Just for fun I did sumit the question to the Myth Busters forum a couple days ago!

Well this is interesting. I did 6 pulls at 30 FtLbs with the straight torque wrench and all fell within the 30-31 FtLb range, well within the expected results. I did 6 pulls with an 18" X 3/8 extension and got the same numbers. I stepped it up to 100 FtLbs, 6 pulls on the wrench gave me readings of 100 to 101. 6 pulls with the extension gave me readings of 96 to 98. There is "some" loss but had I used a 1/2" drive extension I don't think there would have been any detectable. Even at that the variation is still within 5% which is acceptable. Yes, there is some torque loss at higher settings with a smaller diameter extension but it's minor.

Hippie I was expecting something like that, I was waiting for someone to comment on the two wrench test I kept asking about before saying anything but a couple of days ago at the shop here we took two identical S/K 1/2" torque wrenches and two 1/2" drive extensions (6" and 12") and performed the test I kept asking About. The extensions were in a lubed pipe clamped in a vise and were pulled evenly by two people and the results were somewhat different than what you got. All pulls were well within tolerance at about 4% at the upper range with one wrench consistently under the other even when we switched ends of the extensions. this would confirm a slight difference in the two tools. This was done by 6 different people with multiple pulls at many different torque settings. Next we took a piece of 3/8" X 10" mild steel bar and applied torque to about 65 ft/bs at which point the bar began to twist. At about 74 ft/lbs the torque reading was within a couple of lbs even though the bar had twisted about 30 deg. In other words nothing we did gave any error between the wrenches, if you do not believe this try it yourself. How would you explain the error you got other than mechanical? You and I both know that it is a physical impossibility for that extension to have more force on one end than the other.

Wait for it.....wait for it.... you know it's coming.....

Rat. As I said It was about what what I expected Let them believe what they want or maybe we should all call up everyone from Snap-On tools to NASA and tell them they are wrong Anyway you be your own judge.

I'm sorry you don't like the results but the machine doesn't lie. Roll your eyes and nay say all you want, I'm using certified test equipment not a piece of pipe clamped in a vice.

Like I said if it was a 1/2" extension I'm sure there wouldn't have been any difference but I could see the extension physically twist approximately 10 degrees over it's length before the wrench clicked at the 100 FtLb setting. Torque is the application of energy, where did the energy used to twist the extension come from? It had to come from the force I applied to the wrench and if the wrench "clicked off" at 100 FtLbs and some of that 100 FtLbs was being used to twist the extension why would you expect to still have 100 FtLbs at your fastener or in this case the test unit? If you use the correct strength extension then I agree there is no discernable loss, but using the lighter extension there IS loss.

Hippie Why would I expect the energy to still be there? You are confusing ENERGY with FORCE. As I said earlier the total amount of energy will NOT be exerted on the fastener. The total minus what it used to twist the extension to equilibrium is what reaches the fastener but equilibrium is just that, the same amount of force on each end! The amount of force applied to a fastener is what the torque reading is, not how much energy it takes to get there and any extra energy used twisting the extension is gained by extra wrench movement using constant force(torque) . I explained this earlier the extra energy is gained from extra wrench movement and stored in the extension as static energy and will be released when the force is relaxed. You are now using the same old argument that has been tossed around for years but I bet you cannot explain why the wrench quits turning if it has less resistance on the fastener end than force applied to the wrench end. Please explain how I am wrong if I state that the force on one end HAS to be equal to the resistance on the other end or the thing would continue to turn. Simple laws of physics cannot be explained away by mechanical error.

I'm not confusing what the test equipment tells me. 100 Ft Lbs is 100 FtLbs., 98 FtLbs is 98 FtLbs. 2 Ft Lbs. went somewhere, the machine doesn't lie. I don't have an opinion one way or another on where it went, it doesn't really matter, I saw what I saw. 2% at 100 Ft. Lbs. honestly doesn't mean squat for most of us here. You guys can keep on arguing all you want. My torque wrench is still accurate after all these years and I won't use my long 3/8" extension when I'm torquing over 30 FtLbs, that's all I really care about.

