To Stretch or Not to Stretch? (tech)

[jdbubb at verizon.net (Dan Bubb)]

First off let me say I don't know what a K-value is.
If it's the coefficient of friction then I disagree. The gain in clamping force would not depend at all on the amount of friction, simply on the amount the bolt was turned, which is sort of the whole point of the thing. If a bolt is turned a certain amount then it is stretched a proportional amount and since stress is proportional to strain, the clamping force is increased a proportional amount. The desire is for a consistent clamping force irrespective of friction.

I don't see at all how a small error in the initial torque value translates to a big error in the final torque. Friction is proportional to load, so as the clamping force goes up the torque required to overcome friction goes up. Depending on what the coefficient of friction is you will vary the clamping load for a given torque. But if you only torque to a small value the possible variation in clamping load is proportionally smaller. i.e. torquing a bolt to 40 lb-ft will only give you 40% of the variability induced by friction of torquing a bolt to 100 lb-ft. Then turning the bolt an additional angle will give you the same increase in clamping force irrespective of friction. So, the possible variation in clamping load is reduced using the torque + additional angle technique.

As far as a "solid" application, that's the point of the small torque value. To remove any clearances or the effect of items, such as gaskets, that have high compressibility. The low torque value needs to be high enough so that the additional angle provides a linear relationship between clamping load and angle of rotation. That's the only way you're going to know the clamping load which is what matters.
Course, this does not apply to a stretch bolt. As you point out, with a stretch bolt it goes into the plastic range and the actual angle turned isn't that important as long as you get into the plastic range and, of course, don't exceed the bolt's ultimate strength.
Dan


From: "Allyn" <amalventano1@tds.net>
Sent: Tuesday, January 02, 2007 2:57 PM
Subject: RE: To Stretch or Not to Stretch? (tech)


>> Excellent explanation Allyn!
> 
> Thanks Dan.  I got pissed off at myself for not knowing the real deal about this for so many years, and therefore researched it
> myself over the past few weeks.
> 
>> I just have one comment. You reference lubrication and how it 
>> can change the clamping force of a bolt. Does it not make 
>> sense that torquing to a lower value, and therefore incurring 
>> a much lower frictional resistance to turning,  then turning 
>> an addition angle of rotation will give a more consistent 
>> clamping force irrespective of lubrication?
> 
> I assume you are talking about a standard bolt.  If you apply the plus angle idea to a standard bolt, you will end up with a fixed
> 'gain' in clamping force, where that gain would depend on the K-value, which in turn would vary with lubrication, etc.  Wait a sec,
> ok, I see what you're saying.  Starting at that lower value would introduce other errors into the equation for several reasons:
> 1. starting at a lower value, and torque amplifies as bolt head rotated -> small error in initial torque could translate to a large
> error in the final torque.
> 2. very small angle - since the standard bolt doesn't stretch (at the desired tensile strength), and therefore doesn't provide any
> buffer, you would get a very rapid increase in tensile force for a given rotation.  For a short bolt, a single degree of rotation
> could translate to the equivalent of a several ft-lbf error.  5-10 degrees off, and you'd be way out of range.
> 
>> Just because the tightening spec specifies a torque plus an 
>> additional angle of rotation doesn't necessarily mean the 
>> bolt is a stretch bolt. It's simply a way to get a more 
>> accurate clamping force in a critical application.
> 
> This assumes there is absolutely no galling / fouling of the threads.  Anything that interferes with the no-load free running of the
> threads could interfere with either method.
> 
> This also assumes a very 'solid' application, where there would be no expected compliance issues during the additional angle.
> Imagine trying this on lug bolts, where the wheel wasn't fully seated against the hub when the initial torque was reached.  That
> extra angle of rotation would not result in the intended outcome (and you'd probably lose that wheel after a few miles, to boot).
> :)
> 
>> As far as the crank bolt IIRC there are no reduced sections 
>> or other geometric features typical of stretch bolts. 
> 
> You're talking about the ones we deal with all the time, right?  Don had mentioned a stretch variant available, as if there was an
> option.  I'd guess you and I have only ever seen the 'old' type.
> 
> Al
> 
>