expansion, I want to know if anyone knows actual numbers for our application.

I came up with a "rough estimate" that for a 100 degree (F) change, the steel canopy frame (roughly 50 inches long? I'm not home to measure it) will expand .0325 of an inch. The plexi is a little less than 7 times that much, so it would expand about .22 of an inch (maybe a little less). The delta is less than 1/5 of an inch. Keep in mind that this is for a 100 degree (F) swing.

Is my math correct on this? I'm looking at the tech specs for Sikaflex 295 UV and they list the "Elongation at Break = 500%". Does that mean it'll stretch over 500% before pulling apart? If so, that 1/8 inch bead would have to stretch quite a bit to break. Of course, this is assuming that the entire (or most) of the surface of the powder coating doesn't come off at the same time...or that the powder coating was removed down to the metal before using primer (although if the powder coating was secure, it's not required to do so...just an observation).

Click to expand...

At 73 F degrees you have a 50% recovery rate this means if your boding sides must be twice as wide and the mid point of your joint an you will get 200% elongation. Meaning the joint if at the 1/8 your bonding sides must be 1/2 inch and it will break at 1/2". But it will flex on each side of the mid point 1/16" with out break down of the molecule structure. This is all at 73F degrees.
Alan

PaigeHoffart said:

Sorry,

Wrong Bob; I meant Bob "Captain Avgas" Barrow. What did you use to replace the powdercoat?

Paige
RV-8A

Click to expand...

Hi Paige, this is what I did to ensure excellent coating adhesion and performance.

I had the canopy chemically immersion stripped to remove Vans coating. Most powdercoaters will not have this service so you may need to go to a company that specialises in chemically stripping powdercoated products. I recommend this process as opposed to having all Vans powdercoat blasted off (safer...better).

This will leave you with a Rust Grade A finish (Steel with millscale layer intact and very minor or no rusting).

Take the stripped canopy to a high quality powdercoater that has blasting facilities. Specify a Class 3 blast done with a 30/60 mesh garnet at LOW PRESSURE (approx 30 psi) or equivalent. This will produce a complete blast clean with consistent metal colour all over and no visible contaminants.

Specify a primer of Orica/Dulux Zincshield Powdercoating Primer. This is a high zinc containing primer specifically formulated for powdercoating applications over steel. It inhibits corrosion creep under the powdercoat by an anodic process while exhibiting excellent intercoat adhesion properties. It is typically applied at 80 micron film build.

Specify a compatible Orica/Dulux polyester powdercoat top coat such as Dulux Duralloy (Product code 272 Line) or equivalent which is suitable for use on steel in exterior applications.

Provide the powdercoater with a small sample of 4130 tube with a wall thickness the same or VERY similar to the canopy tube and get him to run it through all stages (including blasting) with the canopy. Similarity of wall thickness is important to ensure you do not get false results. You will use this tube sample for batch testing. Without destructive batch testing you can NEVER know that you have an adequate adhesion (because you're not going to destructively test the canopy itself).

Batch test the sample by scribing 10 lines at 2mm centres in one direction and then another 10 lines at 90 degrees at 2mm centres to form a cross hatch (ie forming lots of little squares with each square being 2mm x 2mm). Scribe the lines with a "Stanley" knife making sure that you are cutting right through the coating to the metal. Clean the cross hatch with methylated spirits, dry, and apply a fabric reinforced packaging tape over the squares. Leave enough spare tape to get a good grip to pull on. Let sit for 24 hours to allow the tape adhesive to get a good bond. Pull off as violently as you can. If no squares come off then you can be assured that the powdercoating adhesion is truly excellent and you have a coating sytem that will last the life of the plane. If you lose four or five squares it is still OK. If a third or more of the squares come off then something has gone wrong (REJECT).
Now you have a system that will never fail.

Of course you will need to sand the powdercoat as specified by Sika to ensure proper adhesion of the Sikaflex (and use the appropriate primer). I would also use the batch sample to build up a full joint profile including the powdercoat plus the sikaflex plus some acrylic sheet. Then you can test this small sample to be assured that the TOTAL system has adequate adhesion. HOWEVER I am at pains to emphasise that adhesion is only one part of the equation. The design of the sealant joint geometry is the BIG issue (and the one least understood).

The Sikaflex process is one whereby aircraft structural components are being chemically bonded. That is a demanding regime and it requires quality control at all stages. By definition quality control means CONTROL OF THE QUALITY of all crucial processes. In other words if you bond onto a powdercoated surface but you have no way of knowing the integrity of the powdercoat system then you have no way of knowing that the total system will not eventually fail. Under those circumstances you just live in hope that everything will hold together without ever really knowing.

Of course my advice on how to achieve a 100% reliable powdercoat system in no way signifies that I endorse the use of Sikaflex as a sole use product to bond on an aircraft canopy. As you will know I did not go down that path myself because I had specific technical reservations about long term performance.

In the long run I suspect that some builders will have considerably more success with the Sikaflex process than others (ie some canopies will last a lot longer than others) based purely on how much they understand the technologies involved and on their attitudes to quality control. At the moment everybody is just doing whatever takes their fancy and I doubt there are any 2 canopies exactly the same in terms of design and method. It is therefore logical that some will outperform others.

