What's new
What's new

Bolt snapped earlier than expected

Strostkovy

Titanium
Joined
Oct 29, 2017
I'm assembling a project and just got a torque wrench set (Tekton. Not great, not terrible), and set my 1/4" one to 17Nm (from charts online) for an M6 10.9 hex bolt with 16mm of thread engagement into a 3/4" aluminum plate with freshly tapped, oil free threads.

I had accidentally been tightening the handle, and when the bolt snapped I saw my wrench was up to 20.5Nm. It seems odd to me that not even 20% more torque was enough to snap the bolt. I will say even 17Nm feels tighter than I would go with a wrench. Luckily the threads were totally fine and it was a through hole so I just used a drill to thread the stub out the back.

I know that the actual bolt tension depends on a lot of factors, including material. Is there a better chart or calculation to determine what my torque values should be?
 
Alright, I'm trusting you.

(I say that as I set my torque wrench to 70 foot pounds for an M10 bolt. Oops)

I think that oops denotes the fact that you noticed that the 70 on the chart is in Nm, not Lb/Ft...right?

And those are just ballpark numbers. Adding lubricant to the threads and under the bolt head for instance will mean you need less torque for equal tension.

You might find this interesting:

https://www.fastenal.com/content/feds/pdf/Article - Bolted Joint Design.pdf
 
My "school of hard knocks" lesson in bolt torque was this -

I was installing a cylinder head on an air compressor. Took something like... I dunno, let's call it 50Ft-Lb (arbitrary). I ended up snapping about half the bolts (they were through holes, fortunately, otherwise, I would've stopped at the first one).


Turns out it was "50" ft-lb DRY. Lubricated, it was MUCH less. I didn't lubricate them intentionally. I had assembly lube on my hands from putting the pump together and wiped off the threads before I installed each bolt.

I ended up buying all new bolts, lubricating them, and torquing them to the appropriate LUBRICATED torque.

Hope this helps someone, so that they don't have to learn the hard way!

I hope this helps, Swarovski! :D

Sent using Morse code on - .- .--. .- - .- .-.. -.-
 
yes, it will snap. m6/10.9 torque is 13.2-16.5 nm (vdi 2230-2003). this is based on 90% ys. so 14.9 (lightly lubed)/0.9 = 16.5. uts/ys is 1000/940 mpa, corresponding to 17.5 nm. even starting with the dry value it should snap.
 
Last edited:
Well...I gotta say your torque wrench is kinda POS. But let's assume it's accurate.

You mighta had a defective fastener. It happens. If you told us you repeated that failure 5 times, I'd eliminate the fastener.

Dry thread are always asking for trouble when trying to achieve a set torque value. I've found that using 'proper' lube is critical...most people and most books understand that. But what most people and most books neglect is the issue of where to put the lube. Not only the threads, but also under the head of the fastener.

If you had a good fastener and lubed it properly and used a good wrench...I seriously doubt you'd have this issue.
 
I know people here like to do this.
So how much load, stretching power or tension/spring in lbs on a M6x1.0 bolt at 17 and 22 NM both lubed and dry. (it may have been a 0.75)
Let's ignore a flat washer under the head.
Bob
 
It was a brand new bolt. The aluminum part it was securing is powder coated. I know the friction on the head was fairly low because the washer spun on the powder coat and not under the bolt head.
 
well.. getting the proper value to start with is kinda important..:D

having said that, what was the origin/brand and history of the fastener? that's an ESSENTIAL piece of the puzzle. without that all other data is secondary.

also, its VITAL to understand torque can be a very, very poor determinant of preload on a fastener, as the above references to variation due to lubrication indicate. any critical application that can do so uses direct stretch measurement, such as con-rod bolts. of corse not every application can be checked that way, so a practical test with the actual material, fastener, lubrication, and method of tightening is recommended. torque to failure on a TEST piece, and proceed accordingly.

as I have said (over and over), please, please get the (bibles) books on fasteners, Carroll Smith's and Forbes Arid's and read them. a gold mine(s) that highlights some common mis-conceptions and myths. one that most machinists simply don't get, and are almost completely unable to accept, is how VASTY inferior the most perfect cut threads are to rolled threads.
 
well.. getting the proper value to start with is kinda important..:D

having said that, what was the origin/brand and history of the fastener? that's an ESSENTIAL piece of the puzzle. without that all other data is secondary.

also, its VITAL to understand torque can be a very, very poor determinant of preload on a fastener, as the above references to variation due to lubrication indicate. any critical application that can do so uses direct stretch measurement, such as con-rod bolts. of corse not every application can be checked that way, so a practical test with the actual material, fastener, lubrication, and method of tightening is recommended. torque to failure on a TEST piece, and proceed accordingly.

as I have said (over and over), please, please get the (bibles) books on fasteners, Carroll Smith's and Forbes Arid's and read them. a gold mine(s) that highlights some common mis-conceptions and myths. one that most machinists simply don't get, and are almost completely unable to accept, is how VASTY inferior the most perfect cut threads are to rolled threads.

