Can anyone answer me this? I’ve had quite a debate at work and I’m sure I’m right.

You have 2 identical bolts and lubricate one and not the other. This will decrease the run down torque on the bolt you have lubricated. So say it halves it leaving you with a rundown torque of 5 on the lubricated bolt and 10 on the unlubricated one. You factor this in and add a torque of 20 on top of the rundown torque.

You now tighten both bolts up to this figure, so 25 on the lubricated bolt and 30 on the unlubricated one. Is the clamping force going to be different on the two bolts?

Hmm.. tricky…

If the additional run down torque is simply due to friction, then the clamping force will be less (how much less depends on loads of factors) in the ‘stickier’ bolt for the same applied torque (the additional resistance is still there when you are putting on the torque to preload the joint, which is still working against you in generating an axial preload in the bolt).

Torque is a very blunt tool for generating what you actually want to control – i.e. the axial clamping force in the bolt. Axial bolt force may be +/-50% for a given torque depending on friction at the threads, washer, heads etc etc.

Any use?

http://www.spaenaur.com/pdf/sectionD/D48.pdf

depends on how slippery your lube is – you cant just say ill add a factor into my torque to hope its the same …..

your right though most torque values given especially in the bike trade never specify

Potentially not, but there’s a lot more to it that run down and clamp torque.

Assuming everything else is the same, then there is friction under the head and in the thread, so you would have a change in “lost torque” on the thread as you are turning (like having a nyloc or normal nut on there) and then a change in the friction under the head as you are tightening it.

If all that come out the same, then you still have the normal scatter in the joint, so two joints won’t necessarily give the same clamp load for the same geometry (unless you’ve got really tight manufacturing controls and you can actually measure the stretch on the bolt).

Scatters can be huge (e.g. VDI2230 quotes a 43% variation in preload for certain dimensions of joint using a torque-controlled tightening method). And thats assuming you didn’t accidentally lubricate the mating surfaces 😯

(I have to design these kind of things, bolts are a right pain! check out this 992 page INTRODUCTORY beast: big fat bolt book)

The reduction in rundown torque would purely be due to the reduction in friction from the application of lubricant.

One arguement was that the additional friction has already been factored in by negating the rundown torque but i think it will still have an effect when you are applying the ‘actual tightening’ torque so the unlubricated one will still have less clamping force.

And for the purpose of this then yes we are assuming everything else is the same (although I appreciate this would never be the case in real life)

I like one of the reviews of that book:

As Plato expressed long ago, the wiser you are, the less you realize you know.

Ooooo – VDI2230 – light bedtime reading, esp making sense of some of the English translations from the German source 😉

I think, purely in this simplistic example you’ve described:

One arguement was that the additional friction has already been factored in by negating the rundown torque

No

when you are applying the ‘actual tightening’ torque so the unlubricated one will still have less clamping force.

Yes

….. but as noted above it’s pretty complex in reality.

if the run down torque is the torque before the head hits the plate, then you’d have some lube under the head (probably), so you’d get less losses there too.

The extra 5 is just for the thread loss, but you really have (thread AND head) losses.

So sounds like your lubricated 25 one would have more clamp load than the unlubricated 30 one.

But if you’re using something like a nut runner or impact wrench, then you’ll have very little actual torque control. Whoever designed the joint would (should) have factored that in, so practically it might not make any difference.

As Plato expressed long ago, the wiser you are, the less you realize you know.

Yes… I wasn’t wise enough to go through book 2!

So sounds like your lubricated 25 one would have more clamp load than the unlubricated 30 one.

This was my arguement. Unfortunately our debate was effectively ended by a senior engineering guy agreeing with the other guy.

I realise in practise there is little difference and it is impossible to control to this point anyway in the majority of applications.

I’m sure I’m right in theory though.

This was my arguement. Unfortunately our debate was effectively ended by a senior engineering guy agreeing with the other guy.

I’m a principal engineer – does that trump a senior engineer? Although in an engineering cockfight the one with the best looking equation will usually win…..

I’m a principal engineer – does that trump a senior engineer? Although in an engineering cockfight the one with the best looking equation will usually win…..

Likewise here – I get trumped only by the director I answer to….. and the I completely agree, some of the equation-fights have become legendary 😉

re the op. ime (commerical vehicle wheel joints) you cannot determine what the clamp load will be without testing.

my experience comes from tech specialists at a veh. OEM

what you can state is if you lubricate one thread and not another and do them up to the same torque, then obviously the lubricated one will have more clamp load. this may or may not yield the bolt.

someone said the variation of clamp load v. torque is huge, and ime this is correct. even using test samples sent from the supplier, and using lab. quality torquing equipment the clamp load can vary by the figure mentioned, say 40% or more.

most of the torque (50%) is used up on underhead friction, 40% on thread leaving a small amount for clamping.

as you can see any small amount of difference in the friction co-efficient between the bolt and the nut/washer can lead to a masssive difference in clamp load, and this can easily be enough to exceed the yield of the bolt, or, not have sufficient clamp load. the variation generally comes from the coating, and we experienced some issues with the recent ELV (delete chrome 6) legislation because the old hex cr coating were not only cheap, but they had quite good variation properties as well.

the way to determine what torque to use to get what clamp load involves a lot of testing on a lot of expensive equipement, and you have to do a lot of it to make sure your fastener is both capable of providing the necessary clamp load and also not put the bolt into yield at both extremes of tolerances.

As this is impossible to predict, in bridgeworks we use load indicating washers in order to provide confirmation of the tensile load.