I note that Bob Brown stated in an earlier post that he sanded the powdercoat off completely in the area where he applied the Sikaflex (ie he bonded to bare steel). I would have grave reservations about the long term success of that approach. He has no passivation of the steel to prevent long term corrosion undercreep spreading invisibly under the Sikaflex and thus destroying the adhesion of the Sikaflex to the steel.

Hope I've been of some assistance.

Show Me

If some one will send me a picture of their canopy joint application then I would understand this better and my comments would be of better use to you all. Is there a detail in my preview plans of my RV8? If you all are just placing a bead of caulk in a void that does not have two clear and defined sides with a mid release point you have little or no elongation. With out elongation your adhesion bond to the side walls will fail sooner than later.
Alan

What is Sikaflex?

I'm late, I know, but what is Sikaflex? I gather that it's an adhesive with some awesome strength and flexability. What other applications has it been used for in aviation, or anywhere else for that matter? I'l be adding a tip-up canopy to my -9 and this is another "something" I need to look at and investigate. Will it work on tip-ups?

Thanks!

Incomplete poll

I couldn't take the poll because there wasn't a choice that applied to me. I guess a selection of "OTHER" would have worked, but I would have preferred one that ask:

O Both rivets and Sikaflex.

After around 800 hours of flight, the Doll's canopy cracked at the second rivet aft of the bow on the left side. It was a very very cold day, and I was in the worst turbulence I had ever flown her in. I had already decided to remove the canopy, and add Sikaflex to the gap between the canopy frame and the glass on both sides of the tube, and had already purchased the Sikaflex. Unfortunately I hadn't got around to doing it yet.

When I replaced the canopy glass, I prepared the surface of the glass and frame as if I were going to us the Sikaflex method. I riveted the glass on per Van's procedure and then I added the Sikaflex in the seam to bond the glass to the tube. I believe the Sikaflex will add to the rivets and relieve any stress on any single rivet. I don't have to worry about the engineering of using just Sikaflex as the only bond because the Doll's canopy is riveted on per plans. I also believe I have added the advantage of using the Sikaflex to prevent cracks.

same here

Like Danny King, I have an other on the poll. I have both rivets and sikaflex on the windshield on an RV7a slider. I used sikaflex to glue both sides of windshield in, because it wanted to expand. I also used sikaflex, and screws per the plans on the box. The sikaflex covers the gap caused by the spacers across the bow, and provides added strength.

I added the "Both" choice to the poll.

Dont want your vote to go un-counted.

No comments on the stress actually on the canopy? I'm really not a fan of doing something without physical fasteners, but I think this whole debate is making a mountain out of a molehill. If the canopy doesn't have the pressure on it to remove itself in flight, then why the heck would it matter WHAT you glued it in with?

I mean seriously, there are more people out there flying with nosewheels on RV's that have a proven problem (the old design...) still than Skia.

osxuser said:

No comments on the stress actually on the canopy? I'm really not a fan of doing something without physical fasteners, but I think this whole debate is making a mountain out of a molehill. If the canopy doesn't have the pressure on it to remove itself in flight, then why the heck would it matter WHAT you glued it in with?

I mean seriously, there are more people out there flying with nosewheels on RV's that have a proven problem (the old design...) still than Skia.

Click to expand...

It does make you wonder given the fact that the higher pressure on the outside of the canopy tries to force it closed (hence the canopy cannot be opened in flight). If there were constant forces pulling the plexi away from the frame, I'd probably think differently about any means of fastening (sika, rivets, or otherwise).

Canopy Fastening

I'm taking this approach: When working on the fuselage, if I got a good metal to metal fit I riveted; when I didn't I used epoxy or proseal as a shim, then rivets.

I am setting my canopy frame. I am going to use sikaflex as a shim so I am not pulling or twisting the frame to fasten the canopy. Then I am going to use rivets and screws. Best of both worlds?

Silkaflex

Considering the B-737 interior overhead panels were originally installed with velcro and the skins were glued together, which both proved to be not the way to go,
I would say that time will prove if Silkaflex is up to the
task. I installed my canopy with it and am satisfied with the result. The windshield in my truck is only held in with a like compound and it's gov. approved just like the 737 skins were.
Walt Ray

Once again.

osxuser said:

No comments on the stress actually on the canopy? I'm really not a fan of doing something without physical fasteners, but I think this whole debate is making a mountain out of a molehill. If the canopy doesn't have the pressure on it to remove itself in flight, then why the heck would it matter WHAT you glued it in with?

Click to expand...

Stephen, this matter was fully addressed in my posting #38 on this thread. I refer to a couple of my comments in that thread:

The actual tensile stresses on the sealant under loads normal to the bonded surface will be quite minimal. However the shear stresses in the sealant due to lateral loads from differential expansion will become massive as the sealant bead height approaches zero.

If you don't understand what this all means then I would respectfully suggest that you are not a good candidate for attempting the exercise.

In other words the stresses within the sealant resulting from live loads (wind loads) on the canopy are likely to be insignificant compared to the stresses within the sealant resulting from differential thermal movement of the acrylic across the steel. In other words the stresses normal to the bond surface (resulting from loads at 90 degrees to the bond surface) are likely to be insignificant compared to the lateral stresses (shear stresses) in the sealant (resulting from loads parallel to the bond surface).

