Torque Talk

Abstract

How does bolt-tightening work? A puzzle.

This one sounds obvious, but has been bothering me for the last couple of weeks when it suddenly struck me. Most things in life, details, should make sense, but here is one which does not but had managed to pass me by for years.

Think about what happens as you tighten a bolt. First, the nut moves easily, then becomes finger-tight when it is fully in contact with the plate. At this point, what happens? A spanner will twist the nut plenty after this point when there is not more wiggle in the system. An M5 bolt has a pitch of 0.8 mm, so half a turn with a spanner means that an extra 0.4 mm have to come from somewhere. What I have been trying to find out is how this flex is distributed between the shank and the thread.

The point of putting extra turns on a nut after it is just tight is that forces on the bolt are actually reduced. In an application where two plates are being held together, if 2 kN tries to pull them apart, then we would expect that the tensile (expansive) force on the bolt shank would go up by 2 kN; in fact, if the bolt has been slightly stretched to correspond to a static load, then the force on the bolt does not increase at all (think about it!).

I quickly realised I needed some help to work out the real-world mechanics. It turns out there are indeed complications, including the ratio between the compressibility of the clamped material and elasticity of the bolt shank. A 2 kN preload will compress the plate being clamped by perhaps half as much as the bolt shank will lengthen.

Also, bolt thread flex turns out to be a minor factor, as the majority of the extension is gained from shank elasticity. The load is distributed over several turns of thread (doing the calculations for an M5 bolt on the back of an envelope, I get that projected thread area is 6.2 mm2, while shank area is 13.4 mm2, so for a standard M5 nut the stress is about 57 per cent that of the shank).

That just leaves the practical questions of how to make sure the bolt is correctly preloaded for the application. I have no intention of writing a how-to, so leave the details to the reader. There are plenty of fun gauges and methods to read about though if you are interested.

References

If you are interested in chasing down any further thought, including questions of shear or temperature, some links I found helpful: