Originally Posted By: doitmyself
"analysis paralysis"
I like that term! I suffer from it too.
You hit the nail on the head with your 20% figure, except my research showed it's more like +/- 30%. When I worked for a NASA contractor I was charged with writing our torque specification and did A LOT of research. (This was Space Flight Hardware for goodness sakes!) Weeks of research showed that torque, like you say, is a very poor indicator of bolt clamping force, which is what you're trying to measure by measuring torque. My research showed that the preload variation is +/- 30% for the same torque.
A few years ago, I was investigating a quality deficiency report on a bolt from the driveline of a helicopter. A statistically significant number of these bolts had failed upon torquing during installation. I determined that the unit was using torque wrenches, they were calibrated properly, and the mechanics were doing everything properly.
I conducted laboratory testing on many samples of this bolt, measuring bolt elongation as a function of torque. I also measured the torque required to shear the bolt. Turns out that if you had a torque wrench that was on the high end of acceptable calibration, you had a less than average (but still acceptable) bolt strength, and the turbine engine oil used to lubricate the threads per the TM gave slightly less friction than normal, and you fell on the high side of that +/-30% I mention, you'd shear the bolt upon install. The statistical probability of all those factors was in line with the percentage of bolts that had failed.
Measuring the elongation confirmed the inaccuracy of torque as a measure preload. (For the same cross-sectional area, preload is elongation until the material yields and becomes plastic instead of elastic.)
(FWIW, I recommended that the bolts be made stronger by changing the manufacturing method from cut-threads to rolled-threads and testing for required hardness after heat-treat.)
I agree with your analysis that using something, anything, is going to get you in the proper ballpark and is what's required.
We use clickers at work (by Snap-On) and they go out to cal at a NIST-traceable lab yearly. Never had one fail or come back having failed cal.
At home I use Craftsman clickers. They are okay. Quality is so-so. They seem to work fine. Never have broken a bolt or had one come loose using it. Have broken bolts when I don't.
Beam types are very simple. I'd rather spend $30 on a beam than $30 on a clicker. It's easier to get the beam-type made right than the clickers...