Or put another way, depending on the geometry of the sealant joint, the stresses induced in the sealant on a very hot (or very cold) day when the plane is just sitting on the tarmac could be hugely larger than the stresses induced in the sealant as a result of the plane flying through the air. But bearing in mind of course that these stresses are accumulative.

That is why IT IS IMPERATIVE THAT THE SEALANT JOINT GEOMETRY MUST BE WELL DESIGNED TO MINIMISE ANY LATERAL STRESSES.

I can't explain this in any simpler terms. People who keep referring to wind pressure on the canopy in flight as if it was the main issue are missing the point. People who keep referring to the adhesion qualities of Sikaflex as if it was the only important issue are also missing the point.

And once again.

buck said:

I installed my canopy with it and am satisfied with the result. The windshield in my truck is only held in with a like compound and it's gov. approved just like the 737 skins were.
Walt Ray

Click to expand...

I would refer you to my post #41 on this thread which specifically deals with glass windscreens on vehicles and reveals why there is NO COMPARISON with the use of Sikflex on acrylic RV canopies.

It is obvious that people are simply not reading previous posts...or they're not understanding them.

Interesting

Well, I don't have a dog in this hunt but one of our local RV 6A owners did his tip up with Sika. It looks great and shows no signs of issues or problems even in the extreme AZ heat.

I can see the concern with potential chemical interaction with the canopy. From my experience any negative interaction will occur relatively quickly. On a helicopter I used to fly one of the pilots grabbed the "wrong" can and wiped down a side window causing it to fog within minutes. A little acetone will instantly cause problems as well as other chemicals.

With the Sika canopies I've observed, some 2+ years old, there is no sign of any negative interaction. After all, the front of the screen is glassed in anyway. Where will the failure occur?

Today, I had a windshield replaced in my van, at my house. The installer had his van open and I asked what they use. They use Sika Sierra. This adhesive is heated to 175 before installation. The area is prepped with a prep pad.

I asked if it could be used on Lexan or other acrylic type materials. He said it is used on a variety of non glass windows. He related the Sika Sierra must meet Federal requirement for strength in the event of a roll over. The windscreen is structure and therefore the adhesive must meet structural requirements also. He said the Sika Sierra 1000psi in strength, what ever that means in this purpose.

I've read all of the posts. In review, so far there has not been one problem reported with the use of Sika. There has only been anecdotal evidence that there should be problems.

Although this in alternative method of attachment, it is not not like the alternative engine, where there have been numerous issues. So far, none reported with Sika.

I'm hoping to start another project soon and will definitely be researching this method of attachment. I cracked my canopy on the last hole I drilled. Frustrating.

If I understand correctly, the concern is that the plexi expands in length a lot and the steel very little. This difference has to be accomodated by the sealant bead. A thin bead (small gap) will not allow for the relative movement and will tear loose. I think that this is also a concern when rivets are used. Has any serious thought been given to making a canopy frame out of something whose expansion rate is closer to plexi?

szicree said:

If I understand correctly, the concern is that the plexi expands in length a lot and the steel very little. This difference has to be accomodated by the sealant bead. A thin bead (small gap) will not allow for the relative movement and will tear loose. I think that this is also a concern when rivets are used. Has any serious thought been given to making a canopy frame out of something whose expansion rate is closer to plexi?

Click to expand...

Steve, thank God some-one seems to understand what is going on. I was starting to get depressed.

Sealants work quite well bonding glass to steel (ie car windscreens) because glass and steel expand and shrink at virtually the same rate given a similar temperature change (ie they have very similar coefficients of linear thermal expansion...CLTE).

However bonding acrylic to steel has ALWAYS been regarded as dodgy because acrylic has a CLTE roughly 7 times that of steel. In other words for every 1mm that steel expands in a given temperature increase acrylic will expand 7mm. That huge differential in movement can really stress the sealant bead, particularly if as you have pointed out, the gap is too small (ie the sealant bead has insufficient height). In addition as time passes and the sealant loses some of its elasticity it becomes more prone to rupture or delaminate.

In answer to your question, having the canopy frame made of epoxy or aluminium would lessen that stress because both of these materials have a CLTE that is closer to acrylic.

The problem I find with the whole Sikaflex discussion is that raising technical points is like writing a message in the sand at low tide....at the end of the day the information just gets washed away as if it never existed.

Where's the beef?

Captain Avgas said:

The problem I find with the whole Sikaflex discussion is that raising technical points is like writing a message in the sand at low tide....at the end of the day the information just gets washed away as if it never existed.

Click to expand...

I don't think it is being ignored, but all the emperical data is demonstrating that it is just not a problem.
Just because the two materials have very different CLTEs doesn't automatically mean a bond between them will be a disaster. If the actual amount of movement between the two (probably a few hundred thousandths - I haven't done the math) creates shear forces within the elastic capability of the bond material, then there is no problem. Long term degradation effects are TBD of course, so monitoring and inspection are warranted (like everything else). Again, the emperical data so far bodes well, so the experiment continues.
What we need to take away from this discussion (IMNSHO) is:
- properly prepare the surfaces to be bonded
- provide a reasonable thickness of bond between the materials
- apply the bonding material according to the instructions

The technical points should be used to determine the proper engineering solution, not simply toss the concept.

DGlaeser said:

I don't think it is being ignored, but all the emperical data is demonstrating that it is just not a problem.
Just because the two materials have very different CLTEs doesn't automatically mean a bond between them will be a disaster. If the actual amount of movement between the two (probably a few hundred thousandths - I haven't done the math) creates shear forces within the elastic capability of the bond material, then there is no problem.

