Breakaway torque of threaded fastners.

[turkoej]

We have a case where we torque a lid to approximately 900 ft-lbs using a Hy Torc. There are 48 1.5" bolts on the lid of the cask. It has a helicolflex metal seal comprised of a metal lining formed around a helicaly-wound spring. By design, we do not get metal to metal contact on the lid and flange. We use lubricant on the threads of the bolts when installing them. During the removal of one of the lids, an operator made a subjective statement that he felt that the bolts required a lot less torque to remove them then he expected. Would it be expected that the force required to remove a bolt be significantly less than the torque that was initially applied? We conducted after the fact testing on a lid that was properly torqued to 900 ft-lbs. Out of the 48 fastners, 12 required torques over 500 ft-lbs to remove them and the others were 464 ft-lbs or less. These numbers were based on the pressure reading of the Hy Torc.

 

[tmac]

This might be a lame suggestion, but...<br>
<br>
You might check the stretch in the bolts, over time.<br>
<br>
You also might check the spacing between the flanges.<br>
<br>
If the &quot;Helicoflex&quot; is relaxing, maybe creep, or hysteresis is unloading the bolts.<br>
<br>

 

[AJohnson]

&quot;An Introduction to the Deign and Behavior of Bolted Joints&quot; by Bickford is a great reference to look at for trouble shooting your problem.&nbsp;&nbsp;Two basics that are commonly overlooked could be a place to start.&nbsp;&nbsp;First, you stated that you are lubricating the threads...are you also lubricating the bearing surface under the head of the bolt?&nbsp;&nbsp;Friction generated in torquing a bolt is split about 50/50 between the threads and the bearing surface under the bolt head.&nbsp;&nbsp;Second, a large (~80%) portion of elastic relaxation that occurs in bolts takes place in the first minute after torquing.&nbsp;&nbsp;A simple step is to retorque the bolts to max torque using the same star pattern after having completed the first max torquing.&nbsp;&nbsp;With 48 bolts on the lid, enough time will have passed after the first pass that you can just start over without having to actually stop and wait.

 

[fastech]

Sorry for the delay in response. I am new to engtips and am reviewing some old posts.

Ford Motor Co. conducted an interesting study (and published in Quality magazine)back in the '70s on this very topic. I repeated the study for GM in the late 80's.

You have received some very good answers to this. However, in Bickford's handbook by Marcel Decker, he has a complete chapter on Torque Auditing. (His book is available through the IFI web site.) He proves elsewhere in his book that 45% of the applied torque is lost in thread friction and 35% in underhead friction; thus leaving only 10% of the applied torque available to put stretch in the bolt (or clamp in the joint). From that you can see that if you can reduce the frction loss in either the thread or underhead friction by 10%, you can get a 100% increase in the torque available to put stretch in the bolt.

Back to the Ford &amp; my study for GM, we seen very similar numbers you quote on a variety of bolt sizes (M3 - M24)when torqued in the counter-clockwise direction (for right hand threaded bolts). The greatest loss was always within the first 10 - 30 minutes after the initial torque, especially in joints that contained a gasket material.

In the T=KDF formula given to you, you do not have to divide the &quot;D&quot; by 12 if the bolt diameter is already in inches.

In summary, you can always expect the removal torque to be 50% - 80% of the initial torque applied to the joint. If you want to conduct a torque audit then you must always check the torque in the clockwise direction and closely monitor the degrees of rotation of the bolt head or nut, whichever you are rotating (explained in Bickford's Handbook). You will more nearly approach the 80% level if you precondition the joint by torquing and completetly loosening it at least three times before the final torque.

There is new approach to monitoring torque via an M-alpha plot generated with intelligent drivers and software. Several articles have been written on this approach by Ralph Shoberg and his people, RS Technologies, Ltd in Farmington Hills, MI. He has done a lot of work with Bickford in developing his demonstration figures.

Hope that my experience has been of some help to you. Bill P.
CSFT, Inc.