Liquid Shim Process 1

[JCorsico]

Hello all.

We are working on a design where a carbon composite panel is fastened to a 4130 steel subframe. The subframe has flat mounting flanges for the carbon panel. The carbon will be attached to the mounting flanges using both an epoxy paste adhesive and Hi-Lok pins. Basically, the design is very similar to how a wing skin is attached to the stringers.

The mounting flanges are not perfectly flat (there is minor weld distortion). The carbon panels are also not perfectly flat (manufacturing tolerances). The largest gaps are about 0.030", but most gaps are 0.010" or less. We want to fill these gaps using a liquid shim paste adhesive.

What is the recommended process for installing the liquid shims? Some questions that we have:

1) Can we install the shim paste and the Hi-Loks at the same time, using the Hi-Loks as clamps? In other words, will our epoxy paste adhesive function as a shim? Or, do we need an entirely separate process to create and cure the shims first? And then after the shims are cured, then install the Hi-Loks and more paste adhesive?

2) If we need a separate process to create and cure the shims first, do we use mold release on the carbon panel? Or a release ply? How do we prevent the shim from adhering to the carbon panel?

3) If we need a separate process to create and cure the shims first, how do we clamp the carbon panels in place? Clecos? Does the shim material get all over the Clecos and glue them in place? Do we coat the Clecos in something so the shim material doesn't stick?

Thank you!
Jon

 

[rb1957]

I don't know. It is possible that there are no gaps, that stringers and skins are tooled into position.

But how is this relevant to your question ? You can use liquid shim to fill voids but you can't count on it as a structural adhesive.
This would be similar to counting fay-surface sealant as structural adhesive. Sure it's there, sure it does carry some load, but we don't count on it.

another day in paradise, or is paradise one day closer ?

 

[Ng2020]

If you have a thick adhesive layer it will probably have very limited capability due to excessive peel stresses.

Also I'd design for limit load with a failed bondline, refardless of acutal shim/bond strength, as it's likely to be an unreliable feature.

Both these considerations will drive the number of fasteners required.
Have fun.

 

[SWComposites]

Boeing does not bond metal stringers to metal skin. They may use liquid shim in the mechanically fastened joints, but that shim is not a “bond”. And liquid shim is used between lots of composite and metal parts, to fill gaps so the fasteners do not pull up the gaps and cause problems (like Boeing apparently forgot about leading to no 787 deliveries for over a year now. Cripe). And again this liquid shimming is NOT to bond the joint together.

Further, the 787 wing and fuselage panels have cocured or cobonded stiffeners, so no liquid shim is needed between skin and stiffeners. Liquid shim is used between fuselage frames and skin panels, and between wing ribs/spars and skin panels.

And above I outlined how to apply liquid shim to a joint.

 

[blakmax]

One reason that bonded joints on commercial airliners use both adhesive and fasteners is simple: The FAA has a policy that any bonded joint, the failure of which would cause a risk to flight safety, must have an alternative load path. There is however an example of where the regulators do not enforce that guidance. Helicopter rotor blades are frequently bonded together without additional fasteners.

I agree with Ng2020, design as if the bond had failed. In the example I gave before about the bonded and bolted designs, I forgot to mention that in the bonded and bolted example, the fasteners carried about 400lbs (not a typo four hundred pounds). The fasteners will only carry load when the bond fails.

Regards

 

[WKTaylor]

Blakmax... be very cautious about load transfer... a simplified explanation for a subject with many elements of complexity.

Stiffness plays a huge part every time. IF the joint is designed for metallic fasteners and paste-adhesive 'shimming', You can count on the fact that THE fasteners will pick-up load first. The paste adhesive [liquid-shim] will ensure that there is little/no flexibility/looseness between structural parts due to gaps, ensuring shear continuity/stiffness for the fasteners. however... Epoxy in a fay surface has a far lower modulus of elasticity than tight metallic fasteners... hence epoxy will see a far-lower load-share in all stress states in your joints.

NOTE. This same 'stiffness-rules' clearly applies to metallic fasteners of various alloys. For the same tensile/shear strength, alloy steel and CRES/HRA have = modulus... whereas titanium has a modulus ~66%E of steel/CRES/HRA... and [puny] aluminum is ~33%E of steel/CRES/HRA [at much-lower tensile/shear strength]. Yes... steel/CRES/HRA will 'load-up' at a higher-rate than the titanium and a whole bunch faster than aluminum [rivets], due to substantially higher stiffness. Be very careful mixing fastener alloys.

