Lubrication in fasteners and loss of preload

[Juanda]

TL;DR Summary

Can lubrication be a cause for loss of preload because of the reduced friction coefficient?

I'd like to focus the post on threaded joints used in clean room conditions with stainless materials (aluminum, bronze, stainless steel, and titanium) for aerospace applications that need to survive vibration and temperature cycles.

In general, I believe lubrication is always preferable in threaded joints for 4 reasons.

1. It reduces the uncertainty related to the tightening process if preload is controlled with a torque wrench.
2. It reduces the amount of friction present in the joint so the fastener will not be suffering as much torsion.
3. The reduced friction implies less torque is necessary to reach the target preload which can make its application much simpler and safer in certain cases.
4. Galling is less likely to happen.

Uncertainty during preload using torque

Threads damaged by galling

So it'd seem like lubrication, due to the reduction in friction, is always a good addition to threaded joints. However, the thing holding threaded joints together is friction itself. If it wasn't for friction, the screws would unwind returning to their original length.

In some situations, lubricants must be used even if that risks losing some preload which should be addressed with other measurements. Two examples of such cases are:

1. Working with unlibricated stainless steel can be awful. Galling is a true risk that materializes pretty often in my experience. It may happen even before you start applying the preload when the head contacts the surface because it galled already on the way to get there.
2. Helicoils are a common practice in aerospace applications because they allow the use of smaller fasteners in weaker materials such as aluminum. Locking helicoils have a deformed thread which induces greater friction. This makes it very difficult for the fastener to completely unwind itself until it falls out. I'm not sure of its effectiveness in preventing the loss of preload. Here is a perhaps biased video because it's from Nordlock but I still consider it interesting. I started at 2:14 where it shows the results from the nut with the nylon insert which would be the closest to the deformed thread I was talking about. The point is that the manufacturer imposes the use of lubrication if such helicoils are used.
   
   

The way I currently rationalize the use of lubricants in threaded joints is that, even if it might make the loss of preload easier due to the reduction in the friction coefficient, the reduction in uncertainty during the application is worth it because you'll be able to reach a higher preload more consistently which will result in greater friction keeping the joint together even if the friction coefficient is lower. However, that argument falls apart if the application of preload is through other methods independent of friction such as turn-of-nut or measuring the extension of the bolt experimentally (extensiometric gauges, ultrasonic, micrometer, etc).

Is lubrication something to avoid if possible whenever the joints suffer vibrations which could result in the loss of preload?

 

[DaveE]

In high vibration applications you may need additional support like seizing wire, lock washers, lock-tite, etc.

 

[jack action]

Full disclosure: I'm a big fan of lubricated threaded parts.

You already have described the advantages of lubrification. The only disadvantage you describe would be the lower friction possibly leading to the loss of preload. I even heard people stating that being "welded" by rust over time is a good thing: Nature's perfect locking mechanism. But what is the point of having a removable fastener that cannot be [easily] removed?

The preload is there to set the bolt in tension. Once it is done - unless your parts are made of ice - whatever friction you have should be enough to prevail rotation. Even if you have something as low as a 12 lb.in torque setting, it requires a one-pound weight at a one-foot distance or a 12-pound weight at a 1-inch distance to overcome the friction; that is not nothing. Without vibrations, a bolt or nut would never loosen by itself. I mean, threaded fasteners shouldn't be used to prevent rotation between two parts, only to prevent longitudinal displacement.

The reason a bolt loosens under vibrations is that the parts deform so much that the preload becomes almost non-existent. In such a case, the friction coefficient is worthless - no matter how high it is - if there is no normal force. As @DaveE said, if high vibrations are expected, options exist to solidify your joint and they shouldn't be optional. And your joint is easier to set, easier to unset, and more reliable.

However, that argument falls apart if the application of preload is through other methods independent of friction such as turn-of-nut or measuring the extension of the bolt experimentally (extensiometric gauges, ultrasonic, micrometer, etc).

But you forgot about all your other points which are still valid: less torsion, application simpler and safer, and avoiding galling.

 

[Ranger Mike]

on this forum
https://www.physicsforums.com/threa...e-when-using-anti-seize.1061132/#post-7071179
ifin it is a farm bolt, Loctite. Blue or red is your call.
do not use lock washers. my rabid hatred of these.

 

[Juanda]

In high vibration applications you may need additional support like seizing wire, lock washers, lock-tite, etc.

I'll list some questions and concerns with the options you mentioned.

• Seizing wire: I assume it's the same as locking wire or safety wire (non-native English). I don't have much experience with them but my concerns would be:
  • You need an anchor. If there is more than one screw, they can anchor each other.
  • The screws must be above a certain size. Right? I have trouble imagining this on M6 screws (approximately 1/4 inch) and below.
  • The screws must be compatible with the wire.
  • There must be a free "line of sight" between the screw and the anchor for the wire to be placed there or it must be routed through some space where it doesn't collide with other elements.
  • The operator must be properly trained to install it.
• Lock washer: The ones I like are Nordlock (or similar brands that work on the same principle). I love how they work in regards to maintaining the preload against vibrations. Still, my concerns would be:
  • They damage the surfaces they contact.
  • It may generate debris.
  • Only compatible with certain hardenss of materials.
• Adhesives: I love them and hate them. I have had varied experiences with them. Cons would be:
  • Curing time and curing conditions can be an issue.
  • Often outgassing makes a lot of adhesives unusable for space applications.
  • If the thread requires lubrication, it may interact with the adhesive. Unless you consider the adhesive itself the lubricant.
  • Unthreading the joint can be a problem if the adhesive is too strong. Simultaneously, it can't be too weak or it won't work as intended. (I'm aware of the different strengths of adhesives available such as the classification from Loctite divided into 222, 243, etc.)

