Applying torque

[duncan webster 1Participant @duncanwebster1]

When I worked in a steam turbine drawing office there was a rule not to use anything less than 5/8 bsw. They reckoned that even the most horny handed fitter couldn't overtightening them unless he went and got a scaffold pole to put over the spanner. A lot of the sets went to developing countries so perhaps skilled men were in short supply. Getting that way in the UK.

[Howard LewisParticipant @howardlewis46836]

When Rolls-Royce Oil Engine Division supplied engines for British Rail, three car DMUs, troubles with broken cylinder head bolts eventually arose.

They should have been tightened to 175 lb ft. One of our service engineers visited a depot which was suffering badly with the problem of broken bolts found the fitters tightening the bolts by walking around the engine pushing a pole over the breaker bar on the socket!.

The habits of tightening large BSW fasteners died hard! 5/16 UNF was watchmaking to them.

Howard

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Nicholas Wheeler 1 on 24/11/2022 22:04:48:

[…]

I would suggest that pretty much anyone who doesn't use a torque wrench is overtightening common fasteners by a significant amount – the M6 bolts for K-series or Vauxhall V6 cam bearing caps only need 10Nm, which is easily applied with a couple of fingers on a 10mm spanner.

.

My introduction to such torque levels on cars came in 1972 with the lovely Hillman Imp [effectively the Coventry Climax fire-pump] engine in my secondhand Singer Chamois.

… which justified my purchase of a low-range torque wrench.

MichaelG.

Edited By Michael Gilligan on 25/11/2022 08:35:22

[Nick WheelerParticipant @nickwheeler]

Posted by Michael Gilligan on 25/11/2022 08:31:43:

Posted by Nicholas Wheeler 1 on 24/11/2022 22:04:48:

[…]

I would suggest that pretty much anyone who doesn't use a torque wrench is overtightening common fasteners by a significant amount – the M6 bolts for K-series or Vauxhall V6 cam bearing caps only need 10Nm, which is easily applied with a couple of fingers on a 10mm spanner.

.

My introduction to such torque levels on cars came in 1972 with the lovely Hillman Imp [effectively the Coventry Climax fire-pump] engine in my secondhand Singer Chamois.

… which justified my purchase of a low-range torque wrench.

Those two are why I have a 1/4" torque wrench. Using it for the first time was an eye-opener at just how little effort is required to torque such small fasteners correctly.

I did refit an XK cylinder head using a 3/8" TW, which was much harder work than if I'd used the 1/2" one I had at home.

Anyone with experience of LT/Sprinter vans from twenty years ago will know the consequences of not following the exact procedure for torquing the wheel bolts: the rear wheels fall off! It was a pity there wasn't a similar procedure to keep the side door attached to the van.

[Clive FosterParticipant @clivefoster55965]

Further to Howards last comment I put 5/16 UNF into the bolt stress calculator spreadsheet from Security Locknut at the link given in my earlier post.

For a 2" long grade 8 bolt it reckons 27 flt lb on a dry joint, 20 ft lb lubricated, is enough to reach maximum safe stress on the bolt of 120,000 psi. Takes 52 degrees rotation angle from full contact, finger tight I guess, to full torque.

Bolt stretches by 6 thou.

Gives a clamp load of 5,200 lb per square inch. Equivalent to about 2 tons on my thumb nail. Ouch.

Bit more playing around suggests that tightening to nominal torque / maximum safe stretch is 25° per inch from finger tight.

Frankly its all a bit scary how easy it is to generate, ahem, very large stresses in a bolt with correspondingly high clamp loads.

Clive

[duncan webster 1Participant @duncanwebster1]

Posted by Michael Gilligan on 25/11/2022 08:31:43:

Posted by Nicholas Wheeler 1 on 24/11/2022 22:04:48:

[…]

.

My introduction to such torque levels on cars came in 1972 with the lovely Hillman Imp [effectively the Coventry Climax fire-pump] engine in my secondhand Singer Chamois.

… which justified my purchase of a low-range torque wrench.

MichaelG.

Perhaps they put the good engines in the Chamois. When I met her, SWMBO had an Imp. I spent a lot of time underneath it. Great when it was going, but suffered from lack of development, particularly overheating.

[duncan webster 1Participant @duncanwebster1]

Posted by Michael Gilligan on 25/11/2022 08:31:43:

Posted by Nicholas Wheeler 1 on 24/11/2022 22:04:48:

[…]

.

My introduction to such torque levels on cars came in 1972 with the lovely Hillman Imp [effectively the Coventry Climax fire-pump] engine in my secondhand Singer Chamois.