On the bigger bolts it is sometimes impossible to provide the necessary torque to squash the load indicating washer without lubrication to rotating faces in contact.

porter_jamie: Whilst I agree with sentiment, it is relatively easy to do limited testing and (statically) guarantee you won’t yield the bolt and get the clamp force required for the joint PROVIDED that you not only specify the torque, but also the full assembly procedure that must be used (this included lubrication, material grades for bolts/washers/flanges, torquing procedure, etc etc. Specifying the assembly procedure is the bit that is often neglected.

In industrial situations where you are talking BIG torques, so interface friction is a big issue, you wouldn’t even torque the bolts at all, but use a hydraulic puller to stretch the bolt to the required load, nip the nut up ‘finger tight’ and then let go. Doing like this means friction is a non-issue to the clamping force and the axial preload is controlled directly.

We had an issue on what was basically a 14″ dia pressure vessel a while ago. (14″ dia tube clamped between two plates with x24 M16 tie-rods). Seal blowing at >10 bar. Suspected inconsistent torques, despite repeated sequential tightening. Dab of grease on the tie-rods & all was good. Didnt up the torque, just lubed & sequentially tightened again. Air tight this time. I’d say, for a given torque the lubricated one would generate more force.

what you can state is if you lubricate one thread and not another and do them up to the same torque, then obviously the lubricated one will have more clamp load. this may or may not yield the bolt.

I realise this. The question is (in the theoretical world where everything else is exactly equal), would factoring in rundown torque completely negate this?

This had got really complicate really quickly but I think the general opinion is that I am correct.

We have a chief engineering fellow here on site. I’ll go run it past him.

In industrial situations where you are talking BIG torques, so interface friction is a big issue, you wouldn’t even torque the bolts at all, but use a hydraulic puller to stretch the bolt to the required load, nip the nut up ‘finger tight’ and then let go. Doing like this means friction is a non-issue to the clamping force and the axial preload is controlled directly.

That is really interesting and makes perfect sense.

In my experience, it is good practice for the nutrunners to undo the bolts by 1/4 of a turn after the run down before doing the final torque. The “scatter” hugely decrease after the first tightening as the variation in the surface finish is usually smoothed to a fairly consistent level after the first tightening.

A bit off topic but using yield bolts usually achieve very consistent clamping force in my experience. As long as you get to yield point, the clamping load becomes consistent and is what we use for the most critical situations.

It is a pain to stop people reusing those bolts though.

Also, testing can be done fairly cheaply if you get hold of a ultrasonic bolt meter (~£10K for the meter and couple of quid for each bolt). Still need to confirm the results with strain gauges though as it’s not as consistent. It gives a pretty good idea of where you are though.

Is 30000 ftlbs a big torque in an industrial application ?

30000 ftlbs

😯 How big are your bolts??? 5000 ftlbs is about M40 size (7000Nm for 12,9 ish…)

I’m normally looking at M20-60, plus some >1000mm locknut things (but those don’t really follow the same rules).

(this is for stuff with peak shaft torques in the mega-Nm)

Is 30000 ftlbs a big torque in an industrial application ?

Depends?

Currently I’m running 90mm bolts up to 10,000Nm

bristol bike – i can only speak of my personal experience, which is developing new bolted joints (or actually, carrying over an existing one into a new application). i do know if you change coatings (ie delete cr6) but cannot change any of the other hardware (bolt diameter, thread type etc, tightening strategy) you can run into all sorts of bother!

i know there was a move towards torque to yield for a lot of fasteners. the idea here is you defo get a more consistent clamp load, and because the variation is a lot less, you can actually use a smaller fastener if you want (or more likely situation tbh, you can get more performance out of your existing one). remember the variation on some joints using torque alone (in controlled lab conditions) can be enough to put one bolt nearly into yield, and another one almost without enough clamp load.

the production kit for t2y (dc nut runners etc) is much more expensive though.

edit add – critical fasteners i believe are always tightened using a non-torque method. a recent development i saw was using specially prepared bolts and ultrasonic kit which measures time of flight of an impulse along the length of the bolt – you work out your required clamp load and then test to see what bolt stretch you need to achieve that and then use the same kit to measure bolt length as you tighten it. also useful for checking it has not come undone in service.

I dont torque bolts

I do pipe threads.

Use api pipe dope and a bucking unit to put things together to high torque and generally using other lubes renders my graphs all wonky and a fail on qc. . Metal to metal contact seals at top and bottom of thread require torque graphs monitoring and fairly precise amounts of dope on

In industrial situations where you are talking BIG torques, so interface friction is a big issue, you wouldn’t even torque the bolts at all, but use a hydraulic puller to stretch the bolt to the required load, nip the nut up ‘finger tight’ and then let go. Doing like this means friction is a non-issue to the clamping force and the axial preload is controlled directly.

Sounds cool, do you need a special type of bolt and what does the puller look like?

http://www.f-e-t.com/our_products_technologies/drilling-intervention/tubular-handling/torque-equipment/high-torque-stroking-bucking-unit-/

Apparently one of the ” spanners” i use is good for 250000ft lbs…….

So to answer my own questiom 30k ft lbs isnt all that high.

These are quite clever and are used on structural steelwork / construction projects as a replacement for traditional HSFG (high strength friction grip) bolts.

They have a nice little video: http://www.tcbolts.co.uk/2_installation.html

Bolts seem to have a sacrificial bit on the end which breaks off to prevent re-use (I guess).

If you use the right tool to install the bolts then it does the bolt up to the desired load, then the tool shears off the sacrificial bit to show that the bolt has been set correctly using the proper tool.