Torque Transmission Through Friction Coupling Assisted with Retaining Compound

[JCReynolds79]

Hi all,

I have a tube, which has a bolt down the middle, which when tightened, pulls the end face of the tube against another face. This tube then has to transmit torque, via friction alone. I can workout the possible torque transmitted by firction, but what I am trying to figure out whether adding retaining compound (e,g. Loctite 648) between the mating faces of two abuting surfaces will actually help transmit more torque or not.

given the shear strength of the retaining compound, and the known axial load, can I calculate the torque capacity of the joint with the addition of the retaining compound? Is it additive or is it a case of whichever is strongest (friction or retaining compound shear strength) is the max value?

I think this is the formula for the shear strength of the joint, but I don't know whether the axial load needs to come into it or not...

T = t*PI(ro4 - Ri4)
2ro

T = torque; t = shear stress; ro = outer radius, ri = inner radius

Thanks in advance


Regards,

Jon Reynolds

 

[desertfox]

Hi

The formula you have quoted is the the standard shear stress equation for a tube in torsion, you need the preload force of the bolt between the two faces its clamping, whatever this preload force is, you then mulitply it by a friction coefficient say 0.2, from that you can then work out how much torque you can apply to the tube before the faces slide.
what happens if the torque is applied in a direction tending to undo the screw?

 

[Nereth1]

My guess would be the retaining compound shear strength gives you your new value until such a time as it breaks, then you get the worst of friction between it and itself, or its bonding strength to whatever your tube is, and shortly thereafter when the retaining compound gets rubbed off, it will be the friction of tube to tube.

 

[JCReynolds79]

desertfox, I work out the torque transmission through torque alone with a different formula, based on the area of the contact annulus and normal force and coef.fric. The second formula, for shear stress of tube...I am unsure if that is the right formula to use to determine the strength of the bonded joint. i.e. how much torque could the two tube glued together transmit on their own before the bond breaks (ref. the shear strength of the 648) I see this as independant (at first) from the axial load.

So I can calc frictional torque and I can calc shear torque of the bond...but how do I bring it altogether when I actually bolt the joint together with the 648 in place? Do I get something better than the frictional torque or something different?

Nereth1, I think you are right in a way, but unless I calculated it wrong, the shear torque of the compound alone was rather less than the frictional torque. So does it contribute nothing?

Regards,

Jon Reynolds

 

[Nereth1]

If it is very thin or your surface finish is very poor then it may help just fill the gaps where previously no frictional force was generated, otherwise I don't think it can 'contribute' to the existing frictional forces. It is a barrier between the contacting parts and unless it has a higher coefficient of friction than them then it probably won't do much.

If you are using an adhesive it sounds like you are going for a permanent or semi permanent bond. It would be exceptionally easy to serrate these parts so they key together, or there are a lot of other ways to achieve the same. Alternately are there frictional compounds or coatings that can be used that can actually leverage the very high contact forces you will likely be able to generate with your threads?

 

[JCReynolds79]

Unless the bond shear strength is stronger than the pure friction joint, then I agree there seems to be no point adding it. But and I can't stop thinking about this, if I did add the adhesive, my pure friction calc would surely change, as the assumed coef.fric., which was for metal-on-metal, is now no longer metal-on-metal but something like metal-on-adhesive-on-metal. But even this doesn't sit right because unless the bond breaks, it is not a friction joint, it is a bonded joint...

Regards,

Jon Reynolds

 

[Nereth1]

Also keep in mind that when it breaks and the adhesive layer abrades away, depending on the stiffness of the setup, you could lose some preload and thus some friction unless you retighten it.

 

[BrianE22]

When sharing loads you look at relative stiffness. Friction joints are very stiff. The glue would have a thickness and to develop a shear load on it would require shear deflection. My guess - the glue contributes very little.

 

[desertfox]

Hi

The friction between the abutting faces is sliding friction and should be based on 0.2 * the bolt preload! that gives the frictional force needed to be overcome in order to get the surfaces sliding over one or other.
Once you have that friction force, I would assume the friction force acts at the mid radius of the two faces and use that mid radius to calculate my limiting friction torque.

Hope this makes sense

Desertfox

 

[JCReynolds79]

Desertfox, I think you're missing my point. I already said in my original and subsequent posts that I can calculate the friction torque.