All I know is that these bolts are from Fastenal. We buy a lot from them and don't have quality concerns, other than the odd malformed fastener.

Once I torque to failure, how much should I back off for my actual torque value? I thought there was a bit more safety margin than there is.

I'll get those books. This is my first adventure into getting better at engineering and manufacturing consistent assemblies.
 
10.9's are brittle like that, as the .9 suggests.

Programmed via Mazatrol
I gotcha. It's a 10.9, because it's just a Lil too brittle to turn it up to 11.

Right? :D [emoji869]

Sent using Morse code on - .- .--. .- - .- .-.. -.-
 
All I know is that these bolts are from Fastenal. We buy a lot from them and don't have quality concerns, other than the odd malformed fastener.

Once I torque to failure, how much should I back off for my actual torque value? I thought there was a bit more safety margin than there is.

I'll get those books. This is my first adventure into getting better at engineering and manufacturing consistent assemblies.

buying from a retailer like Fastenal, McMaster, or ANYONE for that mater, other than a trusted primary manufacturer like ARP or Holo-Krome (regrettably we can take SPS Unbreako off that list) is absolutely not adequate to control or assure the quality of fasteners.

as to the percent of failure to preload a fastener to, that depends on a host of engineering parameters such as Saftey factor, consequences of a failure, redundancy, ratio of yield to UTS, cost, weight penalty etc etc.

if you can, and have the test equipment for it, torque to yield on a "representative sample" (I.E. more than one), the point where more rotation fails to increase pinch or torque significantly (well right around there anyway), and 80% of that is a reasonable level of preload (I think?). the consistency of this value (yield) also gives a good indication of the quality of the batch.

bottom line and absolutely most important, never, ever think "oh I got it from Fastenal, it must be ok".


general good advise, unless its head bolts, crank bolts or aerospace, get stuff in a yellow and green box that says "Holo-Krome" made in usa, and be careful about how you maintain consistent friction characteristics of the assembly, torque to recommended values, and you are good.

P.S., the washer or other bearing surface the fastener is pinching is absolutely just as important as the fastener. the best bolt in the world, properly preloaded, is wasted if the washer is soft, poorly fitted, or otherwise deficient.
 
...
P.S., the washer or other bearing surface the fastener is pinching is absolutely just as important as the fastener. the best bolt in the world, properly preloaded, is wasted if the washer is soft, poorly fitted, or otherwise deficient.

As in: split lockwasher.
 
buying from a retailer like Fastenal, McMaster, or ANYONE for that mater, other than a trusted primary manufacturer like ARP or Holo-Krome (regrettably we can take SPS Unbreako off that list) is absolutely not adequate to control or assure the quality of fasteners.

as to the percent of failure to preload a fastener to, that depends on a host of engineering parameters such as Saftey factor, consequences of a failure, redundancy, ratio of yield to UTS, cost, weight penalty etc etc.

if you can, and have the test equipment for it, torque to yield on a "representative sample" (I.E. more than one), the point where more rotation fails to increase pinch or torque significantly (well right around there anyway), and 80% of that is a reasonable level of preload (I think?). the consistency of this value (yield) also gives a good indication of the quality of the batch.

bottom line and absolutely most important, never, ever think "oh I got it from Fastenal, it must be ok".


general good advise, unless its head bolts, crank bolts or aerospace, get stuff in a yellow and green box that says "Holo-Krome" made in usa, and be careful about how you maintain consistent friction characteristics of the assembly, torque to recommended values, and you are good.

P.S., the washer or other bearing surface the fastener is pinching is absolutely just as important as the fastener. the best bolt in the world, properly preloaded, is wasted if the washer is soft, poorly fitted, or otherwise deficient.

All good points but it makes the most sense for these sorts of projects to design it to use the reasonably priced, reasonably strong bolts I can get from Fastenal (I get a good discount since the business I work for buys a $20k in bolts per year).

I think I'll look into the standards they comply with to get the ideal tensile strength, and do some testing and/or calculations to determine what gets to that.

Honestly half the stuff I make could be held together with a threaded potato, but hey, I'm learneding.
 
Once I torque to failure, how much should I back off for my actual torque value? I thought there was a bit more safety margin than there is.

70% of failure torque is a very general rule of thumb. Dial-type torque wrenches with memory pointers are the best at capturing that peak torque, but you may be able to get close enough with a click-type. Good luck!
 








 
Back
Top