Doing some testing at work on fasteners in support of some new standards being written on fastener analysis. A somewhat popular item on ST is “what torque wrench to buy?” The test results so far just substantiate what I’ve always said about bikes and torque wrenches - save your $$$ and don’t buy a torque wrench.
Granted my current testing is on larger steel bolts (aerospace quality NAS fasteners from the same lot, not el cheapo bulk fasteners) and fasteners used on bikes are smaller, but the trends would be similar. 10 fasteners, 5 torque-tension tests on each bolt using a lab calibrated dial gauge wrench and 5 tests on each using an out-of-cal click wrench. First round of testing is without lubrication (lubrication will tend to decrease the error some). Bolt loads measured with the same load indicating washer which is currently in cal and has an error of much less than 1%.
The results? High uncertainty. Uncertainty for a single bolt isn’t great, but from bolt to bolt the variability is high. Bolt load can be expressed as F=Torque/bolt_diameter/K. I know F, Torque, and the diameter, so it’s easy to get K experimentally. The error is at least 30%. So suppose you have some nice carbon bits in your clamp stackup. Based on the uncertainty in your bolt, you could be anywhere from potentially slipping a seat post or handlebar to crushing the carbon (how do you know those bar or seat post designers were accounting for such a big error in the bolt load?).
Note, none of this is particularly earth shattering. Look at any set of test data for torque-preload of fasteners and you’ll see the same thing. High scatter unless you do things like add load indicators in your joint (not practical for bikes).
The best advice for bikes is to simply torque the crap out of the fastener but just not enough to either a) shear the head off or b) crush your clamp material. For bolts on composite components I torque just enough to prevent slipping. If it’s a new bar or seat post, I’ll bring my wrench along to make adjustments as needed. On all the rest I torque until it feels really tight. I certainly wouldn’t trust the torque callout on any 3rd party component, and even moreso with a fastener that didn’t come with the component due to unknown lube state and/or strength. Add to that that you don’t have any bolts on a bike I would consider preload critical except for the seat post and stem/bar (and only for slippage which you can test for) and you’ll find you won’t need a torque wrench.
Agreed, you’re going to get lots of variability from stiction. A question on methodology: did you just tighten the bolt until the wrench “clicked”, or did you do the backoff and retighten to “click” methodology.
I’d be interesting to hear what safety factor carbon parts manufacturers build into the clamping specs. I’m sure they have all measured at what point a stem or seatpost clamp would crush the carbon bar/post.
There is only 1 part on any of my bikes that I actually used a torque wrench on, and that is the center bolt for my Campagnolo Ultra Torque crankset on my road bike. The reason why is that the torque spec is way more than when it feels tight. I practically had to stand on it to get it to its torque spec.
How about the smaller, electronic torque wrenches (not the cheap needle pointer or click ones)? Can they be “trusted”? I use one and am pretty anal about torquing bolts (and I don’t really run any carbon pieces anymore except my steerer tube on my new bike. I do generally find that I need to be on the high side of the range (if given, i.e. 5-8 nm, I do 8 for sure).
I’m asking because I’ve had some major slipping issues on my cross bike when bombing hills/bumps, etc. Nice Thompson stem as well (so I trust the size/machining) but the crappy FSA bar may be not quite up to spec diameter? So many variables.
How about the smaller, electronic torque wrenches (not the cheap needle pointer or click ones)? Can they be “trusted”? I use one and am pretty anal about torquing bolts (and I don’t really run any carbon pieces anymore except my steerer tube on my new bike. I do generally find that I need to be on the high side of the range (if given, i.e. 5-8 nm, I do 8 for sure).
I’m asking because I’ve had some major slipping issues on my cross bike when bombing hills/bumps, etc. Nice Thompson stem as well (so I trust the size/machining) but the crappy FSA bar may be not quite up to spec diameter? So many variables.
No, the problem is not the torque delivery device. You could have a torque delivery device that’s good to .000001%. The major source of error is in the joint - the friction between bolt/nut threads. Now if you have a torque wrench that is off, it’s worse, but it’s error is relatively small in comparison with the other errors in the chain.
In your case, you say the torque call out is 5-8 Nm. If you took some generic fastener out of a box is it lubed, unlubed? You generally develop higher preload with lubed threads (less friction means better tightening), and on average it’s 30-50% more. And then are you accounting for running torque? Is your final torque with or without it, …
Everything you have said about fasteners and torque wrenches is true. It is all pretty basic fastener application that is taught in engineering school. Torque wrenches w/o lube is downright miserable, torque wrenches w/ lube is slightly better, torque to a spec and then turn a certain amount is moving in the right direction, measuring the actual bolt elongation is better still and measuring the actual pre-load w/ an instrumented washer is the golden standard. Having said that I think you are throwing the baby out with the bath water. I still use a torque wrench on my bike because it is orders of magnitude more repeatable than my hand by itself. It’s miserable but it is the best I can do w/ the tools I am willing to or can use.