Click to expand...

This is what I was trying to find out with my previous post...If I'm correct, the plexi will move about 1/5 of an inch with 100 degree (F) change. I'm assuming that the 1/5-inch will NOT be completely isolated to one section of the frame and dispursed over the whole thing, therefore at each end (say front and back?) of the canopy you're looking at 1/10th of an inch of flex at the bond joint. Again, this is for wildly swinging temps (not unheard of, just not real common).

I also agree with the statement that no one is ignoring the data being provided, but the evidence is showing that either that data is somehow flawed, or everyone who's used Sikaflex to date is doing a great job of applying it because it appears to be working quite well. Maybe there's a mix of both at play.

Glued mine but haven't flown yet.

Cap'n-

I understand your concern about a thin glue joint not having enough material in place to absorb/tolerate the differences in expansion between the plexi and the steel. This may be a greater concern on the -8's than the side by sides due to the greater length of the steel frame on the -8. I used the 1/8" tubing spacers "recommended" by posters here and elsewhere on the net but I think that next time I would do away with them.

Now before you start banging your head on the keyboard... I spoke with Sika USA about how to apply the adhesive between 2 perpendicular surfaces. My specific concern was how to apply the glue between the canopy and the side bows on my -7A slider canopy. They said that a fillet type joint was structurally as strong as a 1/8" lap-type joint. The thing here is that a fillet of glue applied into the corner between 2 surfaces has a lot of material and at the thickest point is 3/8-1/2" thick.

The 295UV Sika Flex is able to tolerate something like a 600% elongation before it tears (I'm operating from memory on that stat so I may be off some). That would suggest that the 3/8" thick fillet joint would have to be stretched to 2.25" before it would lose it's integrity. Most people make a fillet joint between all bonded surfaces for cosmetic if not structural reasons so it seems to me that this bond has plenty of adhesive to tolerate stresses from the different thermal expansion of plexi vs steel.

I'm NOT an engineer but I did take a physics class once and even the 1/8" joint seems adequate to me. Let me try to put a pencil to it but someone smarter than me might want to double check my math:

• 600% of 1/8 inch = 3/4" (failure of glue joint).
• Acrylic experiences 7X's the thermal expansion of steel (per Cap'n Avgas).
• To achieve a 3/4" expansion difference between the plexi and the frame, the frame would have to expand 0.125". At this temperature change the plexi would have expanded approximately 0.875 (7 x 0.125). The difference of these two is 0.75" and the presumptive point of failure for our glue joint.
• The coefficient of thermal expansion of steel is 8.6 x 10-6 inches/inch/degF (which I found on the net and so if this number is not totally precise, is at least in the ballpark).
• A 1 degF temp change of a 30" length results in an expansion of 0.000258". (30 x 0.0000086 = 0.000258). I used 30 inches because that's about 1/2 the length of the front or rear bows and the expansion will occur in 2 directions.
• Therefore a temperature change of -20 to +120 degF will result in a change of 0.03612 inches (0.000258 x 140 = 0.03612)

This shows an expected expansion of about 29% of what would be required to fail this joint. And that would only occur if the glue was applied at one of these temp extremes (-20 or 120 degF). If the glue was applied at the midpoint of this temperature range (50 degF) then the stress on the glue at either end (120 or -20 degF) would be half of this since the actual temperature change would be from 50 degF to either -20 or 120.

I think this shows a substantial safety margin. I think you'd be hard pressed to expose even a 1/8" joint to enough temperature range to approach that 600% elongation/failure point. Add to that the fact that most (? all) builders using Sika Flex finish the joint with a relatively thick fillet of glue and I just don't see one of these failing.

Regards,

Thermal expansion

Keep in mind we are dealing with the thermal expansion of steel when used on the slider model and the thermal expansion of aluminum on the tip up model.

dpansier said:

Keep in mind we are dealing with the thermal expansion of steel when used on the slider model and the thermal expansion of aluminum on the tip up model.

Click to expand...

The coefficient of thermal expansion of steel is 8.6 x 10-6 in/in/degF and aluminum is 13.7 x 10-6 so the difference in expansion with the acrylic canopy is less with aluminum than it is with steel. So the slider canopy is a worst case example in these planes.

Sika flex

Captain Avgas said:

In answer to your question, having the canopy frame made of epoxy or aluminium would lessen that stress because both of these materials have a CLTE that is closer to acrylic.

Click to expand...

My tipup frame is aluminum, thank goodness. I believe that if we follow Sika's specified instuctions for this application (1/8" minimum fillet etc.), there should be few problems. As posted previously, with the front fairing fiber glassed/sika'ed and the side aprons rivetted, the only area on the canopy which is not "reinforced" is the trailing edge which has no other support, other than Sika acrylic adhesive. Captain Avgas. I'm enjoying sitting on the sidelines reading these different posts. Thanks for bringing up some important safety issues with regard to our projects.

The boy with his finger in the dike.

rvtach said:

Now before you start banging your head on the keyboard... I spoke with Sika USA about how to apply the adhesive between 2 perpendicular surfaces. My specific concern was how to apply the glue between the canopy and the side bows on my -7A slider canopy. They said that a fillet type joint was structurally as strong as a 1/8" lap-type joint. The thing here is that a fillet of glue applied into the corner between 2 surfaces has a lot of material and at the thickest point is 3/8-1/2" thick.

