# OT - Torque the Nut or the Bolt?

Page 3 of 3
• posted on July 3, 2008, 11:25 am

I would have worded it in reverse to make the point more clear. The purpose of the torque is to get the clamping force correct. It is correct that that should also stretch the bolt the correct amount in order to meet the primary goal, the clamping force. How much the bolt stretches, or whether it goes into yield or even fails, is determined by the size of the bolt. So first you pick the force, then you use that to pick the bolt size. And that is why you see eight sizes of bolt on the same lawnmower - it's not really to make you buy eight different wrenches, that's just a side benefit.
You also alluded to the difficulty of getting the accuracy. From what I've seen, nobody with a torque wrench does any better than 50% plus or minus his desired amount of torque, so which end you put the wrong torque on is probably moot.

<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
• posted on July 3, 2008, 1:37 pm

re: So first you pick the force, then you use that to pick the bolt size.
Unless you have no choice as to the bolt size. Earlier I mentioned that we are forced to single-source the bolt. Taking it to the next level, we are required to purchase a *specific* bolt from that single source.
Once we have the parts, the torque used to install them is up to us. If I could purchase bolts from any source I wanted, things might be different from a torque value perspective, but the original question would still stand - nut v. bolt.
re: nobody with a torque wrench does any better than 50% plus or minus his desired amount of torque
Please elaborate. Let's say I'm using a click wrench set to 200 in- pounds. Are you saying my actual torque is anywhere from 100 to 300? I'll assume that's not what you're saying since my assembly will either be flopping around (relatively speaking) at 100 or the bolt would have snapped well below 300.

<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
• posted on July 3, 2008, 3:03 pm

Yes, but how do you decide how much torque? Ideally, you know the clamping force you want between the parts. Then you apply just enough torque to get the clamping force. If you are forced to accept one particular bolt, you run the risk that to get sufficient clamping force you have to overtorque the bolt, risking failure, or undertorque it, risking loosening in use. Do the calculation. If case 1, forget it. If case 2, better use loctite, a very good lockwasher setup, etc. Torque is there for a reason.

That seems strange to me too. But apparently that's what the studies show. Also the torque can fail to have any correlation to the clamping force.
By the way, torque is usually specified dry or wet. It's been decades since I put a fastener on without lube, so I always look for wet. I don't even like to put in a wood screw dry! <g>
And if I've got any of this wrong please correct me. Engineering school was a .......long..........time ago and I may be misremembering.

<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
• posted on July 4, 2008, 12:30 am

re: Yes, but how do you decide how much torque?
Trial and error over many years of experimenting. To prevent movement, we go with the max before the bolt will fail, based on past experience. In the other application, where we went to allow for movement, that number might vary from installation to installation based on how "loose" we want it - it's a case by case thing.
re: studies show...(the +/- 50% error)
Nothing personal, but I can't accept that without some substantiating citations. If I were truly getting anywhere from 100 to 300 when my click wrench says 200, then there is no way we could have zeroed in on certain specific numbers that prove to work best time after time after time. Why would we see bolts fail at a fairly consistant number that is well below the upper range of a 50% error? I'll bet I could consistantly snap the bolts at less than 10% above what we use as our max number. Finally, the torque values we use work for our application - and the bolt failure points are consistant - across multiple torque wrenches from various manufacturers.

<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
• posted on July 4, 2008, 7:46 am

(I thought you were having some sort of "problem".)
Generally, I think, a condition of "over torque" might be expected to produce a satisfactory result in many if not most "normal" instances.

Whatever you do, don't trot out to the bolts acting all crazy and confirm your assumption.
I'm no engineer, but 10% seems to me as if you may be using the wrong grade bolt... depending, of course, on your application, which, in this case, could scarcely be more vague..

Similar errors of tolerance might be expected to produce similar results.
I have only the slightest of ideas of what you are expecting of a bolt, but my gut feeling is it's too much, and, to answer the original question, again, from which end you torque it isn't going to make any appreciable difference. -----
- gpsman