NOTE. It's late. I hope this still makes sense in the morning-light after coffee.

Regards, Wil Taylor
o Trust - But Verify!
o We believe to be true what we prefer to be true. [Unknown]
o For those who believe, no proof is required; for those who cannot believe, no proof is possible. [variation,Stuart Chase]
o Unfortunately, in science what You 'believe' is irrelevant. ["Orion", Homebuiltairplanes.com forum]

 

[rb1957]

@blakmax ... that is not correct (from my experience). I Know the DHC8 has metal-metal bonding of panels (bonding two plies of skin together) and bonding stringers to skins (without rivets, though some have rivets as well). But, yes, metal-metal bonding is tricky and Very process dependent.

another day in paradise, or is paradise one day closer ?

 

[WKTaylor]

BTW... BE VERY CAUTIOUS about [at least] 2-other aspects/elements of Your design.

You MUST ensure electrical bonding/grounding paths between dissimilar skins, fasteners and substructure for arcing/sparking, corona, EMI and lightning protection. This is a deep/important aspect worthy of intensive discussion, by itself.

Also

Epoxy adhesives are subject to deterioration by exposure to aviation fuels-oils-fluids and moisture: the epoxy MUST be protected from exposure to fuel-oil-fluid-moisture by overcoating with elastomer sealants.

NOTE1. I've worked F-106s and F-16s in a past life. Both are CONVAIR/GD designs and both used fastening and adhesive-sealants for wings and body integral-fuel-tank sealing. HOWEVER the adhesive system(s) used is/were NOT epoxy based. In these designs, GD always relied on the fasteners: the adhesive was always considered as 'shim' and 'fuel/pressure sealing' media that was 100% fuel resistant, which was intended to improve the metallic-structure efficiency [= fatigue-durability and toughness] over conventional assembly with elastomeric sealants. This sealing method is very 'proprietary-process intensive' and not for amateurs. Conventional elastomer sealing is all-but avoided by this assembly method... but is still necessary in obscure situations where adhesive cannot/does-not fill-between precisely fitted details.

Regards, Wil Taylor
o Trust - But Verify!
o We believe to be true what we prefer to be true. [Unknown]
o For those who believe, no proof is required; for those who cannot believe, no proof is possible. [variation,Stuart Chase]
o Unfortunately, in science what You 'believe' is irrelevant. ["Orion", Homebuiltairplanes.com forum]

 

[Ng2020]

Fairly sure the dhc-8 skin / waffle doublers are hot bonded in a controlled environment (autoclave?), id expect there was confirmation of no voids via ultrasonic ndi as part of the build process. Very much nore a controlled process than epoxy paste & clecos on the shop floor. Other OEM's have used similar approaches in lieu of chem milling skins.

 

[rb1957]

yes they are ...well, were. i was just responding to the comment that all joints need fasteners (that the FAA says you can't rely on metal-metal bonding).



another day in paradise, or is paradise one day closer ?

 

[blakmax]

Hi Rb1957

The difference is that the FAA rule applies to joints in a critical load path in principal structural elements. If failure of a bond to a stiffener does not risk structural failure, then the FAA rule is not I believe applicable. Still love to know why that rule is waved for helicopter blades.

WKT,

I understand the issue of stiffness well, and I suggest that thick adhesive joints are far more compliant than the thin bondlines in the example I referred to. Like you, I've worked on GD designs (F-111) and I must say that the procedures used for sealing the wings was appalling. RAAF went through an extensive wing and fuselage deseal-reseal program using very nasty chemicals that cost the lives of a number of servicemen and resulted in multiple compensation pay-outs, even though the original manufacture sealed the joints with a HT resistant sealant. Almost all bonding issues relate to hydration of the surface oxides to which the adhesive is bonded. Stopping that hydration is the key to stopping bond failure. Naturally, another key factor is the design of the load transfer in the joint. Unfortunately, many people still use the average shear stress design methodology, despite the fact that Volkersen showed that was deficient as far back as 1936. Even worse, it is the basis for the methodology embodied in Advisory Circular AC 20-107B. You need to do extensive testing of every design variable (thickness, overlap, dissimilar materials, cure temperature and service temperature and then select an allowable average shear value such that failure does not occur in any of your tests, then use that value for design. This may involve thousands of tests.