Of those options, the locking wire I think is the best one. I guess it's often used for a reason.
Parts could be designed with an anchor in place (even if it makes it a little harder / more expensive to manufacture) to be compatible with locking wire in case there isn't a coupled screw to be used as an anchor.
It doesn't damage the surfaces the way that locking washers do it and you don't have to worry about potential problems with adhesives.
The only negative point I can think of is that you could end with a pointy end from the wire. If exposed to radiation, it could be a starting point for electric arcs but it'd be under control as long as the parts are correctly grounded (easier said than done) so there can't be electric build-up over time.

I have some questions about locking wire.

1. Are screws with holes for locking wire defined by some norm? Drilling them myself feels wrong since I could damage them or, at least, invalidate their prior qualification. I have seen them in McMaster for hex heads and Northest Fastener and supply for head heads and socket heads for American pitches. I'm certain they must be defined by some ISO/DIN or even American norm that contains ISO equivalencies. Having them available as DIN 912 and DIN 933 would be great. (DIN 912 and DIN 933 are old norms and there are newer international ISO versions but those codes are still widely used)
2. Same question as above but for nuts. The usual nut in Europe is DIN 934 / ISO 4032. Do people just drill them? It seems even more dangerous than drilling the head of a screw.
3. What's the smallest size you have seen being used? If they are defined by some norm I guess I will be able to see the sizes there but I'm not sure yet if such a norm exists.
4. Countersunk screws don’t seem compatible with this method. But what about counterbore screws? Is there a ready solution for a scenario like that? If not, I believe a large enough counterbore with smoothed edges might work.
5. Is the locking wire qualified too? Does it follow its own norm (ISO/DIN/American norm)?

 

[Juanda]

Full disclosure: I'm a big fan of lubricated threaded parts.

You already have described the advantages of lubrification. The only disadvantage you describe would be the lower friction possibly leading to the loss of preload. I even heard people stating that being "welded" by rust over time is a good thing: Nature's perfect locking mechanism. But what is the point of having a removable fastener that cannot be [easily] removed?

Virtual high five. We're on the same ship here. I heard that quote about rust too and I find it very funny but, as you said, what's the point then?

But you forgot about all your other points which are still valid: less torsion, application simpler and safer, and avoiding galling.

That is true. I have been pondering over this for quite a while now and I think the trees are not letting me see the forest sometimes.

 

[DaveE]

There in lies the art of mechanical engineering. Lots of choices, lots of applications, none is perfect.

Yes, BTW, when I worked on DoD satellite electronics (not fasteners, we had real MEs for that) there were lots of Mil-Specs to know and comply with. Probably even more now. Also lots of people looking over your shoulder that have done it before; lots of history and experience at places like Lockheed. The crude rule of thumb that I've believed in ever since is that all fasteners need a locking device, unless you can explain why not.

 

[Juanda]

on this forum
https://www.physicsforums.com/threa...e-when-using-anti-seize.1061132/#post-7071179
ifin it is a farm bolt, Loctite. Blue or red is your call.
do not use lock washers. my rabid hatred of these.

I actually read that thread before opening this one. I felt it didn't address the same issues I'm listing here.
The thing I haven't read from there yet is the article @jrmichler shared. I have it open in another tab so I don't forget to read it later.

By the way, why the hatred towards lock washers? Which lock washers are you referring to in particular? There are different versions and I might be misunderstanding you.

 

[DaveE]

Which lock washers are you referring to in particular? There are different versions

Yes, it doesn't make sense that way. It's like saying you hate vegetables. Split, Star, Belleville,... there are soooo many.

PS: I guess the farm problem is corrosion? Finding good spring steel that doesn't rust is tough.

 

[Averagesupernova]

Say all you want about a fastener rusting and when the part fails that it is holding it doesn't matter. Of course I'm taking the example from the other thread linked above about a shock or strut. But, if I need to replace something that is bolted to a casting I sure don't want to be drilling out a bolt that is rusted in. If you think this doesn't happen in the real world I have some waterfront property to sell you in Florida. So the solution is to use anti seize during assembly. Most of the time I'm relying on someone else from a number of years back but when I reassemble it is often with anti seize.

 

[Baluncore]

A thread will be self-locking if the friction coefficient is greater than the thread slope = pitch / circumference, but that assumes there is no extreme vibration.

I like to take advantage of the lubricant properties of Loctite & etc.

 

[Juanda]

A thread will be self-locking if the friction coefficient is greater than the thread slope = pitch / circumference, but that assumes there is no extreme vibration.

I like to take advantage of the lubricant properties of Loctite & etc.

I have heard that before. I'll consider using it with locking helicoils.
The problem is that the friction coefficient is unknown in that scenario although it could be found through testing. Or the turn-of-nut method could be used instead.
Later, removing the fastener might be doable as long as the adhesive is not too strong. The remnants between the threads could still work to prevent galling. I'll have to test this approach in the future to confirm the procedure.

 

[Baluncore]

Later, removing the fastener might be doable as long as the adhesive is not too strong.

Use Loctite where, if you want to unlock it later, you can apply local heat to the nut or shank. 150°C is sufficient to soften Loctite or cyanoacrylate.

Use anti-seize with dissimilar metals, where threads may corrode, or where threads go deep into a casting.

 

[Juanda]

Use Loctite where, if you want to unlock it later, you can apply local heat to the nut or shank. 150°C is sufficient to soften Loctite or cyanoacrylate.

I remember doing that a long time ago. It's especially useful in aluminum parts with stainless steel fasteners because of the different expansion coefficients although my experience is only anecdotical so maybe I was just especially relieved the screw finally came out.

 

[renormalize]

A thread will be self-locking if...

...(fill in the mechanism).
PF with this feature would make the mentors' job so much easier.