Hacker News Reader

Comment by ominous_prime | Hacker News Reader

A problem with installation is that torque is only a proxy for tension. Torque alone is not incredibly accurate, so there are a lot of cases that could benefit from an easier way to accurately indicate tension than by externally measuring the stretch of the bolt (which often is not an option)

View on HN

True, but this caveat is generally pre-empted in design phase through safety factor principles - multiply certain design parameters by a coefficient greater than 1 to account for the inaccuracy in things like torque-tension correlation.

I’m not getting your point here, won’t there always be inherent trade offs?

If you scale the specified torque, you risk over-tensioning.

If you scale the dimensions of the fastener, that has lots of consequences.

Is there some safety factor here you can add “for free” that I’m not seeing?

Otherwise, why is it that you “generally” pre-empt this caveat with safety factor versus coming up with a more reliable way to achieve the right tension?

The general idea is that proper bolt installation is a solved problem…at least in so far as ordinary industrial/commercial/construction applications are concerned. There’s an informed experienced trade base, a well established supply chain, and sound engineering practices already.

To put it another way, the right tension includes a safety factor. And engineering practice specifies testing and inspection and proven materials.

That safety factor and testing and inspection has costs associated with it.

Basically, it is a "solved problem" by designing things stronger than they have to be, and spending more on testing and inspection than necessary.

I'm sure the cases where these things are economical are fairly few, at least today, but if everyone took your attitude no progress would happen at all.

Safety factors; testing; and inspection are engineering.

The Smartbolt doesn't eliminate engineering.

More importantly, safety factors; testing; and inspection are progress. We have robust assemblies built first time based on rational analysis and formal processes. We have sky scrapers, suspension bridges, and space telescopes.

Obviously it doesn't eliminate all engineering. What it could do, though, is reduce costs. The idea is you could potentially achieve the same level of safety with less material or less testing.

Those things are progress in a sense, sure, but...

It's as if someone finds a way to make an alloy less prone to metal fatigue, and you just say "but you can use thicker metal, and test more." Sure, but.... well I don't see the point explaining this further to you, but I assume people reading this that don't see everything in black and white already understand.

Have you ever taken direct personal legal liability for structural fasteners as a licensed professional?

I have.

Does that give you the ability to have nuanced, big-picture, long-term view thinking?

You might as well argue that someone who works in a bank must understand macro-economics, or that someone in the military must have an excellent grasp of foreign policy.

You are steeped in the status quo. Your thinking on this is expected.

There's a fantastic set of two textbooks covering bolted joint design by John H. Bickford. One for gasketed assemblies and another for non-gasketed assemblies, if you want to really _really_ dive deep into it.

At the end of the day you get the final clamping force by multiplying a bunch of modification factors together that have been empirically measured and the bolting procedure has been verified. When you step outside any of the parameters you've verified in the past, you have to substantiate the math with experimental proof. The size of the fastener, fastener style (e.g., stud vs. bolt), fastener length, thread pitch, torquing method, presence of a washer, presence and type of lubricant, fastener material, base part material, and bolt arrangement ALL play into the final bolt assembly procedure. A lot of these are well-known and published.

Nobody is out there on assemblies like this just slinging unsubstantiated numbers around, at least if they're doing their due diligence.

Even where you can't measure the stretch directly, torque-plus-angle is a better way to achieve consistent stretch than torque alone.

In case anyone else needs to look this up, torque-plus-angle means seating the threaded fastener to a base torque, and achieving the target tension from there by further turning the fastener a specified angle. So super duper simple and something that many of us have done instinctively or from instructions, now we know there is a term for this general technique. Cf https://dannysengineportal.com/torque-specification-torque-t...

So "tighten till you hear a crack, then back a quarter turn" has merit after all?

Yes, though instead of the bolt or part cracking it's the click of the torque wrench, and you're going a quarter turn extra not back.

So really the exact opposite.

Well often nowadays it is torque+angle:

https://www.srtorque.com/error-proofing-tools/digital-torque...

Reply on news.ycombinator.com