What I am questioning is whether adding a retaining compound (an adhesive) on that joint face would have a positive or negative impact on possible torque transmission.

Regards,

Jon Reynolds

 

[MintJulep]

Glue thickness = 0 --> Friction

Glue thickness = "very thick" --> glue

Glue thickness = "thick" --> glue

Glue thickness = "thin" --> glue

Glue thickness = "very thin" --> glue

 

[desertfox]

Hi

Sorry I did miss your point however it's possibly because you mention using the surface area of contact, normally unless the pressure is very large, friction is independent of contact area and I don't see any sizes given.
Also in my first post I asked how you would prevent the screw from loosening if the torque was applied in an anti-clockwise direction?
As regards the compound being placed on the mating surfaces I would say not a good idea because you would need a gap between those surfaces which means your bolt isn't pulling the faces tight and I doubt whether the bond between the two surfaces will be as good as the frictional force as you yourself mention.
I would however be tempted however to put locking compound on your bolt or screw just prior to tightening.

 

[tbuelna]

The anaerobic compound does not actually increase the capacity of a clamped interface to transfer torque from the shear strength of the adhesive bond. Instead, what the anaerobic compound does is ensure a high and consistent static coefficient of friction will always exist at the clamped interface. And this permits a smaller factor of safety to be used when analyzing the torque transfer capability of the connection. Friction characteristics can vary greatly, and so any analysis based on friction must use very conservative factors to ensure the joint will never slip. The use of an anaerobic compound at a friction interface might reduce the analysis friction factor required by 50% or more.

 

[JCReynolds79]

Hi,

Thanks all for replies.

desertfox, I agree, definitely will be putting thread lock on the thread. This seems to add a lot to the breakaway torque.

tbnelna, this sounds more like what I think might occur. I do not envisage an actual thickness of compound separating the two metal surfaces, especially as it is clamped up. I envisage it filling all the potential voids, effectively increasing the coefficient of friction of the joint from the basic metal on metal approximation.

Do you have any data/references to help me estimate this increase in coef.fric.?

Is there any danger of the compound breaking down leading to loss of preload?

Thanks.

Regards,

Jon Reynolds

 

[Nereth1]

If it's in the voids only then as mentioned earlier it's stiffness is probably far too low to do a lot of work until after slipping begins, at which point it will be most likely too weak and will break in shear. As mentioned, it's main point :/ that situation would be a failsafe against Loss of preload, but note the potential it has to act as a lubricant once broken (which it may or may not so but could be worth checking).

 

[Frankie57]

Hi
Is it possible to place a annular groove in the mounting face about .002" deep? This would ensure a bonded layer which can be calculated. The remaining metal to metal contact (assuming the glue is squeezed out) could also give a calculated friction torque.

 

[JCReynolds79]

Well I spoke to a rep from Henkel/Loctite today. They say that the adhesive will fill all the voids due to no surface being perfectly flat, but if clamped, you'll still get the same metal to metal contact (all the peaks basically) but with added shear strength from the compound in all the gaps.

So, one can ADD the shear strength of the face bonded joint to the frictional strength of the metal on metal clamped joint.

This has been very interesting. Thanks for the input guys.

Regards,

Jon Reynolds

 

[desertfox]

Hi

If it fills all small voids due to faces not being flat how will you know how much adhesive you have in the joint, in addition if they had said you cannot rely on adding the adhesive would you still buy some?

 

[JCReynolds79]

Ok, in more detail, he said that a percentage of the shear strength should be used. They have many factors to help decide how much of the total available shear strength to use, such as joint type, temperature, surface finish, etc.

He also said, which is intuitive, that the larger the preload, if you are clamping the faces together, the less of a role the compound will play, as obviously the clamped friction capacity will increase with preload whereas the shear strength does not change. So it's a case of deciding whether you get any good returns or not.

In my case the joint is pretty small with relatively low preload, so I would benefit almost 100% from adding some compound. If it was high preload and the compound only contributed say 5% then perhaps one would not bother adding it.



Regards,

Jon Reynolds

 

[desertfox]

Hi

I don't think I would rely on the adhesive personally, I would go with the friction which is also highly variable but I would do some practical tests on the joint to establish a coefficient of friction