… which justified my purchase of a low-range torque wrench.

MichaelG.

Perhaps they put the good engines in the Chamois. When I met her, SWMBO had an Imp. I spent a lot of time underneath it. Great when it was going, but suffered from lack of development, particularly overheating.

[A SmithParticipant @asmith78105]

Interesting discussion. In the early 70s, I cured a rattle from the o/s front wheel of my brother-in-law's Morris Traveller by taking the hub cap off and recovering the snapped of stud that had fallen victim to a pneumatic wrench. Changed the studs on all four wheels.

[Terry BParticipant @terryb]

Having started driving ambulances in 1966 until I retired in 2008 I drove several eras of ambulances. During my service I also trained ambulance drivers and instructors, . The only modifications I was aware of, apart from the electrics, were higher ratio back axles. The J1 Bedfords in service from the 1950s up to the introduction of Ford Transits and Bedford CFs had a 3.5 litre engine and were certainly fast once wound up and could approach 80mph on a straight road, faster than most family cars of the time. The early transits were quick with the initial acceleration, but lacked top speed and it was not until they put the V6 petrol engine in that the top speed improved. The last Transits in service when I retired were capable of 90 mph with good acceleration. The later vehicles with tail lifts were the first in general use to weigh over 3 tons, approaching 4 tons in weight with the equipment. I hope this answers some myths of the times, some one once told me that they had a reinforced front end in case they hit anything this was certainly a myth as you can see of any accidents involving ambulances.

[Howard LewisParticipant @howardlewis46836]

From memory, the torque to tighten 5/16 UNF bolts and setscrews (Without putting them into yield ) was 12-15 lb ft

Stretching memory even further, 3/8 UNF was something like 21 -25 lb ft

Until yield tightening was implemented, the W range 9/16 bolts were tightened to about 120 lb ft, but did not produce a reliably consistent clamp load. hence th change to yield tightening, and the end of our troubles.

Howard

[Clive FosterParticipant @clivefoster55965]

Howard

That torque calculator program is interesting to play with. Gotta say I'm going to pay a bit more attention to stretch in future.

Looks like your W range could have just about gotten away with grade 5 bolts, yield at 121 ft lb, rather than grade 8 or high tensile. Which seems very dangerous as you can bet someone will stuff a wrong un in. Always felt that half the art of design is outsmarting the things stupid people do or the mistakes normal folk under stress can make.

I do wonder how the relative inaccuracy of simple torquing as opposed to stretch measurements varies with absolute clamp load and bolt length. Instinctively you'd think the resistance to turning would go up disproportionately more with increasing clamp load so the higher the clamp load the less accurate torquing becomes. Bolts are effectively torsion springs so longer bots are inevitably more springy which probably makes torquing less accurate.

Presumably the inherent spring an increasing resistance at higher clamp loads is something that needs to be taken seriously when specifying the angle of turn when using the "turn X° from torque Y" method. I've never really understood how the "turn X° from torque Y" is any great advantage over a good torque wrench when tightening to clamp load. All the inherent variability seems to be the same except for swopping torque wrench accuracy for bolt torsional stiffness variation. Different with torque to yield where, effectively, the spring effect is overstressed and you need to be sure the bolt is firmly in the yield region.

Wonder if its different for nuts and studs. Objectively turning a bolt or tightening a nut down on a stud should be the same except for the thread length in a nut being constant as opposed to increasing as you screw a bolt in. Which difference should be more theoretical than real given (typically) half a turn or less from touch to tight and the end loading stress distribution that means only the outside few threads in a hole carry the load

But we all know that sometimes spinning the job works better than spinning a drill.

Clive

[Howard LewisParticipant @howardlewis46836]

The Rolls-Royce C range oil engines used aircraft practice (as you would imagine ) for the Big End bolts.

They were fully machined waisted bolts, and were tightened to produce a defined stretch.

Needless to say, we never had a rod come adrift!

Strangely, on other places on the engine, non critical fasteners, no torque or tightening instructions applied. It was left to the individual holding the spanner, or socket wrench!

[J HancockParticipant @jhancock95746]

NEVER let a tyre change outfit near a locking wheelnut , EVER.

[David George 1Participant @davidgeorge1]

I have a little useful book.

It has torque settings of many older vehicles cast and vans as the book is a little out of date but some very useful information.

David

 

 

 

Edited By David George 1 on 26/11/2022 15:29:51

[Sam Longley 1Participant @samlongley1]

Posted by Howard Lewis on 25/11/2022 21:00:23:

Until yield tightening was implemented, the W range 9/16 bolts were tightened to about 120 lb ft, but did not produce a reliably consistent clamp load. hence th change to yield tightening, and the end of our troubles.

