Torque to apply w/ Prevailing Torque nut. 2

[gt6racer2]

Having consulted the earlier posts (thread31-78159 and others ), I found a general concensus that the running torque of the nut before contact should be added to the desired clamp load torque.
I can understand this approach for a nyloc or such, however, I'm using a nut with defomation ( squashed out of round). As we get some elastic changes to the thread geometry on torquing, does this same rule still apply ?
BTW, I plan to make a study - just wondered if any of you have been here before ...

 

[MintJulep]

Yes it does.

 

[dvd]

You should buy the prevailing torque fastener standards from IFI if you are doing a study, there is some good information in them:

http://www.indfast.org/publications.php?cat=3

I also recommend that you have a talk with the fastener engineer at your prevailing-torque locknut supplier. This person should be familiar with your concerns. I don't know who you are buying your fasteners from but I have dealt with these folks:

http://www.emhart.com/products/grip/crownlock.asp

 

[HeviiGuy]

gt6racer2:

I assume that your assembly conditions are "perfect". By this I mean to say that you have very tight control over the friction of the assembly. If you have any doubt about the friction between the nut and fastener, between the nut and washer or between the washer and spotface, then, asking for a torque value is like asking somebody to whistle in the wind for you:

Torque cannot and should not be used in the context of a fastener's "tightness"!

This is to say that unless you are verifying fastener preload by some other manner (for example, by using strain gages or by measuring resultant elongation), the preload of your fastener and thus the reliability (and safety ?) of your joint is at risk.

Ciao,

HevïGuy

http://www.heviitech.com

 

[gt6racer2]

Heviguy,
Thanks for your comment. You hit on my big concern - I'm really not sure how much clamp I get for a given torque ( my "bolt" is a machined rod with rolled thread )- and the uncertainty over how to handle the running torque adds to the concern. (In addition, the running torque varies hugely). It's a high volume application, and I have no "feedback". My original question is actually to help me justify why we shouldn't be using the torque prevailing nut.
BTW - I'm recommending a change to std nut plus pre-applied loctite to remove some of the variation. I need some "backup" as the joint is complex ( 13 clamped pieces of various materials ) and hence has some risk of clamp load loss due to contaminant, part issues etc...

 

[HeviiGuy]

Good morning gt6racer2,

Indeed, feedback is necessary. It's surprising and quite dismaying to see so many critical bolted applications where the importance of feedback of some kind isn't considered, let alone even understood.

There are a number of ways to provide feedback ranging from, as I had mentioned in my earlier post, strain gaging to ultrasonic load verification to the use of "go, no-go" fasteners which change colour as clamp load changes. Strain gages are, by far, the most accurate way of acheiving this. Unfortunately, their use often isn't very practical as one can imagine. UT devices address this deficiency by allowing users to measure elongation (thus also measuring clamp load) either in a lab, on the production floor or, in the field. Where precise accuracy isn't required, the aforementioned load-indicating fasteners are an ideal compromise.

Food for thought!

Ciao,

HevïGuy

http://www.heviitech.com

 

[Screwman]

You may want to consider changing to a torque/angle system where you tighten to a relatively low snug-fit torque and then the gun is programed to rotate the nut a set number of degrees further to achieve the final assembly. This method removes a significant amount of the variation resulting from the frictional variation in the joint. It also enalbes a quick QC check: target angle and excess torque= excess friction; possible cross thread and lack of lube. Similar for a low torque reading.
The only disadvantage of torque/ angle is that it requires the use of fixtured drive guns and not hand held ones.

 

[dvd]

Skidmore-Wilhelm bolt tension calibrator is recommended to determine bolt tension as a function of torque. Call and talk to them.

http://www.skidmore-wilhelm.com/bolt_testing.htm

 

[HeviiGuy]

Wait a minute!!

Skidmore-Wilhelm "calibrators" or any other such devices are not suitable to determine the relationship betwen torque and tension in practical terms!

Whereas it can determine the tension/torque relationship based on one specific condition, this type of equipment cannot account for myriad variations in friction encountered in actual practice. For example, let's say that such a "calibrator" indicates preload "A" is acheived with a torque value of "B". It should be plain to see that if the friction in the joint is any greater than the friction was in the test fixture, torque "B" will not result in preload "A".

Indeed, these devices are good in limited cases (calibrating torque wrenches is one such case). However, to rely on thier output if one needs to know actual preload in critical applications as suggested is outright folly :-(







Ciao,

HevïGuy

http://www.heviitech.com

 

[dvd]

HeviiGuy:

I was working off of memory from AISC bolting instructions. I thought that was one of the recommended methods.

http://www.aisc.org/Template.cfm?Section=Engineering_FAQs&template=/FAQ/FAQListAll.cfm&FAQID=979

 

[HeviiGuy]

Hello dvd,

I apologize for the offense if inferred. I am sure that your intention was nothing less than genuine and well-meant. Your recomendation is understandable based on the wide prevalence of misleading documents such as the one from AISC.

When will industry and associated organizations stand-up and say "Enough is enough; we're through with ethereal engineering. The emperor has no clothes!"

