loosening torque

A GA to Tornado and a raspberry for beating me to it.

Although there are bolts (head, lug) where you know you will remove them and want to be able to do so, in each case you should take extra care to make that possible. A bolt such as in a framed connection in steel should, by default, be considered ∞.

I would think about the same as the tightening torque (perhaps a bit less) if it has been tightened recently and there's no corrosion etc. If it's been on for years in a bad environment it's anybody's guess - the bolt might break before it turns.

Cheers...........Codey

Under normal conditions it is never less since when bolt is tightened the friction is a movement (sliding) friction and when one tries to get it loose the friction is static one and static friction (after a longer time it even sticks together) is always greater than movement friction.

It will always be more to start the loosening process, physics being what they are.

Under "normal" conditions the unscrewing torque is depending on the materials in contact between 1 and 1.3 the torquing value. Value depends on time lag between torquing and loosening action. For this reason it is forbidden to estimate the assembly torque by the unscrewing value. However many car manufacturers do it.

If corroded a bolt can ask for 2x torquing to get free so that the risk to break it quite high.

A bit off topic, but still about torque.

I have several times advised folks in the antique vehicle hobby about main and rod bearing fastener torque. In the old vehicles (machinery too) before torque values were established, and before fastener grades were used, a torque value is still wanted for redoing bearing fasteners after they have been disassembled. That is, the torque value originally used is totally unknown. I have suggested that, before any disassembly, they break the fastener loose some small fraction of a turn--that is maybe 1/8 turn; then read the torque required to tighten the fastener back to where it was. (This, of course, means that the fastener has to be marked before anything is done so the amount of movement can be seen.) This, then, is the torque to be used when reassembling at a later time.

I have also used the torque value from the charts using Grade 2 fasteners, figuring the originals are approximately that strength.

Your suggestion is RIGHT with a "but", since the bolt was tightened a long time ago the tribologic conditions are not the same as at the tightening itself. This means that the torque which will be applied could be wrong.

I suppose ultimately the loosening torque that is acceptable is the one that does not break the bolt or does not break the rest of the object.

What you are asking for is perhaps "what is the testing torque that can be applied so that the bolt does not untighten when I try to test to see if it is tightened enough" or is it "what is the torque that can be applied so that the bolt untightens but does not break the assembly".

Thanks for asking a great question.

As one who has often battled with vendors on the question of proper torquing methods and calibration, these thoughts add greatly to My bank of knowledge and ability to ability to ensure compliance with standards and specifications.

Thanks to the 2 members who presented interesting answers also.

The wealth of knowledge and yes humor at times makes cr4 the best online forum for tinkerers and engineers today.

A very excellent question. The loosening torque will always be more than the tightening torque as long as the correct tightening torque was originally applied. If corrosion is present, the torque required to loosen, may exceed the yield of the bolt and shear off. It would be beneficial to know what the loosening torque is so that shearing doesn't happen. If the bolt doesn't loosen after the torque has been applied, some means of "helping" the bolt can be tried, such as heating or penetrating oil.

The quality of the job done in setting the fastening torque as well as the quality, material and physical type of fastener and washers used as well as the torque tool, AND the parts joined, will play a roll in the loosening torque.

What scares me is that all to often an engineer has gone to great trouble to calculate and specify the tolerances of flatness and fit of a joint and specked a bolt type but when assembly comes around no-one bothers too much about things like:

Clean faces of mated parts

Partly cured epoxy coating

Poor quality threads on bolt & nut

Using inappropriate Tools like pneumatic impact wrenches

Using torque wrenches that have not been calibrated correctly

Not cleaning nuts and bolts

Not applying high grade lubricants DESIGNED for the bolt material, like using coppaslip on stainless, or no lube at all. Then when the nut galls thinking they have hit the torque load required.

Not applying a service regime in line with the type of application that sets time frames for re-checking the torque on fasteners after initial assembly and intro to service/use. Eg after 1 hour run, 1day run, 1week run...... RE-TORQUE to specification....ALL the bolts.

Thus back to your question - any deviation of engineered specification will result in the obvious deviation from the initial fastening torque, either reduced or increased. A well designed AND executed torquing of a joint should result in a very similar loosening torque +/- 1%. Any thing outside that range would indicate a failure of design or application of design parameter/s. The greater that deviation is the more concern to correct the situation. One thing I learned in aviation is that ignoring such things is a dangerous ego trip that......'ll bring you down to earth with a bump! (my bad pun for today)

PS I am not an aeronautical engineer.

Hi Bushdriver, What you mention is the lack of understanding quality requirements and even more lack of training for the operators. Less training = less quality = more risk Unfortunately management, all over the world, underestimate bolting problems and do not understand the effect of correct training. They believe that for such a "low complexity " technique cost of training can be saved! How much it does cost in after-sales nobody quantifies !

My original post was meant to be facetious, but not entirely so.

As others have pointed out, loosening torque should be reasonably related to the earlier tightening torque. On assemblies that are regularly taken apart, this may well be the case.

However, on assemblies that remain attached for a long time, especially in adverse environments, bolts/nuts are apt to become "frozen." In some of those cases, ya do what ya gotta do to get them apart. Penetrating oil, heat, hammering the nut from the side to relieve the engaged threads; then cutting/gouging. Making a bolt unreusable is better than scrapping a whole assembly that won't come apart.