Howard

I have been reading this thread with interested- seeing as I was the one that started it.

A couple of things have confused me & I may have missed them in the text.

I can grasp the point about striction when applying torque. I can understand the principle of rotating a bar a certain number of degrees rotation to acheive a certain torque- although the starting point may be a bit variable surely?

But I do not understand stretch- well I understand that a stud or bolt will extend in length as torque is applied. But how on earth can one measure that when it is buried in the clamped work piece?

The next point, & once again I apologise if I have missed it – What is yield tightening? I know that mild steel reinforcing rod can be placed under tension. It will give then stop giving, then actually become "stronger". That is part of the process of making high tensile reinforcing rod.

But how is this applied to a bolt or stud? Or is yield tightening something else?

Apologies for the ignorance but I am genuinely interested. I will ask about wheel nuts later.

Edited By Sam Longley 1 on 26/11/2022 15:43:31

[Nick WheelerParticipant @nickwheeler]

Posted by J Hancock on 26/11/2022 14:53:40:

NEVER let a tyre change outfit near a locking wheelnut , EVER.

Better advice is to never refit locking wheelnuts. They are all terrible and likely to cause you some major inconvenience at some point.

[Clive FosterParticipant @clivefoster55965]

Sam

There are some rather clever, and distinctly expensive, bolting systems with provision for measuring the stretch when the far end is buried. The simplest way being to drill the bolt almost all the way through so an internal micrometer or special gauge can be pushed down it. Some rather fascinating offerings for really big bolts that I don't understand. I guess when you are bolting a bridge or similar structure together its advisable to get things right.

Smaller things are generally arranged with both ends of the bolt accessible so direct measurement of stretch is easy. I believe Bristol were one of the earlier exponents of this on the bolted up crankpins for the master rod on their radial engines. With up to a couple of thousand hp transmitted via one crank pin its worth taking a bit of trouble.

In general bolts and similar steel rods have two yeild points. One where the extension starts to seriously diverge from the linear expectation of Hookes law followed by a second one at a little higher stress where the material finally gives up and becomes plastic. There is usually a short, flattish, section of the stretch curve between the two sections.

Stretch bolts are made from specially formulated and heat treated material to have a relatively long extension region after the initial yield with comparatively little change in the stress needed for further stretch. The bolt is still elastic in this region so will recover if you don't stretch it far enough to go into the third part of the curve up to the final plastic yield to breaking.

It won't recover past the initial yield which is why you can't, in practice re-use them. There are theoretically arcane methods of doing so but you'd be mad to try.

The extra angular turn needed when fitting stretch bolts is to ensure that the bolt is safely in the stretch zone and able to respond to changes without running out of elasticity.

For example the old all alloy push rod V8 engine in my Range Rover has two lengths of stretch bolts to hold the cylinder head on. The manual calls out 50° turn to apply the correct stretch, near enough 10 thou / 0.25 mm. The differential expansion rates of the steel bolts and alloy engine stretch the short bolts by about 5 thou / 0.12 mm and long ones by about 9.5 thou / 0.23 mm between cold and running temperature. Objectively the Rover engine runs the stretch bolts fairly close to the limits. It doesn't take much overheating to drive the bolts out of their recovery range.

Hence the reputation for head problems after overheating when serviced by folk who shouldn't be let loose on anything beyond a Morris Minor.

The oft touted use of non stretch studs (ARP et al) has its own issue. The differential expansion pretty much doubles the clamping forces on the gasket and pushes the stud very close to their yield stress. I'm less than convinced that accepting the very wide variation in clamping force on the head gasket is a better thing than the extra care needed with stretch bolts. Pushing hot running temperature up by 10°C and forcing fast warm up to meet emission regulations was not, in engineering terms, a good thing for the motor.

