Checking Bolt Torque 2

[jimo14]

What/How is the best way to check a bolt's torque after it has been tightened? Is it to 'break it loose' and read the peak torque value or to tighten the bolt even more and record the value after it first moves?

 

[TheBlacksmith]

I would rather loosen all the bolts and retorque them according to the manufacturer's specs. The torque to loosen a bolt may not correlate very well to how tight it was, either due to friction or the fact that a lot of fasteners loosen very quickly. Turning the bolt farther may result in overtightening the bolt, resulting in bolt or substrate failure. Finally, more manufacturers are specifying "angle torquing", where the fasteners are tightening just enough to draw up the parts, then turned an additional 180-270 degrees.

Blacksmith

 

[ivymike]

I had a discussion of this subject over dinner one time with some engr friends of mine who worked in production at an engine manufacturer. The only thing I remember is that torque to tighten isn't the same as torque to loosen a bolt, which made it very difficult to verify reliability and repeatability of a bolt tightening machine. Sorry, I can't remember how they got around the problem.

I'm pretty sure that additional tightening would be the wrong way to go, if the bolt torque is near a limit.

 

[Hush]

To get a true reading you would have to tighten the bolt more until you overcome the break away torque (it starts moving) then quickly take your reading. But as Blacksmith points out this probably wouldn't be a very good idea. ALso if this is an existing install that has been sitting for awhile there are other factors that would throw your reading out to lunch, lubrication breakdown, oxidation creep, to name a few.

 

[brad]

Ivymike's statement is correct. I remember this discussion in manufacturing processes class.

The real answer (as I recall) to the original is--there is no way. As ivymike states, the torque to loosen is almost always different than the torque to tighten.

When torque is specified, that is the final tightened torque. If you want to achieve this torque, you either need to use a torque wrench (at least for the last turn of the installation) or else use a calibrated pneumatic wrench (which gives you the calibrated torque automatically).

That's my recollection (but then, I could be wrong, it's been awhile since I've had that particular class).

Brad

 

[Buhrmann]

Theory predicts that the loosening torque must always be less than the tightening torque.

But what about marking the bolt head angle, then losening it and retightening it to the same position it were before you loosened it. (That is if the bolts were not plastically stretched during the first installation like what TheBlacksmith mentioned)

If the friction coef between the nut and flange, and on the threads remain the same between the first tighten and the measurement, then it should be reasonable accurate.

Friction is sadly something that is not very reliable, especially with time.

Teo

 

[jackboot]

I am offering this- but not as an advertizement to the company who makes these bolts. However, they work and work well.

http://www.vfbolts.com

The old formula T=K*(Fi)*d (T=torque; K=coef. of friction; Fi= preload on bolt; d= nominal diam of bolt)works well- as long as you know exactly what K is.

In my experience: torque wrenches are inaccurate and it is hard to determine K accurately.

So don't use torque wrenches. Use preload by either measuring the bolt before and after make-up (strain calculations) or by using one of these bolts on the website.

The Maxbolt at the website has a built-in strain gage. You can monitor the preload of the bolt anytime-just look at the little indicator. If your equipment warrents this type of bolt you will find it well worth the cost.

 

[djtim]

The gist of what has been written is correct. I will add a further complication for you. With the usual caveat (the design of the specific joint) many joints will go through a relaxation so that the load you created by the original tightening of the bolt will be reduced at some time later. There are many reasons for this and I will leave that as an exercise for you. At my company when we torque critical fasteners we actually do a 2 stage process. The first is to torque to a value near the final value. (This value must correctly chosen.) Then we wait some period of time (an hour is not unusual) and retorque with the second torque value being the final desired torque. Can you truly know what the load is with standard hardware? Nope.. you have to get fancy as mentioned in the one post. Not practical for most situations.

 

[WKTaylor]

Folks... from an aero's perspective:

In large/critical structures we typically specify the following:

Install temporary fasteners or clamps to pull structure TIGHTLY together.

Install all threaded fasteners "finger tight".

Install rivets [and swagged-collar lock-bolts] undriven, between temp fasteners/clamps. NOTE: tape-over heads to prevent pop-out, allowing rivet to stabilize hole.

Torque To final value [or frangible-collar-break] in (3) stages [50%, 75% and 100%, MIMIMUM] using a typical crossing pattern (perhaps the term "criss-cross pattern" is more "friendly/descriptive&quot.

Note: wait at least 10-miuntes between torquing stages for joint settling and any "sealant squeeze-out].

NOTE: FINAL-TORQUE value is determined by adding the "free-running-torque" of the nut on the male threads to the theoretical required torque value [shear or tension from charts for each thread style]... resulting in a value that will be different for each fastener combination: especially with self-locking hardware!!!

Drive rivets & swagged-collar lock-bolts AFTER securing all threaded fasteners.

Fillet-seal around edges of nuts/washers and structure using sealant-squeeze-out; or apply tamper-evident paint across thread, nut, washer and structure [evidence of movement]. If necessary [IE: vibration environments], we also use cotter-pins and/or lock-wire.

NOTE: for some joints where installed torques are deliberately low but critical [IE: bearings on shafts, small-headed bolts, etc], we allow a "slight over-torque" to settle parts, wait a short-while, then back-off and final-torque to value. A theme-and-variation to this is "over-torque and back-off slightly concept" is to back-off until a thru-hole in male threads aligns to a "nearest" castelated cut-out in the nut... then install a cotter-pin or lock-wire.

