Lab torque test question 1

[teejer]

We recently had some bolts torque tested. The numbers came back 60% higher than the book value for a grade 5 bolt. The method that the lab used was as follows: They increased torque and measured the tension in the connection as the torque was applied. They stopped applying torque once they measured no increase in the tension as they rotated the head of the bolt. They said that this was the yield point for the connection. In all of my reading on torque testing for fasteners, I don't see any mention of this method. Is this a valid method and would this be very accurate? I am concerned since this bolt will be installed and removed many times in its life. Thanks.

 

[unclesyd]

It appears that they tested the bolt to yield instead of a specific bolt stress or clamp force. This is not a bad number to have just as long as you don't use it day to day operation.
All torque tables have to make an assumption as to what the basis of given values are. As stated it can be YS or UTS or a % thereof, bolt stress, i.e 50,000 psi, or a clamp force, i.e. 50,000 psi.

Comeback with your application and to what specification you working under.

 

[byrdj]

the "book" values I carry in my hard hat are per the desired Kpsi loading and thus vary greatly (20 to 60K). In general, most bolting I deal with is less than 50% of yeild loading. So Finding bolts torque to yeild being 60% greater than book makes sense (but still need to defined "book" to determine if correct)

 

[btrueblood]

"No increase in tension as the bolt head is rotated"

This implies that the connection is well beyond yielding of either the fastener or the female threaded part, and very near to ultimate load for one or the other half of the bolt/thread pair.

Yielding would be when the tension vs. rotation curve is offset by 0.2% or better from the approximate straight-line fit of the rest of the data.

 

[teejer]

The bolt is used with a lifting device that does fall under the ASME B30.26 standard, but the standard does not call for a specific torque requirement or test for the bolt. I am doing this not out of a specification or requirement, but I want to give the customer a torque value to work with in the field. The bolt and lifting device will be attached and removed many times in its life, which also complicates things and makes it all the more important to get a value right.

I oversimplified my explanation earlier. The bolt is a 5/8" grade 5 bolt that threads into a 304 stainless block. The listed yield strength I have for a grade 5 bolt is 92,000 psi. 92,000 psi*.25*pi*(.625^2) = 28,225 lbs in tension theoretically for the bolt to yield. The lab reported yield beginning at 16,000 lbs of tension at 280 ft lbs. The tension value is lower likely because the 304 stainless will begin to yield before the grade 5 bolt. I probably need to calculate the thread pull out strength of a theoretical 304 stainless nut to see if that tension number makes sense. I know that 304 stainless nut has a yield strength of about 30,000 psi, which is only 30% of grade 5, but you have to factor in all of the engaged threads.

The torque value still puzzles me a little bit since the torque value to acheive the 16,000 lbs of tension on the connection was 280 ft lbs. That number is what is way over the calculated value (Torque = Friction Factor x Tension Force x bolt major diameter) for a simple grade 5 unlubricated zinc plated connection. However, I did find in another engineering forum that the friction of a stainless steel fastener may be 2 to 2.5 times that of a plain steel fastener. If you solve for a friction factor with the numbers from the lab test, Friction Factor = (280 Ft. lbs x 12 in/ft)/(16,000 lbs * .625) = .336. That may explain why it took so much torque to reach yield. Please let me know if you think that might be the explanation.

 

[unclesyd]

Book values are just a set of calculated torque values calculated from the physical properties of a fastener taking nothing else about the bolted joint into account except lubrication. There are a myriad of other considerations to take into account in the real world when specifying a fastener and it's torque value .


Portland Bolt and FanDisc have some torque tables. Portland bolt has some very good general information on Torque and tension.
I think Portland Bolt puts it best as they label their book values as starting values.

http://www.fandisc.com/tti.htm

http://www.portlandbolt.com/Documents/Portland-Bolt-Bolt-Torque-Chart.pdf

If you want to go a little deeper try this one.

http://assist.daps.dla.mil/quicksearch/basic_profile.cfm?ident_number=70455

Just about everything you wanted to know about bolting from NASA.

http://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19900009424_1990009424.pdf

 

[unclesyd]

Here is paper that discusses thread stripping by Dose

http://www.hexagon.de/dose/dose-1e.pdf

If the above link doesn't work go here and then to the paper.

http://www.hexagon.de/sr1_e.htm

As you state there will be a lot of make and break I would definitely use Helicoils.