Clive

Edited By Clive Foster on 26/11/2022 16:54:32

[SamsarandaParticipant @samsaranda]

In my teenage years, like most youth of the 60’s, I was obsessed with motorcycles, I remember one particular vehicle that I owned was a 500 cc single cylinder Matchless. The big end failed and I had to strip the engine to replace it, when subsequently reassembling I think I was fitting the cylinder head and tightening the bolts with a ring spanner and suddenly one bolt sheared whilst I was tightening it, the top of the bolt took the spanner out of my hand and they both hit the garage ceiling as there was some considerable pent up force locked into the bolt when it let go. Needless to say that I didn’t possess or have access to the correct requisite torque wrench which might have saved the bolt from letting go in the manner that it did. Later in life I spent many years working on aircraft of all shapes and sizes, torque wrenches that were correctly calibrated became the norm, particularly when fitting the large main wheels on transport aircraft where the wheel bearing preload was critical to prevent embarrassing failures. Dave W

[David AmbroseParticipant @davidambrose86182]

Sulzer marine engines used to have plain nuts with Tommy bar holes on the main bearings. A hydraulic jacking device was screwed onto the end of the stud, over the nut, which was pumped up to a pressure that gave the right tension. The nut was then hand-tightened, and the pressure released.

[Tim StevensParticipant @timstevens64731]

Two memories from my (too brief) time at BSA in the late 1960s:

Their twins, and Triumph equivalents, had through big-end bolts in light alloy conrods, that were measured for length and tightened to a specified increase in length. The nuts were all-metal stiff nuts, (which vary a lot in initial stiffness) so this ensured that the applied stress was independent of nut-stiffness.

The big-end nuts (on the ends of the taper-ended crankpin) on the BSA Gold-star singles were tightened to 180 foot-pounds. This required a special crank-holding jig, bolted to a bench, which bolted to the wall – and a 4 foot tube over the end of the spanner.

Happy days

Cheers, Tim

[Nicholas FarrParticipant @nicholasfarr14254]

Hi, the biggest torque wrench I've used had a 1" drive and was about 6 ft long, the handle was about a third of the length and was hinged onto the business part, which would move once the setting was reached. It probably weighed about 3 Kg. I can't remember how high it would go, but probably 1000 or so NM. It was mostly used for torquing M24 12.9 high tensile hex head bolts with high tensile Nyloc nuts and hardened washers under both the head and nut, there were about 30 or so of these bolts holding two parts of a big rotary kiln together, but every so often, one or two would snap and had to be replaced as soon as! We had to put up a tower scaffold to replace these bolts and it wasn't much fun tightening them up at 650 NM. as one of us had to use the torque wrench and the other guy holding the other spanner to stop the bolt turning. Not something you really wanted to be dragged out of bed for at about 2 O'clock in the morning.

Regards Nick.

[HopperParticipant @hopper]

The studs holding down steam turbine top casings I worked on had a small hole drilled almost all the way down the centre for an extended dial indicator plunger to drop down into and measure the stretch while half a dozen of us pushed on a very long scaffold pole slipped over the spanner handle.

Initial tightening was done with a flogging spanner and 28-pound sledge hammer that came with its own 6'6" Albanian body-builder lad to swing it. I was the one who had to press my boot down on the flogging spanner to stop it flying away and hope his aim was as good as his weight lifting.

Edited By Hopper on 27/11/2022 02:25:56

[Howard LewisParticipant @howardlewis46836]

Yield tightening is when a fastener is tighten to just beyond the elastic limit,(begins to yield ) and takes a permanent extension. It is the most efficient use of the material.

The 32 spindle Unbrako machine that we used monitored torque and rotation, so that once the rate of torque increase vs angular rotation decreased, (i.e.no longer linear, the machine stopped ) from the intensive work that we had done, on every spindle, we knew that each bolt had increased in length by about 0.002"

This meant that each fastener now had a tensile load applied of 9 tons.

If you overtighten a fastener, you feel that the rotation is taking place, but then resistance (torque ) is no longer increasing at the same rate. You have loaded it beyond the elastic limit and put it into yield.

If you persist, it will fail. (As a youngster, I failed several 1/4 BSW bolts in this way.)

Monitoring the stretch in a fastener, as long as the material characteristics are known, provides an indication of the tensile load within that fastener.

The clamp load is determined by the coefficient of friction between the two members of the fastening (Delendent on the materials and the surface finish, and the depth of then thread)

A 1/2 x 40 ,tpi thread in brass will fail at a lower torque than a 1/2 UNF (20 tpi ) in even mild steel.

Howard

 

 

Edited By Howard Lewis on 27/11/2022 12:45:36

[Fowlers FuryParticipant @fowlersfury]

[duncan webster 1Participant @duncanwebster1]

I know it's not intuitive, but within sensible bounds tpi doesn't affect the tension torque relationship that much. The reason is that somewhere about 80% of the applied torque is absorbed in friction, and the finer the thread relative to the diameter the higher is this percentage. I doubt 1/2" 40 tpi is sensible bounds. I have a very erudite paper on this, if anyone wants a copy send me your email address via pm and I'll scan it.

[Author]

Posts