Click to expand...

Dear Jim, please don't be offended (I truly don't mean to offend) but I think this might be a case of a small amount of knowledge on your part being dangerous.

I am imagining that you are referring to a "fillet" bond as a bond between two perpendicular surfaces (ie the bond edges are at 90 degrees to each other and touch at the root of the fillet and the "fillet" itself is triangular in cross section).

If that is the case then that is virtually the worst case scenario for an elastomeric bond between two parts with differential movement because it places INFINITELY high stresses at the root of the fillet. The sealant bead will fail almost immediately at the root (ie the point at which the 2 bonding surfaces converge) regardless of the side dimensions of the bead.

As a sole means elastomeric bond that configuation is useless. To make that configuration work you would need to run a closed cell polyethylene foam backer rod along the root before applying the sealant. This will prevent the sealant bonding to the root (and it will not bond to the polyethylene). I have attached a drawing to explain this better.



I would suggest that when you asked for advice from Sika one of two things (or possibly both) occurred.

1. You did not give them a drawing to fully detail exactly what it was you were proposing to do. You simply rang them up and everyone became confused trying to discuss a very technical matter over the phone.

2. You were talking (as many builders seem to do) with the absolute bottom of the food chain in scientific knowledge at Sika and received very poor information. The bottom of the food chain at Sika is the pimply faced kid just out of college who has absolutely no science background but goes by the title of "technical representative" when in reality he is nothing more than a salesman trying to increase his quota by flogging off a few more tubes of goo to whoever.

*****************************************

To all and sundry. I have to give this away now and leave you all to your own devices because it is never ending and I'm starting to feel like the boy with his finger in the dike.

In the final analysis I have come to the conclusion that the enthusiasm with which builders embrace Sikaflex as a SOLE means of canopy restraint is inversely proportional to their scientific understanding of the subject.

My best advice is to not use Sikaflex as the only means of restraining your canopy. Back it up with mechanical fastening of some type as insurance. That is the conservative thing to do.

In my opinion there WILL be Sikaflex canopy failures.... eventually. I do not know if the first one will be in a day or a month or a year or several years...but I am 100% certain that there WILL be failures given sufficient time and exposure.

I say that because it is quite obvious that most builders are not technically up to the very demanding task of designing and fabricating a fully bonded canopy with materials exhibiting such vastly different coefficients of linear thermal expansion. In particular the understanding of optimised sealant joint geometries is an absolute science in itself and RV builders who use "intuitive" principles are destined for disappointment. Those who blindly follow others who have no expertise in the area will obviously fare no better.

It is also not possible to predict how the impending failures of canopies without any mechanical fastening might manifest themselves. At the one extreme they might be local and lend themselves to repair if they can be detected early. At the other extreme they might result in something much more catastrophic.

Thank you to those people who emailed me personal messages thanking me for my technical input to the subject.

Cheers

How about this...

First, thanks to Capn Avgas and the others who have contributed to this debate. I certainly understand the issues here better. It would be nice if Capn can continue to contribute... But, alas, it does take a lot of time to type well-thought out posts.

How about this... Would it be helpful to design in a few early indicators of pending failure. By that I mean, could we select a few spots where we would purposefully use a smaller fillet for about an eigth of an inch or so. These weaker bonds will hopefully show the signs of stress well before the rest of the joints fail, thus alerting us to the need for proactive measures. Kind of like wear bars on tires. Is this practical?

It will be obvious...

apatti said:

First, thanks to Capn Avgas and the others who have contributed to this debate. I certainly understand the issues here better. It would be nice if Capn can continue to contribute... But, alas, it does take a lot of time to type well-thought out posts.

How about this... Would it be helpful to design in a few early indicators of pending failure. By that I mean, could we select a few spots where we would purposefully use a smaller fillet for about an eigth of an inch or so. These weaker bonds will hopefully show the signs of stress well before the rest of the joints fail, thus alerting us to the need for proactive measures. Kind of like wear bars on tires. Is this practical?

Click to expand...

Pending failure will be easy to spot with regular inspection. The 'weaker' spots in your concept will only show signs of stress if the 'stronger' ones do as well. The softness of the Sikaflex means you will see evidence of movement well before the material is close to failure.
I hate to be contrary, but Capn Avgas's explanation is way off base. I think the Sikaflex folks have been doing this for a while now and do have a clue about how their product works.
A proper bond would have not only a complete fillet, but bonding agent between the two parts as well, in a decent gap. As the two parts move, due to thermal expansion (or any other force), the Sikaflex 'stretches' to accomodate the movement. A nice fillet reduces the peel stresses that would tend to pull the Sikaflex away from the surface at the edges.
If in doubt, bond a couple of scrap pieces of plexi together with Sikaflex and judge for yourself how well it works.

Cap'n Avgas,

I feel your pain. If I understand your position, it is not that Sika is an unsuitable product for this application, only that there has been little true engineering put into the Sikaflex use. Your point that the shear stresses go to infinity as the gap goes to zero is absolutely lost on many. It's like folks saying that you can't tear a phonebook in half, but we've all seen it done by concentrating the stress to a small area.