Hippie, I did not take exception to your findings but I just could not help but think here we go again when you started talking about the thinner extension and flex absorbing the torque when that quite plainly is a physical impossibility. Machines are fallible due to either malfunction or input error but the laws of physics and mathematics are the laws of the universe and are inarguable. I keep asking for someone to show how an extension can have less force on one end than on the other but so far no one has made any attempt to do so. Also let's not forget that as has already been pointed out the manufacturers themselves state clearly in their manuals that this error does not exist.

Hippie-
I'm not going to fault your equiptment or findings,but I am in agreement with Old red that both ends of the extension should,by basic laws of physics,see the same torque.Since at 30 ft/lbs,the readings of both wrench and tester agreed,I feel that the discrepancy at 100 ft/lbs is indicative of a mechanical loss,probably due to friction in the method of support for the extension.
If a loss of torque is possible in this situation,then we would also have to accept a loss of engine torque is possible through the driveshaft and other power transmission shafts of the cars we build and drive.For example,a motor putting out 500 ft/lbs of torque at the flywheel would be losing 2% through each shaft the power was transmitted through on it's way to the differencial,then an additional 2% through the axles.We know that this is not what happens.
The 10 degree twist in the extension is indicative of additional energy in the form of a longer arc of motion needed to reach the desired torque in the system,not additional torque applied to the wrench.
If we could support the wrench end of the extension in a frictionless bearing,located so that the centerline of the extension was perfectly perpendicular to the plane of rotation of the tester,by all known laws of physics, the amount of torque exerted on the wrench will exactly equal the torque transmitted by the extension on the tester.

George

OK, I have been thinking about this and I may have a plausible theory(NXS will love this ) Up until now I really had not given all that much thought to this since the physics involved are obvious and this has been hashed out thousands of times over the years, BUT, Hippie seems to have uncovered a discrepancy of some sort that appears to defy logic so the question is, how? If one is using a "clicker" type wrench then there may be an explanation for the error that does indeed involve twisting. A "clicker" type wrench does not measure torque in the sense that a gauge will take a direct reading it simply releases or "fails" suddenly at a predetermined setting and once this happens it no longer "reads" anything, if you continue to pull it would not show any increase in "reading". So if using a "springy" extension and applying force at an increasing rate the extension will deflect according, that is it will deflect at an increasing rate and since the resistance at the fastener end will not "see" the full amount of force being applied to the wrench end until it reaches equilibrium it will lag behind slightly and may be the cause of the problem. Since the clicker wrench only "reads" one spot and only at the instant it senses the setting it will snap when it reaches that setting but since the resistance is lagging behind due to the increasing rate of deflection it(resistance) will not have reached an equal amount of force at that instant. This will have no effect on a gauge or beam type wrench or any other type of torque measuring device that takes a continuous reading of the resistance(torque) and in no way changes the principles involved in applying force through a shaft(an extension in this case)but it does identify a design flaw with the "clicker" type wrenches and is thus a mechanically induced error. While this is only a theory I have come up with it does seem, unless I am mistaken, to explain the discrepancy assuming Hippie was using a clicker type wrench.

Opinions on this?

Yes, I was using a clicker so I can buy that. I think you may have something there. Either way the little variation I experienced wouldn't really have much affect on what most of us here would be torquing. You can get that much or more between two different people using the same wrench.

This whole thing has been ridiculous for more reasons than one but the worst reason is that we have been in disagreement over an error to small to matter anyway. The problem is when someone tries to correct the "error" by adding 10% or even more(NXS never said it was that much)and indeed I have seen this done. I guess that is why I take this so seriously since it usually is "corrected" by a large amount and with any kind of wrench. I have an acquittance that is an engineer with a mining equipment company and I will run my theory by him and see what he thinks.

man this thread grows fast!

All that is affected by adding a socket extension is the accuracy...typically a torque wrench will display less torque the longer the socket extension is. The accuracy loss is always negative.

Torque wrench FAQ by Norbar.

Heres a snippet;



Please note the last line.

The old wives tale surrounding "extensions" and torque wrenches is related to what different people call an extension...classically an extension changes the length of the handle but most people think socket extension.

Savvy?