Guys:
As usual , the original thread feeds a number of peripheral issues
Couple of things...I wouldn't worry much about the 94-96 ft-lbs discrepency..it just doesn't matter. In fact, the whole torque number routine is not optimal...as the deciding factor is to "close-in" on clamping force, and torque settings are the nearest-good ( bad?) approximation of getting to that clamping force target ....note all the variables like lubed / dry ( itself the cause for 30-40% changes in target torque values)....that come into play.
Ok
...let's talk torque wrenches. A good idea is to have a click type ( easier to use and it ratchets)...backed-up by a cheap ( but almost always accurate) beam type, just as a quick calibration. The beam types don't lose their accuracy, unless mishandled, whereas the click type can get out of calibration...and very badly out of calibration, too.
--- Wil Ferch