DGlaeser said:

I hate to be contrary, but Capn Avgas's explanation is way off base. I think the Sikaflex folks have been doing this for a while now and do have a clue about how their product works.
A proper bond would have not only a complete fillet, but bonding agent between the two parts as well, in a decent gap. As the two parts move, due to thermal expansion (or any other force), the Sikaflex 'stretches' to accomodate the movement. A nice fillet reduces the peel stresses that would tend to pull the Sikaflex away from the surface at the edges.
If in doubt, bond a couple of scrap pieces of plexi together with Sikaflex and judge for yourself how well it works.

Click to expand...

There's nothing like accusing a man of being a bullsh*tter to draw him back into the fray.

So here we go again.

The problem with this discussion is that many people are totally confused about the difference between an adhesive and a sealant.

For everybody's information.....a sealant is designed to allow movement at a joint....whereas an adhesive is designed to restrict movement at a joint. And it is important to know which is which because the optimum geometry of a joint bead will differ depending on whether it is acting primarily as a sealant or primarily as an adhesive.

Sikaflex is an elastomeric SEALANT formulated with the specific intent of allowing movement at a joint. In the case of an RV canopy the Sikaflex acts as a "structural" sealant in that it is also expected to transmit loads. That the Sikaflex "adheres" to the surfaces does not make it primarily an adhesive. Virtually all sealants are "adhesive" by nature but they are not PRIMARILY adhesives. Epoxy is an adhesive (and it may act also as a sealant...but it is not PRIMARILY a sealant).

It is relevant that in his post #78 Jim McChesney claims he asked Sika "how to apply the ADHESIVE between 2 perpendicular surfaces" and was advised to use a full depth fillet. And that would be correct advice for an ADHESIVE....but incorrect advise for a SEALANT.

I have done a few quick (and rough) drawings to indicate a cross section of the canopy to side bow joint on the RV7 slider with 3 possible joint geometries utilising Sikaflex. You will need to left click on the thumbnail to bring it up to a reasonable viewing size.

Drawing A shows the detail that I proposed in my post #82 using a polyethylene bead. I recommended this because it's probably the simplest solution to implement.

Drawing B shows an alternative joint geometry using a polyethylene bead (between the bow and the acrylic) which would work just as well but would require moving the acrylic away from the bow by up to 1/4" which might prove difficult without serious mods to Vans canopy design.

Drawing C shows a Sikaflex joint geometry that is destined to fail. It's a good geometry for an adhesive but a very poor one for a sealant. Infinitely large stresses will arise at the root of the fillet as the side bow and the acrylic move differentially.



The fact that I make these points should not be construed by anyone as an indication that I recommend using Sikaflex to bond RV canopies without an additional system of mechanical attachment.

szicree said:

Cap'n Avgas,

I feel your pain. If I understand your position, it is not that Sika is an unsuitable product for this application, only that there has been little true engineering put into the Sikaflex use. Your point that the shear stresses go to infinity as the gap goes to zero is absolutely lost on many. It's like folks saying that you can't tear a phonebook in half, but we've all seen it done by concentrating the stress to a small area.

Click to expand...

Hi Steve, basically you're correct but I'll clarify my position. I believe that Sikaflex is unsuitable to use as a sole means system of RV canopy attachment. The large expansion and contraction of acrylic makes it an absolute bugger to work with in bonded glazing systems. And it becomes more problematic as the area of the acrylic panel becomes larger (and an RV canopy is a VERY large piece of acrylic). That's why professional designers use mechanical restraint on anything other than small acrylic panels used in non critical applications. Look at formed acrylic skylights for domestic and industrial use....they're ALL mechanically restrained using the picture frame method. Its the same for Pipers and Cessnas and any other certificated aircraft. And I'll bet dollars to donuts that Dick VanG's new RV12 is not about to appear with a Sikaflex bonded canopy.

Using Sikaflex alone to secure an RV canopy is REALLY pushing the product to the edge of its envelope and to have any reasonable chance of long term success would require that all joint geometries be truly optimised. And it is obvious from this thread alone that very few RV builders have the technical expertise to achieve that goal.

Flip-floppin' views?

szicree said:

Cap'n Avgas,

I feel your pain. If I understand your position, it is not that Sika is an unsuitable product for this application,....

Click to expand...

Huh? Did you miss the first half of this thread?

Doing double duty...

Captain Avgas said:

The problem with this discussion is that many people are totally confused about the difference between an adhesive and a sealant.

Click to expand...

There is no reason the same material can't be BOTH.
From the Sika website:
"Sikaflex? 295UV is a moisture curing polyurethane sealant/adhesive that exhibits outstanding UV resistance. This high strength elastic adhesive is specifically formulated for use with plastic windows. Sikaflex? 209 is a black primer that develops tenacious adhesion to acrylic and polycarbonate. Used as a system, Sikaflex? 295UV and SikaPrimer? 209 provides very durable adhesion to plastic windows." (I added the bold)
The ability to flex is a built-in stress reliever. As long as the load is within the strength capability of the material, nothing happens (rigid or flexible). If the movement is within the elastic range of the material, it returns to it's original shape when the load is removed. Your wing does this every time you pull a few G's.
A joint made with 'super glue', which is very rigid/inelastic, will hold with virtually no movement right up to its load limit, then fail totally. A sharp edge driven between 2 surfaces super-glued together can pop it right off.
Use 'Goop' (another adhesive/sealant - but not for canopies ) for the same job and the joint will flex a lot before it fails, it will fail much more gradually, even if you pry it off. (assuming good adhesion in both cases of course)

DGlaeser said:

There is no reason the same material can't be BOTH.