Note: The concept of "checking torque" in a joint is alien... we use indicator mechanisms and inspections to verify that a "job" is finished.... and indicators that joint may have become loosen in-service.

Regards, Wil Taylor

 

[electricuwe]

The important issue for a bolted joint is usually not the torque but the force in the bolt. If a reliable bolted joint has to be manufactured one method is to apply a lubricant on the thread and below the head of the bolt so that friction is lower and tightening torque and force correlate more reliable.

For very critical bolted joints (e.g. in the automotive industry) the following method is used:

During tightening of the bolt the angle of rotation of the head and the torque are monitored electronicly. The automatic wrench stops, when the system indicates that the bolt reaches the limit of elastic stretching (The German term for this is "Streckgrenzen-gesteuertes Anziehen" but unfortunalty I don not now the English term)

 

[Hush]

I’ve used the following method on occasion to determine required torques and friction coefficients for different thread lubes:

Drill a stepped hole down the centre of the bolt. Measure the length from the step to the bottom of the hole. Apply a set torque and measure the change in length to determine your bolt load. Of course this doesn’t provide any help with the original question.

 

[gb420917]

The two ways we tighten large bolts is :-
Angle of turn after all bolts are tightened with an initial preload (torque).

By extension meaturements.

This isn't an advertisement for a web site but it has some good articles on bolting and some tutorials.

http://www.boltscience.com

 

[sprintcar]

We found a torque calculator here

http://www.futek.com

go to their Bolt Calculator
Not all the info is there, but if you email them with exact specs, they will add it for you

 

[electricuwe]

From the site

http://www.boltscience.com

I have learned that the method I have described in my post dated Feb 8th is called Yield Controlled Tightening.

 

[alamin]

Check out this website:

http://www.rotabolt.com

 

[nashjp]

Some methods to check current torque-

1. Return to mark check - Mark the bolt and abutment with a reference line, then loosen the bolt, retighten the bolt up to the reference line again while monitoring torque. You want the dynamic torque and not the static. (This method has pretty good correlation with the original tightening torque unless corrosion or galling have taken place)

2. Snug check - While recording torque (you need an instrumented torque transducer and recorder) tighten the bolt 3-8 more degrees. Look at the recorded data and pick off the valley after the original static break away torque. (This is easier to do if you also record turn angle and plot turn angle vs bolt torque.) The additional 3-8 degrees will not effect the bolt joint in a negative manner.

3. Ultrasonics - Measures bolt stretch. - Measure the current installed stretch then remove, measure stretch again. (special preparation of the bolt head may be necessary - this method can be very accurate, or frustrating and @#$ depending who does it.)

As the other posts stated, static breakaway torque usually has no correlation with tightening torque. Also if you are seeing what value was used on original assembly, joint setting (which will depend on the joint design) will probably provide a lower torque value unless you have a good solid abutment.

Good Luck,

 

[diamondjim]

Because torque is such a poor indication
of the tension that is in the bolt, you
really should be using more sophisticated
methods that measure tension and not torque.
Turn of the nut method in reverse would give
you some indication of how tight the bolts
were. Ultrasonic method would tell you more
exactly what the tension is. Do you need
exact information for some type of failure
analysis or is this just a curiosity? With
the turn of the nut method, you would have to
match mark the bolt and part, loosen it, take
it again to snug position and determine the
angle that it would take to take it to the
original matched mark points.

 

[mechconst]

I'm not sure what the application is, but I run across this in the field all the time in structural bolt ups, using high strenght bolts: A325 and above, 5/8" dia. and larger. Other methods are: 1) use load indicating washers. They have a series of raised points on the washer that flaten out when the bolts at the right tension. 2) Use a Skidmore device to check and see what torque is required to achieve the desired tension in the bolt. This is a bench device that you use to test about 3 or 4 of the bolts from the batch and develop an average torque value, since the torque will vary depending on how a bolt was machined and how much oil is left on the threads among other things. I think the Skidmore works on compressing a hydraulic cylinder. 3) Use twist off bolts. They actually have an extended part on the bolt that will twist off at the correct torque. It requires a special tool. Hope this helps.

 

[rmp319]

Find the Ninth Edition of AISC Manual of Steel Contruction and look in Section 5 for Bolds, Threaded Parts and Rivets. This manual offers information on Torque, washers, tension etc. Mostly geared to Structural Steel but has a wealth of information for general bolting requirements.

 

[boo1]

Mechconst the Skidmore-Wilhelm Bolt Tension Calibrator is the most commonly used bolt calibration device in the US. The device is a hydraulic load cell. Tightening the bolt creates pressure in the unit. The inter-changeable bushing in the back of the Skidmore sits in a piston. Between the piston and the body of the unit is hydraulic fluid which is compressed by the tension in the bolt. This hydraulic pressure is then read on a dial gage, scaled to read in terms of bolt tension in thousands of pounds (kips), rather than psi of hydraulic pressure.

Because of the slight movement of the piston, the unit underestimates the actual tension that would be put into a bolt for a given rotation. Some of the rotation is taken up by piston movement. In the structure, the steel does not compress and all rotation is used to tighten the bolt.