I was just about to purchase a rather pricey t wrench…
So the general consensus seems to be lube, paste and feel. I suppose this would apply only to those with enough mechanical experience to actually have feel? (although the comment regarding Campy UT cranks makes good sense - but the farm Snap-on could handle this)
I was just about to purchase a rather pricey t wrench…
So the general consensus seems to be lube, paste and feel. I suppose this would apply only to those with enough mechanical experience to actually have feel? (although the comment regarding Campy UT cranks makes good sense - but the farm Snap-on could handle this)
Right. If you are PJ-Superman-Campbell, you won’t know your own strength and can break things.
While I understand your reasoning and recognize that a 30% variability coudl be rather high in many circumstances, I’ll have to defer to the classic SNL line “Jane you ignorant slut.”
Like the above poster mentioned torquing to ‘feel’ or to ‘tight’ requires a calibrated hand. While 30% variability from bolt to bolt can result in some fairly varied torquing, there is nothing like torquing a 5-8nm bolt to 20nm because you deem that to be ‘tight’. Not everybody has a good understanding about what 5nm feels like or what 15nm or 40nm etc, and there is nothing like crushing a $300 drop bar or $1500 aerobar because your hand thinks tight is 20nm. not to mention crushing a carbon steerer because you over tightend a stem.
I know, I know those could be extreme cases, and apply mostly to carbon bits, but with all the seatpost collars that we see cracked or sheered bolt heads I’ll be using my torque wrench to get me in the right ballpark. 30% variability is far better than 300% in my book.
I limit my torque wrench use to certain parts on my car where the repair manual makes a big stink about proper torque being required and even there I use it mostly just because it makes me feel cool (ie I’m a poser in the garage). I’ve been working on bikes since I got my first Stingray back in the early 70’s and I’ve never stripped a bolt or had anything fall apart because it was under tightened. But, I am a sucker for tools so if I had the cash to buy a really nice torque wrench suitable for bike work I’d probably get one.
Come to think of it, the only time I’ve had a critical fastener come loose was after the first time I used a torque wrench on my wheel lug nuts. Apparently, torquing to spec is not the same thing as making sure the wheel is properly seated on the hub, something that tends to happen automatically when you use the jump on the lug wrench method of torque input. The nuts were tight, just not tight against anything solid . . .
Yes, tried the Tacx paste - but that is for carbon parts mostly. Didn’t help. Changing stem/bar combo did. This was nice machined aluminum stem (Thompson) and an aluminum bar (FSA). I don’t run carbon ANYTHING on my cross bike (not even seat post, frame or spacer). I won’t run carbon bars/stem on the road bike either. Just frame/fork.
I didn’t use torque wrenches for bikes or cars until getting carbon bars. Since the price of experience for wrecking carbon bars is high, I bought my first torque wrench. As Ti T’war observes, ‘just enough’ is significantly more variable than a torque wrench…
I think the “takeaway” here should be to not torque to a specified level without some sort of lubrication on the threads and under the head…be it grease, anti-seize compound, or loctite (yes, loctite acts as a lubricant until it “sets”.)
That said…about the only bolt I consistently use a torque wrench on my bike is the crank arm retaining bolt on the S900/CinQo crankset when I swap it back and forth between bikes…
I think the “takeaway” here should be to not torque to a specified level without some sort of lubrication on the threads and under the head…be it grease, anti-seize compound, or loctite (yes, loctite acts as a lubricant until it “sets”.)
That said…about the only bolt I consistently use a torque wrench on my bike is the crank arm retaining bolt on the S900/CinQo crankset when I swap it back and forth between bikes…
hmmm…I’m not really sure that’s going to help things, is it? Lubricating the threads/bolt head while using a torque wrench is just going to result in higher actual forces on the bolt.
That’s the other reason I’d never use a torque wrench on bicycle parts: torque spec reading are very dependent on fastener lubrication (yet torque specs rarely/ever mention amount or type of thread lubrication).
The difference in lube vs. unlubed fastener torque can be more that 50%! Add in the variation which Tigermilk addressed, and I’m thinking that a well-lubed bolt, a torque wrench and a well-intentioned but clueless operator = busted up sh!t…
Everything you have said about fasteners and torque wrenches is true. It is all pretty basic fastener application that is taught in engineering school. Torque wrenches w/o lube is downright miserable, torque wrenches w/ lube is slightly better, torque to a spec and then turn a certain amount is moving in the right direction, measuring the actual bolt elongation is better still and measuring the actual pre-load w/ an instrumented washer is the golden standard. Having said that I think you are throwing the baby out with the bath water. I still use a torque wrench on my bike because it is orders of magnitude more repeatable than my hand by itself. It’s miserable but it is the best I can do w/ the tools I am willing to or can use.
Agreed. I’d rather be 30% off than 300% off, which is what I’ve seen many people do going by “feel.” There should be margins built around parts to get them within say 50%. I’ve been able to dial most of my parts in specifically to the torque wrench I use. So maybe your stem and seatpost slips the first time. Then you recheck. I agree that it’s not the ultimate precision, but I’ve gotten to the point now on all my bikes where I know what I need my torque wrench to read in order to not break a part and have it hold.