Click to expand...

I agree with that, automotive windshields have been using the same type of stuff for years to both glue in as well as seal the windshield to the structure of the body.

Sikaflex looks good now, will inspect in the future.

I have just finished my slider canopy with Sikaflex. To make sure the bonding to the frame is not depending on the Vans-Powder coating, I carefully removed the powder coating in the "bead area". To avoid corrosion, I made sure the Sikaflex primer covered this area well and some of the powder coating. Additionally the Sikaflex and the frame were painted with a 2K-paint, afterwards. Although I am no expert, I do not have any concern for corrosion at this moment. If indeed the frame starts to corrode, this does not mean that my canopy will suddenly fall off.

To allow the plexi to expand, I made sure I have +/- 4 mm thick beads between the frame and the plexi (in some area's it's even more). Also between the side bow and the plexi I have a 3-4 mm bead. I guesstimated the bonding area and estimated that it would take approx. 6 tons to tear the plexi from the frame. Obviously the frame, that is only held down with 3 plastic blocs in the rear and 2 plastic rollers in the front (oh, yes and the fiberglass bow of course) will not withstand these forces.

Anyway, a slider canopy does have some form of additional mechanical attachment of the canopy bubble: At the front there is the Glassfiber bow that goes over the windscreen and front edge of the canopy, holding it down. In the rear there is the rear skirts that are riveted to the sides of the frame and go over the bubble at the rear. Additionally there is the side skirts that keep the bubble from going out sideways.

I was wondering what happens to a classic canopy, when one or more rivets fail (maybe due to expansion of the plexi?) will the "zipper effect" rip out all the other rivets as well? With the Sikaflex, if an area is going bad, that does not mean the canopy is going to fly off on its own the next minute.

I did plan on a thorough regular inspection of the canopy anyway (hey, would that not be every minute you fly? how can you "not" look at your canopy every time you fly?).

Everybody needs to do what he thinks is good for his situation. At the moment I am confident about my canopy. If problems occur, I can still drill and put the rivets in additionally.

Regards, Tonny.

I don't post much here, but I read alot. I have an RV-8A that I used SikaFlex to fasten the canopy to the rails. It requires one to pay close attention to proper processes. Probably more than the average builder is willing to do.

But we are all in the experimental business. I just don't see a statistically signifiant catastrophic failure mode that is going to impact safety of flight. The technical posters have good arguments, but the consquence of degraded adhesion over time probably does not justify all the concern. I spent 20 years as chief engineer/structural chief in the SPO on the A-10, F-111, and A-7 fleets. What we are doing does not even come close to the risks in many modifications I have authorized over the years. But then I also had a SPD whose philosophy was that was why we put ejection seats in the planes.

We always stress tested our airframes to several lifetimes in huge wiffle trees to validate the structural integrity. They still fell apart. I do not think any RV airframe has ever been streess tested to validate its structrural integrity, but we still fly them happily, at least I do. Everybody needs to grab a cold one and relax.

Let's see the drawings

DGlaeser said:

There is no reason the same material can't be BOTH.

Click to expand...

Dennis, instead of more words why don't you just post a cross sectional drawing of what you actually did at the canopy-to-side-bow on your RV and open the way for some critical evaluation of the methodology. If you're proposing this detail for others to follow (which you seem to be doing) then surely you wont mind putting your theories in a format (a drawing) that others can understand and comment on.

The same advice goes to others hovering on this thread that have finished a bonded canopy and recommend it to others. It's time to stop hiding behind cheap words and bravado. If you really want to foster understanding of the bonded canopy system in a technical sense there has to be drawings presented. Comments like "it worked for me" benefit no-one. Without drawings to discuss and analyse builders will just keep making the same mistakes over and over....the understanding of the technology will remain at the same amateurish level.

I note however that in post #62 Alan Cook (a professional sealant contractor) asked for proponents of the bonded canopy to submit a "picture of the canopy joint application" for assessment. But no pictures or drawings were forthcoming.

If builders are truly comfortable that their sealant methodology is up to scratch then they shouldn't mind submitting drawings.

Got pics, not drawings . . . (and I have a Tip-Up)

Captain Avgas said:

Dennis, instead of more words why don't you just post a cross sectional drawing of what you actually did at the canopy-to-side-bow on your RV and open the way for some critical evaluation of the methodology. If you're proposing this detail for others to follow (which you seem to be doing) then surely you wont mind putting your theories in a format (a drawing) that others can understand and comment on.

Click to expand...

No problem! I didn't do drawings, but I do have pictures:
http://www.wideopenwest.com/~glaesers/BondedCanopy.htm
Comment away!

Cheap words and bravado, indeed.

Captain Avgas said:

The same advice goes to others hovering on this thread that have finished a bonded canopy and recommend it to others. It's time to stop hiding behind cheap words and bravado.

Click to expand...

I think a similar challenge went out to you to do some destructive testing with Sikaflex. I'd be very interested to see how it fairs in some pull, shear & twist testing with various applications (done right vs. done wrong).

Move forward. Contribute something meaningful.

