Torque question

[REVz]

Simply...why (x)Nm + xx degrees, e.g, 40 Nm + 32 degrees.

Does it relate to some sort of increasing tightness curve once threads reach a certain torque, making the last measurement impossible to convert to an equivalent torque number...?

 

I know someone out there can explain it quickly, so that even I could understand it...Thanks

larry

[Joe Frickin' Friday]

Simply...why (x)Nm + xx degrees, e.g, 40 Nm + 32 degrees.

Does it relate to some sort of increasing tightness curve once threads reach a certain torque, making the last measurement impossible to convert to an equivalent torque number...?

 

I know someone out there can explain it quickly, so that even I could understand it...Thanks

larry

 

In tightening a bolt, your goal is to achieve a certain amount of tension. The applied torque correlates reasonably well with the final tension, and it's so darned convenient to measure (compared to measuring bolt stretch or something else) that it's the only spec used for most fasteners (like wheel lugs, oil drain plugs, etc.).

 

A problem arises when you need a very carefully controlled amount of bolt tension, and that's when the inaccuracies of the torque-versus tension correlation get in the way. Variations in thread friction due to lubrication, grit, damaged threads, surface condition under the bolt head, and even the speed with which the mechanic tightens the bolt, screw up that relationship.

 

The good news is that the geometry varies little from bolt to bolt, so if you turn two bolts by the same angular amount, you advance them by very much the same distance, and therefore increase the tension by very much the same amount, regardless of state of lubrication, grit, damaged threads, or whatnot. So instead of specifying one big torque, you first specify a small torque (to pull everything together), and then a certain angular turn. This is more of a PITA, but when you need the exact right amount of tension, it does a lot better than just one big torque spec. This is pretty common on cylinder heads, where you need enough tension to guarantee a sealed head gasket, but not so much tension that you distort the head and compromise the seal (or crack the head). The other case where a torque+turn spec is used is on torque-to-yield bolts, in which the bolts are tightened so much that they actually are permanently stretched a bit. This lets you get the maximum possible tension from the bolt, without having to go to a bigger (heavier, more expensive) bolt. You'll see these on connecting rods (at the crank end) and on flywheels. When you're taking the bolt that close to the limit, you need to hit the desired tension/stretch as reliably as possible. Presto: the torque-turn spec serves a lot better than a simple torque spec.

[REVz]

Somehow I had only thought of the angular velocity part of the equation, but the many other vartiations you list all make sense. Seems like there might be an added chance for variation though.e.g., ...the initial calibration of the torque wrench and then add the calibration of the angular measurement. So, if a reliable result is needed then reliable, repeatable, methodology must be utilized.

 

Thanks Mitch, good explanation.

[cgs500]

If you want to be real accurate http://www.rotabolt.com/html/rotabolt.html

[Joe Frickin' Friday]

Seems like there might be an added chance for variation though.e.g., ...the initial calibration of the torque wrench and then add the calibration of the angular measurement.

 

Well, you need to snug things up before you can use the angle-turn to generate tension. But even if you use a crappy torque wrench that's off by 50 percent, the error in tension created during that step ends up being very small compared to the total, final tension you produce after you apply the angle turn.

 

And as for the accuracy of the angle-turn...you can get a degree-wheel to fasten on the end of your ratchet, and if you're paying attention it's easy to get within one degree of arc; on 90 degrees of turn, that's not much more than 1 percent error, far better than you can expect in a torque-wrench-reading-versus-bolt-tension relationship.

[REVz]

good answer again. Are you sure your middle name isn't Archimedes or a least "Fricken Archimedes?"

 

thanks