Retorquing the head?

Not open for further replies.
Apr 15, 2004
Country flag
My new engine has had its first heat cycle and now it's time to retorque the head, according to the running-in procedure I'm following.

I'm kind of new to this retorquing thing however, and I have to confess I'm not quite sure how you're supposed to do it. :oops: Should I:

a) loosen all the fasteners in sequence and retighten them in sequence?

b) loosen and then retighten each fastener one at a time?

c) just get out the torque wrench and make sure they're all torqued to spec?


d) blow the whole thing off? :twisted:

All most of us has ever done is go back and redo the cycle noting how far they turn as you go. don't forget the 5/16 studs underneth And although you will be sorly tempted don't over tighten them. Re lash the valves. Put on a hundred and do it again.
When re-torquing, sometimes it is preferable to slacken each bolt (or nut?) about a quarter of a turn first, as that shows the fastener isn't stuck or corroded and that you aren't trying to torque up a bolt that is unlikely to move, although that would apply more to fasteners that have been done up for some length of time, but for me it's = answer B.
Hi Deb
All i ever do is ride it to get it hot , let it cool down and then retorque in the correct sequence. Recheck valve clearance .Therefore the answer is C

Torque is properly measured ONLY when the bolt/nut is turning. Once the bolt/nut stops turning, it takes a significantly greater amount of torque to break it free from the static/stopped position. Therefore, you do NOT just put the torque wrench on the tightened bolt/nut and see that it does or doesn't move at the given torque spec. You must first loosen the nut/bolt slightly and then retighten to the given torque spec, one bolt/nut at a time. Do this again in another few hundred miles.
Thanks guys! Lots of good info. Where would I be without this forum? Fumbling around in the dark breaking things I suppose!

Just a few more comments on fastener torque:

Torque is the moment or distance to center times the force; it may be measured with or without the fastener rotating. What you’re after when you torque a fastener is preload or bolt stretch. Torque is just the method of measuring the bolt preload. Many factors affect torque and thus preload. The finish on all mating surfaces and the type of lubricant have the most dramatic affect on bolt preload.

It is good practice to check head bolt torque soon after installation of the head and gasket to ensure that bolt preload has not been lost, owing to gasket compression. Now, I believe the Weatherman’s concern is that when checking head bolt torque, there may be instances where torque is found to be per spec but in reality, bolt preload has been lost. This false reading may be due to a sticky “crust” that formed on the threads or head of the fastener, which increased torque for a given preload. So, to prevent this false preload reading, he recommends loosening the fastener slightly to break any crust and then re-tighten to spec.

I’ve re-torqued head bolts by loosening first and retightening and I’ve also re-torqued just by just tightening – both methods have worked for me.
With all due respect, the basis for my position is not related to a crust but is much more elemental. With most materials, the coefficient of static friction is higher than the coefficient of dynamic friction. That is, all else being equal, it takes more force to move an object from rest than it does to keep the now moving object moving. Here are some links that describe some of the fundamental physics of friction. ... 2.html#sta ... 2.html#plo

Therefore, the fastening torque placed on a threaded fastener can only be properly measured while the threaded fastener is in motion and not while it is static. This is why I prefer click-type torque wrenches for most general torquing applications; I can concentrate on maintaining a smooth, constant motion of the torque wrench until the pre-set click is heard and am not distracted from providing the smooth, constant motion to the torque wrench by visually trying to read a pointer and gauge type wrench.

And yes, torque is really not the end-all of what we are trying to accomplish; it is only a predictor of what we want. In torquing a head, we are really trying to provide a predetermined clamping force on the head with respect to the cylinder(s) to prevent leakage. If the head gasket has compressed, paint/powder coating on the head under the fastener has compressed, or other components have taken a set after one or more heat cycles, the clamping force exerted on the head by a fastener may have been reduced, even though a check of the torque on the static fastener shows that the fastener does not move when the specified tightening torque is applied to the fastener (due to the higher coefficient of static friction). However, by slightly loosening the fastener first so that the torque needed to move the fastener from rest is lower than the specified tightening torque for that fastener, then rotating the fastener in a smooth, constant motion until the specified tightening torque is reached, will show that the fastener has now been rotated beyond where it was in the first static position, thereby reestablishing the desired clamping force.

When using a threaded stud and nut combination, assuming no damage to the threaded stud or nut, the desired clamping force from each stud can also be determined by the preload, or stretch, of the stud. In some critical instances, such as connecting rod studs in racing engines, the correct tightness of the nut on the stud is not determined by measuring the tightening torque exerted on the nut, but by measuring the stretch of the stud. This is a more accurate determination of the clamping force exerted by that stud because it is less affected by such factors as threads that are buggered, rusty/corroded, dry/unlubricated, dirty, etc. that reduce the proportion of delivered/measured torque that is actually applied to rotating the nut. However, measurement of the stretch of a fastener is not a practical approach in many instances to determine the clamping force desired from that fastener.