I prefer the Hart-Smith Load Capacity method which actually uses realistic adhesive design properties and the design methodology takes into account all of the significant design variables, so that you can actually calculate a realistic value of potential bond strength. Then you design the overlap to provide that load capacity such that the joint should never fail PROVIDED that the processing is valid. Now this design method has the potential to eliminate hundreds of wasted tests performed to manage the average shear method. If I can demonstrate that the bond will never fail at 1.5 Limit Load, then there is no load case that will cause failure of the joint by design. Every test will result in fracture of the adherend. So why undertake thousands of tests that only fail the adherends and tell you the strength of the adherends which I can calculate from basic material data.

Now whether that approach is valid for very thick adhesives is open to debate.

I agree about the importance of the other issues you raised.

Regards

Blakmax

 

[rb1957]

ok, sure, but the point is ... liquid shim is not an adhesive.

another day in paradise, or is paradise one day closer ?

 

[Sparweb]

Okay, so how do you adhere two components that don't fit perfectly together?
For example, I cannot imagine that Boeing and Airbus are manufacturing wing skin / stringer assemblies that have 0.005" or less gaps across the entire wing surface. There must be gaps. How are they filling the gaps before bonding?

Maybe there is a seed of confusion in this subject, too.
Many structural assemblies have a variety of "fit-up" issues that need to be resolved, often at the factory level because they are hard to predict at the design level. So there are many standard practices and "rules of thumb" related to what's an acceptable gap in a joint versus what needs to be shimmed. Often very context dependent. IIRC you can find methods and advice in Bruhn, Niu, and Flabel's texbooks, and plenty of others. Those sources address only fastened joints, of course. For bonded joints there are more recent sources (Hart-Smith, I believe). Some OEM's also specify acceptable gaps and shimming techniques that they find acceptable in various cases. If the OEM includes shimming with a liquid-shim product (I can think of a couple of examples) then you can tell that those are joints the OEM has found tolerant of misalignment. Up to a limit.

If you do find liquid-shimming in an OEM's structural repair manual, applicable to joints with fasteners, that does not mean that the joints where it can be used have any extra benefit from the liquid-shim as a bonding agent. Its only function is to support the fasteners and bearing surfaces of the components well enough that the joint functions the way is should.

I believe you will never find liquid-shimming in bonded joints in an OEM's structural repair manual. I can't make sense of that. If the joint has a substantial gap/mismatch needing a shim, there are other methods to address it. You can bond filler materials into the joint - but sometimes that would be very tricky. Setting aside the details which would vary widely from joint to joint, you should usually solve gaps in bonded joints with bonded solutions.

I hope that clears things up a little.

 

[JCorsico]

I believe you will never find liquid-shimming in bonded joints in an OEM's structural repair manual. I can't make sense of that. If the joint has a substantial gap/mismatch needing a shim, there are other methods to address it. You can bond filler materials into the joint - but sometimes that would be very tricky. Setting aside the details which would vary widely from joint to joint, you should usually solve gaps in bonded joints with bonded solutions.

Bingo! There must be some way to bond components that don't have a perfect fit up. That's the crux of my question.

Thank you all again!

 

[SWComposites]

JC - why do you keep talking about “bonding” components? Liquid shim is not a bond. Liquid shim is used to fill gaps in bolted joints.

 

[rb1957]

I think that's his question ... if you have a structural adhesive bonding say stringers to skin, how do you treat gaps ?

Do you fill the void with liquid shim and then bond ? I don't think so, as this would put the liquid shim on the loadpath between the skin, the adhesive, and the stringer. But then I don't know what Boeing do for this specifically.

Do you have mechanical fasteners as the primary joint load transfer ? There are still issues with the shim (increasing the bending).

I'm confused by why the OP is asking about Boeing bonding composite primary structures and how this relates to his question.

From the original post, I think option 3 is the correct approach. As expressed in other posts above, yes use release agent on the carbon panel and the clecos (if concerned, never heard of a problem removing the clecos).

Note that Boeing's process for their build is very different to what you're doing (or what we think you're doing); much more controlled, much more testing and validation.

another day in paradise, or is paradise one day closer ?

 

[JCorsico]

JC - why do you keep talking about “bonding” components? Liquid shim is not a bond. Liquid shim is used to fill gaps in bolted joints.

