Failure of a bonded joint that has a thread 3

[elinah34]

Hello,
I have a question about a thread which exists in a bonded joint.
I am adding a simplified picture of such a joint, that we have some issues with:

Please pay attention that the thread goes a little bit over the second lower bonded Ultem part.
After manufacturing this joint (Including inserting the Tangles Helicoil, a special Helicoil for Plastics which doesn't require exerting any force for inserting it to its place, but only threading it to its place) everything was fine, and there was no evidence of separation in the interface of both Ultem parts while examining it visually and ultrasonically.
After tightening the bolt exactly as described in the picture above, exerting 10 Lb*Inch, we found out that it caused a separation between the 2 Ultem parts (the Adhesive didn't succeed holding them together).
We can't understand the root cause of this failure because the tightening force should attach the parts to each other.
We thought about a potential failure mechanism, but we aren't sure about it – the external load is never uniformly distributed along the thread's teeth (see an example below), and maybe this may cause a relative movement?

 

[TugboatEng]

Ultem 1000 has a very high coefficient or thermal expansion. If the adhesive doesn't match, temperature changes will load the adhesive in shear.

 

[Stress_Eng]

Can you supply a picture of the joint condition you actually have? You say the thread goes into the 2nd layer. Is it the helicoil threads that go through to the 2nd layer? Does the bolt protrude past the end of the helicoil, or does the bolt threads stay within the 1st layer and the helicoil goes through to the 2nd? A picture would explain the situation. How thick is the adhesive layer? What thread type do you have, and in particular what is the pitch (pitch relative to adhesive thickness).
One other thing that comes to mind. The 1st layer will be compressed under preload. As you go radially outwards from the thread root, the 1st layer will progressively return to its original thickness. Will this change in adhesive layer thickness cause an issue?

 

[elinah34]

Hey TugboatEng
The adhesive has a quite close CTE to Ultem 1000.
In addition the joint wasn't thermally loaded but was left in a room temperature.

 

[elinah34]

Hey Stress_Eng
I will take a picture of it and upload it.
The helicoil engage the pre-machined thread in the second (lower) plate, but as seen in the picture above - the engagement is minor and is about ~ 1mm.
As seen in the picture, the helicoil engage a little bit (~1mm) the second (lower) plate and the bolt's threads don't go beyond the helicoil.
The adhesive layer thickness is 0.3 mm.
The thread type is also indicated in the picture and it's 0.138"-32.
I think you mentioned an interesting thing I didn't think of, though I can't imagine that's the root cause of the failure. When the 1st plate is compressed in the axial direction the hole tends to expand in the radial direction and this may lead to a stresses in the interface between the 1st upper plate and the adhesive layer.

I am adding the picture from the CAD again:

 

[TugboatEng]

You said the thread is pre-machined in the bottom plate. How are you ensuring that the bottom plate thread is in time with the top plate thread? As you tighten the bolt it's possible that mistiming or pitch mismatch caused by adhesive thickness is causing the end of the bolt to push on the last thread.

Is there a reason the helicoil extends below? They can be cut with wire cutters if they're too long.

Otherwise, consider chasing the threads with a tap after bonding.

 

[Stick.]

Having the helicoil extend into the bottom piece by basically a single thread just doesn't look right and makes me wonder if it has anything to do with your issues. I'd be concerned that rather than that single thread holding anything together, it may just serve to push the joint apart. Even if just as an experiment, I'd try cutting the helicoil (and the bolt if needed) shorter so that nothing threaded extends into or engages with the bottom plate.

 

[IRstuff]

Seems to me that the 1-mm engagement of the helicoil on the bottom layer is problematic. Tightening of the screw on the helicoil results in an upward force of the helicoil against the bottom layer's top; I think this results in pulling the material around the thread hole upwards, bulging the Ultem, forcing the two layers apart.

Additionally, the helicoil will actually prevent the two layers from being pulled together; it's one of the reasons why decking screws often have no threads near the head, since that allows the threads that engage the bottom layer to pull the head against the top layer, forcing the layers together.

I would suggest trying an experiment where the threads on the top layer are removed altogether, putting the helicoil only on the bottom layer and see what happens.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[elinah34]

Hey TugboatEng,
The pre threaded hole was drilled through the 3 layers - Ultem/Adhesive/Ultem, and threading it was done at once, so I can't think why there might be misalignment between the upper/lower theead.