DGlaeser said:

No problem! I didn't do drawings, but I do have pictures:
http://www.wideopenwest.com/~glaesers/BondedCanopy.htm
Comment away!

Click to expand...

To Dennis: I can't get the URL you provided to open.

To Sonny: Dude, you're caught in a ground hog time warp and for some inexplicable reason you just can't move beyond the question of short term adhesion. I tested Sikaflex over a year ago and I'm not aware that I've ever suggested it had poor short term adhesion to acrylic. At the moment we're discussing optimum sealant joint geometries. Move forward. Better still, contribute something valuable....post a couple of drawings of your canopy sealant details for discussion.

The last straw

Don Jones said:

I agree with that, automotive windshields have been using the same type of stuff for years to both glue in as well as seal the windshield to the structure of the body.

Click to expand...

It's obvious that I'm getting totally exasperated by the impossibility of there being any meaningful improvement in knowledge on this subject through this thread (or any thread for that matter).

The above quote is the straw that has finally broken the camel's back. Why don't people read all the previous emails before they make comments. I mean the irrelevance of glass car windscreens to acrylic RV canopies was dealt with in my post #41 and then AGAIN in my post #72 and then AGAIN in my post #75. I discussed it so many times I was almost going to apologise for repeating myself.....but no need....within a few days some-one else who has not read the thread properly will make the same uninformed comment about car windscreens. If I had any hair I'd be ripping it out.

And if it's not about car windscreens we'll get another inane comment about the Wright brothers and the spirit of invention. Or we'll get another reminder of the fact that, after all, this IS the EXPERIMENTAL category (as if that was a good excuse for misinformed and dangerous design modifications).

See, I told you I was getting exasperated.

My final advise to those who want to perservere with a Sikaflex canopy is to study up on sealant joint geometries and glazing systems. There are hundreds of books and papers on this topic. It's an absolute science. Give yourself at least a sporting chance of long term success.

In the end there is no "Sikaflex canopy". Rather, there are dozens of different Sikaflex canopies. Each one of them incorporates different design decisions, and most of those decisions are completely intuitive in as much as they are not based on any technical know-how. For the most part the only thing they have in common is that they are using Sikaflex. And in the main it's really amateurish backyard shed stuff....no-one even bothers to prepare a drawing of their details so they have nothing to compare with other builders. Nor do they have anything to refer to professionals for appropriate technical feedback. As a result there is no possible critical examination of methodology between parties.

But anyway, good luck to everyone. I'm definitely outta here this time.

I expect this thread will now quickly die because there is simply not enough technical expertise out there among people who have actually built a Sikaflex canopy to provide the thread with the scientific nurture it needs to survive. But we'll see.

I listened!

Captain Avgas said:

I expect this thread will now quickly die because there is simply not enough technical expertise out there among people who have actually built a Sikaflex canopy to provide the thread with the scientific nurture it needs to survive. But we'll see.

Click to expand...

Bob,
Before the thread dies, I'd like to publicly acknowledge that you helped convince at least one person, (me) to go with the old fashioned method of attachment. I finished my canopy last summer and your valued experience and technical knowledge of the subject was enough, even with the limited posts back then, to make me realize that there was just no sense arguing with physics and that there was no long term experience with the Sikaflex system used in this specific application. It was very tempting to go the "new" route reading of all the cracked canopies. I successfully cut and attached my canopy the old way using info gleaned from a variety of builders on this site without it cracking. It can be done! And yes, I certainly realize that it still may crack in the future.

I hope you can take some solace in the fact that at least one person is listening to what you say. To the rest of you guys who get flustered when dealing with technical subjects and those of us with lesser experience and knowledge, please, please continue to add your contributions when you have safety concerns... Thanks for putting up with us!

Time Warp?

Captain Avgas said:

To Sonny: Dude, you're caught in a ground hog time warp and for some inexplicable reason you just can't move beyond the question of short term adhesion. I tested Sikaflex over a year ago and I'm not aware that I've ever suggested it had poor short term adhesion to acrylic. At the moment we're discussing optimum sealant joint geometries. Move forward. Better still, contribute something valuable....post a couple of drawings of your canopy sealant details for discussion.

Click to expand...

Wow Bob, don't take the high-ground and be polite and professional. All I was doing was throwing your suggestion about doing more than "talk" back over the fence to you. It's clear that you have no intent to do more than provide your limited book knowledge on this subject, but getting pissy about it doesn't help convey your point.

I wasn't aware that this issue had specific topics to be discussed in any particular order, so is it me or are your nasty comments just meant to divert the focus from you having to do something (other than talk) back to someone else? I suggest you move past the argument stage and try to have an intelligent conversation.

i'm speaking for the rv-7a tipup canopy with aluminum and (slightly modded) frame.

there's at least 4 different joint geometries involved. so even if one of them would fail at a time, that thing's not going anywhere yet!
i continue to have a very good comfort level in the installation and if problems pop up with the high time / early installations, there's always time to drill a few boltholes later... to be honest i think, other than making a test article of the exact same situation / application, neither the anti-sealant-book-knowledge nor the forever-ok fractions can claim they're right.
we fortunately don't build our airplanes for a nuclear electromagnetic impulse, nor do we need to worry about cabin pressure, so even if it isn't the ideal engineering solution, the field experience so far seems to be quite good and up to the job. that's all it needs to be.

i will try to make sketches of all situations tonight.

bernie