I think that's his question ... if you have a structural adhesive bonding say stringers to skin, how do you treat gaps?

Correct. How do you bond two parts that don't have a perfect fit?

Also, I don't understand why this discussion is making such a big distinction between a liquid shim and a bond. In many cases, the same adhesive is designed for use as a shim and for use as a bond. In fact, the adhesive we spec'd is designed for that dual use.

From the original post, I think option 3 is the correct approach.

Got it.

 

[SWComposites]

Because liquid shim is NOT intended to bond, only fill the gaps in a bolted joint. It doesn’t matter what the supplier calls the stuff.

If your joint does not have fasteners, and you are attempting to bond two solid (cured) materials, the gap filling will only be one of your many problems.

But you originally stated that you are joining a cured composite part to a steel part using HiLoks. Therefore any “adhesive” used in the joint should be considered a liquid shim, and should not be bonded to one of the parts.

Boeing does not bond two cured composite parts on the fuselages and wings. Either the stiffeners are cocured with the skin, or the precured stiffener is cobonded to the uncured skin in the skin cure cycle.

 

[rb1957]

"How do you bond two parts that don't have a perfect fit?" ... not your original problem statement. "How do you fill the gap between parts that don't fit up exactly ?" ... Liquid shim is an answer to this question.

"I don't understand why this discussion is making such a big distinction between a liquid shim and a bond." ... and therein lies the problem.

we (I?) thought you're off-base when you started talking about the shear strength of liquid shim. Maybe you said it (maybe we (I?) thought it) but I got the idea that you were trying to use the liquid shim as an adhesive, and include the shear strength in the joint's load transfer capability.

The primary differences between structural adhesives and liquid shim are ...
1) thickness ... structural adhesives are very thing and liquid shim typically isn't (0.02" isn't "very thin" ... < 0.005" is).
2) processing ... liquid shim is typically room temperature and structural adhesives usually aren't. Yes, there are room temperature adhesives, but see the next comment. Structural adhesives usually have involved and highly controlled processing.
3) testing and supporting data ... structural adhesives will have a lot of testing to validate their strength.

another day in paradise, or is paradise one day closer ?

 

[WKTaylor]

JC... my 2-cents... just to be clear about why I asked You to define Your epoxy paste material being used as liquid shim.

The most critical elements for using epoxy paste adhesive used as 'liquid shim' is to verify fillers and compression strength and environmental resistance to ensure it is up to the task, in-position, for the long-term.

The second-most critical element is to ensure that 'working life' of the mixed-paste epoxy is long-enough to delay cure-to a semi-solid material [in-transition to a solid plastic state] so that structure assembly can be completed as intended. In the case of 'small jobs', using Room-Temp cure materials are typically acceptable. For large/major Assy jobs... requiring hours-long working-life... an elevated temperature cure paste adhesive is likely mandatory.

In all cases, use of 'temporary fasteners' to bring the structure into tight alignment for shim-cure [after squeeze-out/filleting/wiping-away-excess/etc of the 'wet-paste']... may be essential. THEN before epoxy cure [oven?], replace temporary fasteners with permanent fasteners... otherwise delay installation of permanent fasteners until after epoxy cure.

NOTE1.
I have directed re-assembly of large elements of structure, using elastomeric sealant in the fay surfaces. In many cases I've mandated C24 or C48 long-cure sealants. In these cases, I recommend that the shops have a 'witness board' where they place a [generous] sample 'daub' of sealant from each sealant batch [label it to the batch!]. They then test the sealant about every 1/2-hour by stirring it with a clean wooden spatula. At some point, the wet-sealant will suddenly stiffen and become harder to stir-back-into-itself... a clear indicator that the sealant is at the end of its 'working life'. This is a system that automatically validates the mixture for individual 'batches' [A+B variations] and the local ambient temperature... and gives the shop 'fair warning' that no-further wet-use is possible... it is in final stage of transitioning to 'elastomer'.

Regards, Wil Taylor
o Trust - But Verify!
o We believe to be true what we prefer to be true. [Unknown]
o For those who believe, no proof is required; for those who cannot believe, no proof is possible. [variation,Stuart Chase]
o Unfortunately, in science what You 'believe' is irrelevant. ["Orion", Homebuiltairplanes.com forum]

 

[JCorsico]

Thank you all, this has been a very helpful discussion

Jon