Some years back one of our customers suggested we install pre-calibrated torque wrenches of all sizes throughout our assembly area; basically one for every possible fastener, a set for every assembler. While there was no quality problem leading the request, it was part of an ISO-9000 implementation wherein everything seemed to need a procedure. The costs, both in terms of original install and subsequent bi-annual calibrations would have been prohibitive, especially in light of the fact that we were solving no particular problem.

Instead I set about designing a test intending to corelate the 'break-away' torque with the initial fastening torque. This would allow a) inspecting/auditing that fasteners had indeed been installed properly (i.e., within an acceptable range of 'too little' to 'too tight') and b) that my experienced (i.e.; 'certified') assemblers didn't need torque wrenches to accomplish their jobs.

My engineer devised an aluminum block with three-each of four separate fasteners sizes that covered 95% of the fasteners we used and set about collecting data. Each set of three identical fasteners was torqued (all dry) to the same value and of course, that value was different for each set. I knew the breakaway torque would be higher than the applied torque but I was unprepared to see how wide the range of break-away torques was. A fastener initially torqued to say 30 in-lb. might require anywhere from 35 to 55 in-lb. to break it free. It also didn't seem to matter whether it was the same fastener in the same hole or if I swapped things around, the variability never improved. We chaulked it up to variability in the fasteners and threaded holes, as well as when and where exactly in the torquing did a specific thread begin to compress against the mating thread in the hole.

While I ultimately installed the procedure I was disappointed that a clearer corelation hadn't been identified and that so much variability existed. At the time I remember researching and finding that Honda was using specially designed, and VERY expensive (for the time), torque-wrenches (for its cylinder heads and connecting rods) that incorporated ultra-sonics to measure the exact elongation of a torqued fastener which, of course, speaks much more accurately to the compressive strength a torqued fastener is exerting. Honda, as had we, had learned that when it really matters, 'applied torque' during the tightening process is no where near a good enough indicator of clamp strength. And we had learned that break-away torque is an even worse indicator of initally applied torque.

This might improve with a better class of fastener and thread so even my experiences need be viewed cautiously. Ultimately, find a way to measure bolt elongation if it really matters!

I am very glad that you reported your experience with the dispersion of the break-away torque.

This is in fact a well known aspect and as I mentioned it is basically forbidden to use this procedure for a quality check. I made myself such measurements and based on them I gave the 1.3x tightening torque as limit. You came to a maximal ration of 55/30=1.83 which could be the consequence of used materials, loading amplitude of contact surfaces and the fact that the assembly was made "dry". In real life there are slight traces of "lubricants" and those reduce the width of the dispersion.

I am an electrician, not a mechanical engineer and did not realize there is a difference between tightening torque and loosening torque. My problem is that it took me 45 ft/lbs to loosen an oil drain plug on a subaru forester. l tightened it to 45 but should have been 31. By using your 1.3 multiplier, it will take me @ 60 to loosen. Do you think the bolt can handle this after 5000 miles or should l drain it down and replace crushable washer as soon as posible? Thanks.

If the manufacturer specified a torque it has to be respected since it is based on many tests.

The breakaway torque depends as mentioned before on several factors one of them being the time the bolt was under tension in the assembly and which were the applied loads.

Friction is an unpredictable phenomenon which depends on the materials, the surfaces roughness, the assembly lubricant (if any), the sliding speed at assembly, the time the 2 surfaces stay in contact and of course under which pressure and how variable it was during service time.

Even if the friction is the same the breakaway torque is not equal to the tightening one, it depends on the pitch/mean diameter value. For hard surfaces and high quality bolts repetitivity is quite good but not for soft parts since the pressures can go over the yield limit for the weaker material and the tribological conditions are totally changed.

This very complex phenomenon is the reason manufacturer specs have not to be changed.

I wouldn't expect any problem. It will be a bit harder to remove than if you'd tightened it to 31 ft.lb (not ft/lb BTW), but I'm sure you're strong enough. It's not a critical application, the plug is only holding itself in, not taking any loads.

By all means use a torque wrench if you're not confident to do it by what feels right, but I've never used one for a drain plug and I doubt the average mechanic does. In any case we are told that torque isn't a good indication of bolt load, bolt extension should be used, and that isn't practical for a drain plug.

Thanks Codemaster.

Interesting, it's good to hear some results from an actual set of tests.

Codey

Hello Doob et al,

I usually don't frequent this forum; Your bolts are too small and it's usually a hot-bed of fervent Torque Luddites . Nevertheless, this topic caught my attention. In particular, your findings caught my attention.

Of course it's silly to try to correlate breakaway torque to applied torque. It's even sillier than trying to correlate applied torque to bolt load! "Torque", as I'm fond of saying, isn't a measure of bolt load; It's merely the resistance felt when trying to turn a bolt head (or nut). As you've so eloquently pointed out .

You also correctly mention that elongation measurement is the ultimate way to ascertain load but state that one would "need to find a way to measure bolt elongation". For those who are interested, here is a way to do just that... Bolt Elongation Measurement

It does rather depend on the amount and condition of corrosion between it and whatever it is screwed into. Also:

• A Nord-lock fastener takes more <...torque...> to undo than to do up, which is why they don't come off easily.
• A Nylock fastener takes more <...torque...> than a regular one.
• A castellated fastener will take more <...torque...> than a regular one.
• Some fasteners are fitted with split pins that need to be removed first.
• Etc.