Ciao,

HevïGuy

http://www.heviitech.com

 

[hogan666]

HeviiGuy ,
Would you happen to have a reference on a more appropriate document?

 

[HeviiGuy]

Hello hogan666,

Many such documents abound both on the internet and in hard copy. An oft-quoted tome is Bickford's "An Introduction to the Design and Behavior of Bolted Joints". Even within the vast posts of the Eng-Tips forums, there have been many astute comments and observations regarding this insipient yet all-too pervasive myth.

The crux of this matter is so painfully obvious:

The "torque" used to "tighten" a fastener is not a valid metric to describe said fastener's "tightness". Torque is simply a measurement of the resitance to force when a nut is turned down the fastener's thread!

If the friction in the joint is any less than anticipated, it's possible that the bolt could be stretched beyond yield. On the other hand, if the friction is greater than anticipated, less force is available to stretch the fastener. In this case the joint may be dangerously loose.

When one relies on torque alone, one must "guess" at what the joint's real-life friction factor will be. In so doing, everybody is forced to take a leap of faith by "guessing" whether the joint will be safe.

Woe is the engineer who will have to defend his "torque specification" (or even his use of misleading DTIs!) on the stand in the face of an enginering-astute litigator acting on behalf of someone seriously maimed or even killed as the result of a failed joint.

Ciao,

HevïGuy

http://www.heviitech.com

 

[hogan666]

HeviiGuy,
I do understand the difficulty in attaining any meaningful tension from a torque reading. I guess what I'm asking is what would you recommend be performed in addition to a rotational capacity test?

 

[HeviiGuy]

Hello again hogan666,

Bolt load can only be verified after the fastener has been tightened. It's impossible to verify load prior to the act. Yet, when asked to rely on "calibrated torque wrenches" or the result of "torque/tension calibrators" this is what we're told to do.

One method of definitively verifying load is through the use of strain gauges. Granted, this is highly impractical in the field, let alone on the production floor or sometimes even in the lab. Luckily, there is another method: Ultrasonic measurement of a bolt's stretch. Since bolt elongation is directly proportional to the bolt's stress, it provides a quantifiable and accurate verification of load. In trained hands, it's a very effective process for a wide range of fastener sizes and most configurations. Since the UT transducer is only placed on one end of a fastener, "through" and "blind hole" applications can be measured in this manner.

This isn't "new technology". We've used it for over 15 years in the field.





Ciao,

HevïGuy

http://www.heviitech.com

 

[hogan666]

HeviiGuy,
OK I can see your point on UT. But this technique is also limited. Due to the length/diameter of the bolt you will have issues with mode conversion. Not to mention the need for a unique focused, and most likely high frequency, transducer. This, as you have mentioned with other testing methods, would not be practical in the field. IMO. 3/4"x6" sure, but 3/8"x6 that would be tough.

 

[MintJulep]

HeviiGuy,

I'm no fan of torque control as an indicator of fastener tightness - and have highlighted the potential flaws associated with its use many time here.

However, there are many joints that can safely tolerate a very wide variation in fastener preload and still function as required. In these cases torque control is usually "good enough".

Part of proper joint design is an assessment of how much variation in preload the joint can tolerate, and following on that, proper production and assembly process development must determine a method of preload control that can consistently achieve preload within the acceptable range.

So let's take a step back and ask GT6racer: "How much preload do you need, and what variation is acceptable."

From there we can get into what methods of control might be able to achieve the necessary goal.

 

[CoryPad]

Using ultrasonics on a 3/8" diameter bolt that is 6" long? In the field? That is child's play. Equipment to do this is available for purchase (StressTel/GE, MicroControls), or the service is available (Archetype Joint).

Regards,

Cory

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

 

[HeviiGuy]

While not exactly "child's play" ;-), UT is indeed practical in the field. Even at this relatively small size.

Although our focus is on relative behemoth bolts, we've been involved with quite tiny fasteners as well. It does get a bit tricky: one has to know exactly how to set things up and what to look for. As long as a properly-sized transducer (physical and frequency) is chosen, the diameter isn't usually the major factor but rather, the length is. Actually, to be more specific, the effective grip length is important. With very short grip lengths, the resultant elongation of the fastener would be low. In cases such as this, the resolution of the equipment becomes very critical. This is where experience becomes a key point in interpreting accurate data.

Indeed, as MintJulep states, some joints can tolerate a degree of variation in actual preload. In these cases, load verification such as this is overkill. In other cases where the potential cost of joint failure is high, it only makes sense to verify that the bolt has been tightened to the required level. It's all about risk analysis!

Ciao,

HevïGuy

http://www.heviitech.com

 

[gt6racer2]

Thanks for all the comments - although a lttle off from my original question of how to account for the effect of a torque prevailing feature.
To Mintjulep - just for interest - I would like a given preload with no variation, as the end function of the parts I'm clamping ( a stack of valve shims to a piston )is affected directly by the clamp load. Of course that's not so practical, especially given I'm assembling at a rate of around 20/minute. So it's a more a case of optimising value (performance/cost) - and hence my interest in the effect of the torque prevailing feature vs std nut and loctite. We are currently using torque/angle control.