 

[elinah34]

Hey Stick,
You wrote "I'd be concerned that rather than that single thread holding anything together, it may just serve to push the joint apart."
In which scenario (and why) the single thread might push the parts apart?

 

[Stress_Eng]

A bit of info on the thread joint, is the adhesive layer above the last bolt thread, and therefore within the thread engagement length, or below and outside the thread engagement?

 

[3DDave]

Since the bolt doesn't join the two Ultem pieces why would there be any assumption it would hold them together?

It looks like a large ratio between the stiffness of the bolt and the stiffness of the Ultem is leading to distortion that doesn't normally happen with bolts inserted into metal.

Steel modulus of elasticity 30,000 ksi
Ultem modulus of elasticity 519 ksi


That's 50:1; the Ultem cannot create a non-uniform load distribution. Every bit of the helical coil insert in contact with the bolt participates equally.

So this is compressing and spreading the Ultem uniformly along the entire length of the bolt, and that means a large amount of shear strain at the glue line. Unless the glue line is very thick and the bond sufficient to resist the shear, it's going to detach. This same stress will apply if there is a second layer or not. The second layer converts the strain into a stress.

Try this with a single sheet of Ultem tapped and with the helical coil insert installed and with the bolt tightened. See if the Ultem opposite the fastener head stays flat and, if possible, put a linear strain gauge radial and near the hole to see how much deformation there is.

I'd also look at the strain resistance of various thickness of the adhesive for the amount of strain measured with the strain gauge.

Also, if this is a locking helical coil insert, the stress might be concentrated as the gaps between the screw thread and helical coil are eliminated to make that locking feature. That would take some FEA work to understand quantitatively.

 

[Stress_Eng]

Yes, the glue is below and outside the thread engagement. Looks like you need to determine joint displacements / strains under preload, be that by test, FEA or calculation. The helicoil will make modelling more involved. That last helicoil thread is going to be restrained by the lower plate thread, while the rest of the coil threads will be within the compression and radial loading fields. Sounds like an interesting task!

 

[Brian Malone]

elinah34 I am missing how the two Ultem parts are fastened together by anything other than the adhesive. You stated the Helicoil's last coil slightly engages the lower Ultem piece. But if the lower piece has not been explicitly tapped with the Helicoil tap then the insert cannot engage this item. It will merely interfere with it unless you have what appears to be an edge break on your lower bore which is purely a clearance pocket and should be greatly larger in diameter than the Helicoil OD to insure there is no interference between parts that will create a locked in stress. From what I see your Ultem parts are glued together, and the Helicoil insert is for providing secure threads for mounting your thermistor. The thermistor screw does not create clamp force between the Ultem pieces and neither should the Helicoil. If you are experiencing separation then the screw must be contacting material below the lower surface of the upper Ultem piece and jacking the two pieces apart as it is tightened. You should make the clearance pocket in the lower piece larger in OD - regardless of it being as you have described match-machined.

 

[MintJulep]

There are 6-32 Heli-Coils available that are only half the length of your current choice.

That might solve the problem.

This is of course the problem solving technique of "try stuff and see if it works", which is generally less preferred than actually figuring out the problem first.

 

[Brian Malone]

elinah34 I just saw your reply to TugboatEng - so, your two Ultem pieces are adhesively bonded, drilled, then threaded for the helical wire insert to prepare for insert installation, and have the inserts installed. Are you fully cleaning the chips from the threading op fully? Are chips dead-heading in the lower pocket and getting compressed by the thermistor screw?

 

[LittleInch]

Can't you just use a slightly smaller bolt or add a washer under the bolt head??

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[elinah34]

Hey 3DDave (Aerospace)
I agree that the threaded adhesive may fail due to shear, but I am not sure why it would spread radially.
If it was a local teeth shear of the adhesive, I guess I wouldn't clearly see a separation.

 

[elinah34]

Hey Brian Malone (Industrial)
The lower piece was almost entirely tapped with the Helicoil tap and hence the insert can engage it.
Indeed, as you mentioned - the Helicoil insert is for providing secure threads for mounting your thermistor. The thermistor screw does not create clamp force between the Ultem pieces and neither should the Helicoil.

 

[elinah34]

Thanks MintJulep (Mechanical)
I know this Helicoil has other available lengths - 1D and 1.5D (I used 2D length).
I guess that taking shorter Helicoil is what I will finally do, but I really want to understand what caused this failure .