Torque wrenches and drive extensions.

[fredvance]

Funny this should come up today. Ihave done tests like Willie did in the past. Yesterday I was helping a friend put a header on his bike, on some of the bolts I had to use an 8" extension. He questioned me about the ext. and its effect on the torque accuracy. So I did a couple of the bolts  each way and checked it. They were the same.

[willieworld]

for you disbelievers if your theory doesnt fit the facts there is a very good chance your theory is wrong                                                 willie buchta

[Milwaukee Midget]

for you disbelievers if your theory doesnt fit the facts there is a very good chance your theory is wrong                                                 willie buchta

Torque transfers on an extended axis undiminished?  I really don't know - could well be.  But I suspect what we're thinking, and what probably coaxed you to create this test, is that there is torsional flex in the extension which might diminish the torque being transferred to the nut.  And you may not have disproved that.

Kudos for doing the work, Willie, but there remains a possibility, and I stress, possibility, that there is a flaw in your test.

Your procedure started with the constant, the torque wrench without an extension.  This established our benchmark - the marks you made on the nuts.

You then tested the wrench with an extension, and at the same reading, wound up with the marks lining up.

There is a potential variable not accounted for here - bolt stretch.

If the bolt stretched on the first procedure, the second procedure could coincidentally replicate the results of your benchmark with less actual applied torque to the nut.  By reversing the procedure - extension first, no extension second - you'll have eliminated the potential stretch variable, possibly proven its existence, discovered that flex in the extension may have an effect on the clamping pressure, or simply opened the worm can even further.

In the end, there probably isn't a gnat's nads worth of difference, but a truer test might be to use plastigage between two machined pieces being bolted together, because what we're really trying to discover is clamping pressure, not where a nut indexes with respect to a bolt.

And yes, I'm out of plastigage.

Dr. Willie, thanks for getting me thinking - I am grateful for that!

Respectfully yours,

Chris Conrad

[willieworld]

there is also the possibility the earth will be struck by an asteroid in the next 10 minutes and none of this will matter  LOL    if anyone can think of a test they would like done (im not a scientist just a fabricator) that would be neutral i would be glad to do it  if possible    thanks  willie buchta

[RichFox]

For some years one thing I did at UAL was check and calibrate torque wrenches. Not surprisingly I used a torque tester. I know that if you add an extension that would move the socket out farther from the fulcrum the wrench will read wrong. I never tested with a common extension. I don't think it would make any difference.

[wobblywalrus]

An extension on a hand held torque wrench should not reduce the torque if the wrench is adequately supported.  Not so with impact tools.  A springy extension can absorb the impacts and the nut/bolt will not receive the full shock.

[willieworld]

let me just add this---i think the crows foot off-set wasnt enough to make much of a differance---although the difference could maybe be measured with the right tool---i think the point is its not rocket science its just torquing a nut or bolt  whats really important is if all the nuts or bolts and washers are all clean with no burrs and they are  ALL oiled or they are  ALL dry ------ and the tool is used proper and constant  ---i think it was JACK that said ---tighten it till it breaks then back it up 1/4 turn                                  just some thoughts  willie buchta

[Dean Los Angeles]

With a deflecting beam torque wrench you do have to watch that you don't drag the pointer by being off axis. Parallax errors are common if you can't see perpendicular to the scale, and that's most places on a car.

The best for most applications is the click stop torque wrench. It gives repeatable results. Any time you are trying to read a number you have a tendency to go just a little bit more to make sure.

[55chevr]

I don't doubt that there maybe a loss of value in torsional torque using an extension (micro ounce). But the loss if the extension were dead on the axis is probably not measurable on the wrench as it is so minute. If the wrench were cocked off the axis the reading would be more affected.
Joe