Torque Audit on Fasteners with Loctite 1

[BATman2]

I have a situation that requires a supplier to "check" if a large (60mm) nut has been "tightened". Ideally, I would like to confirm installation torque (400 N-m), or at least something very close to that. The nut has Loctite 243 applied and is in a very tight space. I am concerned that damage can occur to the Loctite if I approach the max. torque during an audit. Short of removing and re-installing, how can I check the torque without damaging the Loctite? If no rotation occurs, I would expect the Loctite to remain intact and retain it's full strength, but I'm not so sure. (I recognize pre-load is what counts but I can not measure that at this point. I also recognize the the proposed torque audit process is imperfect).

Thanks
BATman2

 

[dik]

You might want to check with the Locktite rep...

 

[BATman2]

Dik,

In process.

Thanks,
BATman2

 

[MintJulep]

We've discussed torque audits here before - thread725-193606 - with general agreement that they provide no useful engineering information. And that was for joints that rely solely on friction.

With cured loctite involved, I think it is impossible to derive any information at all from a torque audit of the completed joint.

The cured loctite will produce some value of break-away torque required to shear the bond. This will be true even if there is zero preload in the joint.

So any audit torque that is less than the break-away torque can at best test the strength of the loctite.

It is possible that there could be some value of audit torque that would not overcome the combined effects of the loctite bond and the friction in the threads, yet will produce motion of the nut and bolt (as a single unit) with respect to the joint members. There are so many variables involved, and so much potential variation in each, that the uncertainty would be huge.

If the nut moves with respect to the bolt, then you have broken the loctite bond and destroyed the design integrity of the joint.

I have no idea what loctite's stress-strain curve looks like. From hands-on experience, when you exceed the break-away torque typically get an audible "snap" as the bond shears, and then the joint is loose. I don't know what happens to loctite as you approach it's ultimate strength. Loctite is essentially a plastic, so I would guess that there is a region of plastic deformation. Loctite tech rep might be of some help here (but I doubt it).

Basically, I see no way to do any sort of meaningful test on a joint with cured loctite.

If you suspect that the joint may not have been tightened I think the only thing to do is take it apart and put it back together properly.

 

[Screwman]

I agree 100% with Julep. Trying to perform any type of torque audit on a joint with cured thread adhesive is a futile exercise. You will find no correlation between the inspection torque and either the seating torque or the clamp load.
On these joints you must document the installation and then leave them alone.

 

[BATman2]

Screwman,

I fully agree with respect finding the actual torque. I'm basically looking a method to check for untighten nuts that have Loctite without damaging a properly installed joint. At this point, the supplier is experimenting as follows:

1) Measure torque required to break Loctite only (hand tightened). This should be lower then max. torque applied during the "audit".

2) Measure the torque needed to break a properly installed nut with Loctite. Do this for both an undisturbed joint and a joint that has been "audited". If both are the same within reason (10% let's say), I have some assurance an audit will not damage the Loctite.

During the "audit" a line will be scribed in to the nut and bolt to confirm rotation did not occur.

With the above, the only conclusion gained from the audit is that there are no joints being held by loctite only (hundreds need to be checked). We have reason to believe that if the nut has a torque greater than "loctite only", then the required torque was applied.

Clearly, the actual torque on the nut can not be determined. Full disassembly is clearly the best approach, but this is not an easy option to apply at this time.

Thanks very much,
BATman2