Tightening torque specifications set by design engineers for various threaded fasteners are approximate predictors for obtaining the desired clamping force. When setting these torque ranges for various threaded fasteners, the engineers take into account the type, size and material of the fastener, the materials of the components being clamped, intervening washers, and other factors to determine what torque range most accurately predicts the desired clamping force from that fastener. That torque range becomes an inaccurate predictor of clamping force when the threads are buggered, rusty/corroded, dry/unlubricated, dirty or otherwise contaminated or the mating faces of the threaded fasteners, head or washers have become galled. On the other hand, the desired clamping force may never be reached if threads are on the verge of pulling out or if the bolt/stud is fatigued, cracked or has been stretched beyond elastic deformation to plastic deformation.

To summarize, to properly torque a fastener, make sure the threads are clean, and not deformed or contaminated. Chase or repair the threads if they are damaged. The threads should be lubricated or have anti-seize applied. Excepting for lock nuts, in most instances, the nut or bolt should be able to be tightened by finger up to the point of setting the torque if everything is in proper condition. If not, then a significant part of the specified tightening torque is being lost to friction and is not being used to properly set the desired clamping force. The torque should be set while the fastener is in a smooth, constant motion and the motion only stopped upon reaching the specified tightening torque. If a fastener needs to be retorqued, it should first be loosened to the point where the torque needed to again move it from rest is below the specified tightening torque, then retightened in a smooth, constant motion until the specified tightening torque is reached. This should be done one by one so that any seal between components is not lost by loosening all of the fasteners at once.

Certain high stress applications where the fastener is considered to be one use only may have different requirements, but that doesn’t apply to most Norton head fasteners.
How do most people torque up the nuts that have difficult access? It's difficult enough to find a spanner that will fit the long nuts under the exhaust ports that go in from the bottom and the other at the rear. let alone get a torque wrench in there.
The 'old way' (before most home mechanics had torque wrenches) was to use a spring balance to pull on a spanner/wrench and can still be used where there isn't room to use a normal torque wrench, although the distance between the fastener and balance needs to be known so that the amount of force applied to the spring balance can be corrected, unless of course the distance between the fastener and spring balance (if measuring with lbs?) is exactly 1 foot (if torquing in ft/lb?).

Torque = force x radius so a force of 1lb applied to a lever at a distance of 1 foot = 1ft/lb.
If the spring balance was positioned 6 inches along the lever then the amount of pulling force would need to be doubled, and at a radius of 2 feet the force required would need to be halved to give the same result.
So if you wanted to torque a fastener to, say, 20 ft/lb, and the distance between it and the spring balance was 6" then the spring balance would need to be pulled to a show a reading of 40lbs, and at a distance of 2 feet away the spring balance would need to show a reading of 10lbs.

Now every home mechanic has a torque wrench! So 'torquing up' has become a 'must do' job, however motorcyclists managed to do without them for 70-odd years!
Many instruction manuals and handbooks issued well into the 1950's do not contain much (if any) torque information as there wasn't much point putting it in!
Hey LAB – I’ve got a “fish scale” for torquing those hard to reach fasteners. I've gotten lazy and don't use it much anymore, however. Instead, I estimate the torque of those awkward fasteners by feel - just like the old days.

And as you’ve alluded to, a torque wrench is not much more accurate than using the "feel-method" to preload fasteners. The following information was extracted from the 26th Edition of Machinery’s Handbook:

Feel: +/- 35% accurate
Torque Wrench: +/- 25% accurate
Turn of Nut: +/- 15% accurate

Another interesting tidbit: Most, if not all, Norton fasteners were made up at the factory dry. So, the torque values published in the factory shop manual are probably based on un-lubricated fasteners.
Lots of good information on this thread.

Certainly in industry we always back off first when check-torquing although I think that when re-doing Commando heads, it is not so much a matter of checking than re-tightening.

In my experience, a newly fitted flame-ring gasket will always compress to the extent that the fasteners will turn 1/4 to 1/3 turn. In those circumstances, even if not backing off first, the fastener will be turning nicely before the desired figure is reached.

I am also sure that I have read that factory fittings were torqued dry but nothing would persuade me to omit my smear of copper grease.

I think that with some of the Commando fittings (clutch and alternator etc) the torque settings are useful to ensure that things are tight enough but the head torque fiqures are more to guard against over tightening and stripping.

I have changed head gaskets away from home using a short tommy bar on a 3/8" drive socket and a 6" long ring spanner for the nuts underneath without a problem. If the threads are in good condition, one would have to be a bit of an animal to strip anything using a short ring spanner and it's quite easy with those deep nuts to feel the threads starting to stretch.
Not open for further replies.