There are also those extensions with universal joints on them. And they could allow an increase in effective torque from the end of the ratchet. However, again the required bracing of the ratchet head plays a role: any extra torque from the increased moment arm of the handle is exactly counteracted by an opposite torque from the head. For instance, suppose the U-joint allows you to displace the head by 10 centimeters, and you are applying 400 newtons of force (both on the handle and the head). You get an extra positive 40 N-m from the handle displacement, and negative 40 N-m from the force applied to the equally displaced head. These two numbers cancel no matter how the ratchet is displaced, as long as there is no net force on any axis (as must be the case if the extension has extra degrees of freedom).
I read page six of the article in “Aviation High School Revision B
Basic / Advanced Jets”. That is almost word for word what is written in my “Black Standards Aircraft Procedures and Practices Handbook”. I could look it up, next week, in the FAAs part 43-13 if anyone cares.
For what its worth, I tested the effect of extensions on my own torque wrenches as I had them certified before I used them on aircraft. The inspection department has a torque wrench tester. I used this along with 3", 6", 9", 12", & 24" extensions. I tested at 25, 50, 75, 100, 150, 300, 450, 600, 750, 1000, 1150, & 1300 in-lbs settings. I used these settings and these extensions as they would bracket what I used on real aircraft. I used a 1/4" wrench and sockets for below 150 in-lbs. Above that I used my 3/8" wrench and extensions. For my 1/2" wrench and its extensions, I used the 1000 in-lbs and up settings. So these three settings were tested with both 3/8" and 1/2" wrenches.
As expected, the shorter extensions at the lower torques made no perceivable difference. The 12" and the 24" extensions at and above 1000 in-lbs made a measurable difference of 2-3 in-lbs. This error is within the allowable error for my torque wrenches. Please note that all of my torque wrenches are of the “click” type.
So, yes there is a small amount of undershooting with extensions, BUT it is very minute. I never worried about it after these experiments.
Off set with a crows foot or similar devise, is dealt with by making sure that the crows foot is perpendicular to my torque wrench.
In my over 20 years of working in the aviation maintenance field, None of my torques were found to be under torqued. Note that ALL of my torques were double, and often triple checked.
I do not recall EVER seeing a “beam type” torque wrench used in repairing or building aircraft. IHTH, 48.
We’re talking about the same extensions, right? These things? They shouldn’t change the torque applied to the bolt at all. You should be able to have an extension five feet long, and although it would twist up pretty good, the torque you applied to one end of it with your wrench should be countered by the exact same torque at the bolt head (lest the extension begin spinning wildly). 2-3 inch-pounds at 1000 inch-pounds is an error of 0.2-0.3 percent. That’s pretty small, but it still defies physics; it should disappear with repeated testing of multiple torque wrenches using multiple load cells in a randomized testing sequence.
There could be a factor here that we did not identify yet. With high torque and a long extension, perhaps there’s a change in the way one uses the wrench that causes it to slightly undershoot the torque.
Maybe you’re more likely to wind up out of square under those conditions and Willys technique caused him to be out of square with the effect of shortening the lever arm. Maybe someone else stays square, or is out of square lengthening the lever arm.
With a 25" wrench, the lever length only needs to be off by 1/10th of an inch or so to get the tiny change he noted.
Excellent post!
I, for one, will take this as the definitive answer to the OP.
I wonder if the wind-up of a long extension in conjunction with the click mechanism of the torque wrench might account for the discrepancy.
Yes, those are the ones. I only did five or so tests at each setting with each wrench. So I suppose that is multiple. My statistics instructor would want many more tests, perhaps hundreds of them.
As Cheesesteak mentioned, I would only have to be 0.1" off over the two feet of extension that I used. That is probably the explanation for this minor discrepancy.