In the late 1970s, I was the marketing manager for a company that made
ultrasonic nondestructive testing equipment. We made ultrasonic
thickness gages that would measure metal parts to better than 0.001
inch. One project we did was to work with a fastener company to develop
for NASA a length gage for large fasteners used on the SPace Shuttle,
measuring stretch for preload. We followed up with one for large bolts
used in bridges. Both worked like a charm and could be used on blind
bolts, as it only needed access to one end of the bolt.
As part of this project, we learned a lot about bolts. What Mark is
saying is completely in line with our R&D results.
Jim
Mark Palmer wrote:
>
> Listers,
>
> A bit of engineering perspective, to try to clarify things for those who are
> interested (everyone else can delete & go on with more fascinating
> pursuits).
>
> Rod bolts, like any threaded fastener, require a certain amount of PRELOAD
> in order to reliably hold the joint together. This "preload" is a FORCE,
> measured in pounds (or newtons, if you're anywhere but the U.S.). The
> amount of required preload (force) depends upon the loads (force) that the
> joint will see during operation. Too little preload, and the fastener will
> fail due to fatigue as the joint undergoes cycles of load/unload (most
> frequent failure mode). Too much preload, and the fastener yields
> (comparatively rare).
>
> HOWEVER, we can't easily measure the preloaded force in the fastener. That
> is, we can't measure the force in pounds directly -- we can't hang some fish
> scale on it & measure the pounds force in the fastener. So we have to
> measure something else, to get an estimate of the preload.
>
> The easiest, and most reliable, is to measure the length of the bolt.
> Easily done with a micrometer (assuming bolt does not go into a blind hole).
> We measure the length of the bolt before installation, and after
> installation, and the difference is bolt "stretch". The engineering term
> would be deformation.
>
> Fortunately, since the material exhibits elastic material (up to a point
> ...), the deformation ("stretch") is directly proportional to the amount of
> force (pounds) applied to the fastener. So, by measuring stretch, we know
> the amount of preload we've applied. We know whether we have reached the
> target amount of preload for the joint. "Stretch" is a VERY ACCURATE method
> of estimating preload.
>
> On the other hand ... torque is a pretty awful way of estimating preload.
> There are just too many variables -- principly, the coefficient of friction
> in the threads and under the head of the fastener. If you could somehow
> magically measure actual preload (force), you would find that you could
> torque 10 fasteners to the exact same torque and have widely differing
> preloads. Not a good thing on a rod bolt or bearing cap.
>
> As an aside (this whole thing's an aside, actually) ... in critical
> large-scale fasteners, like you find in civil work and heavy industry, the
> preferred tightening method (if you can't measure stretch) is the
> "turn-of-the-nut" method. Tighten to snug, then continue to tighten to a
> specific degree of rotation of the nut. or bolt. This is actually more
> accurate than torque, in terms of reaching target preload. Even civil
> engineers regard torque as the least reliable method.
>
> So ... the right thing to do, is tighten the fastener until you've reached
> the recommended "stretch", which will assure that you have reached the
> target preload. Use whatever torque you need, to reach that stretch. If
> you somehow inadvertently exceed the recommended stretch, take the fastener
> back out, re-measure in the unloaded state, and compare to the original
> unloaded length. If you have caused a permanent deformation (bolt now
> longer than original measurement), throw the fastener away -- you have
> loaded it beyond yiel, into the plastic deformation range, and damaged the
> bolt.
>
> If, on the other hand, you are faced with a fastener in a blind hole, you
> can't measure stretch ... so you have to measure torque, and HOPE that the
> indicated torque has produced the target preload.
>
> Regards,
> Mark Palmer
> _________________________________________________________________
--
Jim Hayes Fotec/Cable U
hayes@mediaone.net http://www.CableU.net
jeh@fotec.com http://www.fotec.com/
All generalizations, with the possible exception of this one, are false!
|