What you always wanted to know ... (torquing heads)

[StephenB]

... and never dared to ask:

So, I have done this before, torquing down a head on the /5. Never experienced a leak afterwards, so I must have done it right. What did I do: used small increments such as 1/4 turns in a cross-wise pattern with a clicker wrench set at 10, then 18 then 26. So, 3 "clicks" per nut, if you want. The idea behind is to tighten the nuts very evenly with the least possibility for warping the heads or incur any of the other possible problems when tightening unevenly. So, this worked for me but has really nothing to do with torquing heads in three stages. I could set the clicker at 26, and then go round in a cross wise pattern until they all click.

I am not sure what the stageing is trying to achieve other than the above but let's assume one would do the following:

1) tighten nuts handtight
2) set clicker to 10ftlbs
3) tighten first nut in one go to the full torque, then the second, third and fourth nut in a cross pattern
4) set clicker to 18ftlbs, repeat step 3
5) set clicker to 26ftlbs, repeat step 3

Step 3 frightens me: if all nuts are handtight and I tighten one of them to 10ftlbs, things could get warped quickly, the gasket deformed, all sorts of bad hings can happen. So, this can't be right!

Poses the question: how is torque staging done correctly for this application?

( I am discarding the two nuts on the top and bottom studs for the moment, they don't need to be part of the cross-wise torquing in my opinion.)

[Duane Ausherman]

I generally do them in two stages, 15 and 23. I don't ever go to 26 lbs, as IMO that is too high and just increases the chance for pulling a stud out. I prefer 23 lbs.

I never use a clicker type torque wrench for this. I have both 3/8" and 1/2" clickers, but only use them for out of sight applications. They are less accurate and I still check them against my beam type once in a while.

Most important is that I pull them up to torque very slowly.

These heads aren't prone to warping like the older /2 were.