Paper on using Helicoils look especially at Nitronic 60.

http://www.emhart.com/products/helicoil/techtop/tt10.asp

http://www.emhart.com/pdf/Heli Coil/HC2000 Rev 4.pdf

 

[byrdj]

when it comes to rigging, I get some one else to provide the requirements, but some of the things I've observed with lifting swivels
a 5/8" only gets torqued to 20 ft-lbs (just a consitant snug) and good for 3000lbs (only 1500 if used in SS). where a 5/8 shoulder eye bolt gets hand snuged and good for 5000

it these rigging uses, the clamping torqe is not a significant factor

 

[teejer]

Thanks everyone. Btrueblood, that was what my gut was telling me about the 0.2% point on the yield curve. I explained to the lab before the test that I wanted the torque, tension and strain measurement on the connection. They have not provided me a stress strain curve, which I think is in order here. To byrdj, thanks for your feedback on the rigging hardware tightness. The rigging hardware is installed on the side of a piece of equipment and then loaded on an angle, so getting as tight as possible will help the situation. I know that the stainless block friction will result in higher torque values, but I think the values that came back seem more like the ultimate strength and not the yield strength. I think some more testing is in order, especially with determining the effects of re-use on the bolts. They will be attached and removed possibly 100's of times in their life.

 

[byrdj]

I once made a bracket for flipping an awkard casing, my design was a bolt on trunion, while calcs showed 30K psi bolting loading would provide sufficeint clamping force to keep it from slipping, I added a solid pin into a fit to also carry the shear so the clamping force could have been negelable

 

[IceStationZebra]

If you have a choice I would design the fasteners to be in shear so the assembly torque will not be so critical. This would also allow you to specify solid pins to carry the shear load. To Byrdj's point somebody someday is not going to assemble the joint correctly.

If you do stick with this design I would strongly recommend a plaque stating the fastener specifications and torque value be conspicuously placed.

ISZ

 

[Tmoose]

"Rigger's Handbook" says a 5/8 inch forged steel eyebolt is good for 3500 lb working load at vertical pull. At 75 degree angle the working load should be reduced by 45%. Non shouldered eyebolts are only suitable for vertical lifts because. Do not use Shouldered eyebolts for lifts at angle less than 45 Degrees.


You probably need to use "hoist rings."
Or is that the "lifting device" you mentioned?

http://www.carrlane.com/Catalog/ind...C3B285356464A2D020609090C0015312A36515F574259

 

[teejer]

This lifting device is neither type, but is a unique lifting device that gets attached to the side of a piece of equipment. Even though it is a unique application device, I was also thinking that a swivel hoist ring would be a better idea for this situation. Thanks for the confirmation.

 

[rmw]

Teejer,

I am with you on your Oct 8 18;31 post. Something is real wrong with the 280 ft-lb/16,000 lb ratio, lubricated or not. I have a bunch of torque calculators bookmarked on my 'day job' computer and I don't want to chase them up now (at home, watching the ball game) so I'll run them on Monday.

I just finished a round of this kind of fun for 5/8" fasteners and one I had to evaluate was A193 B7 and I don't remember it getting anywhere near 280 ft-lb of torque.

I can remember that being higher than the torque limit for the highest pedigree fastener I evaluated (A540 B2X CLX - sorry I have to be vague for propriatary reasons). I can remember that at 240 ft lb I was getting right up under 90% of yield for that fastener. The tension at that point was right up under 22K lb and the bolt stress was approaching 100K psi real fast.

Something seems out of whack with your lab's numbers, notwithstandng SS's tendency to gall.

rmw

 

[teejer]

rmw,

Thanks for the reply and for the information on galling. I read up on galling and it looks like the rpm of the installation tool may actually increase the likelihood of galling. I read that if you back down on the installation rpm, it can reduce the galling problems. They also recommended some lubricants, which is out of the question here. Galling could be why the friction value was so high, which throws off the torque to tension relationship. I will inquire as to the rpm used in the lab.

On another note, I did some calcs using the yield strength of 304 stainless (31,000 psi)x the internal thread stress area for a 5/8" bolt x 0.577 (shear to yield relationship from Von Mises maximum distortion energy theory)and the result comes out very close to 16,000 lbs, so it seems that the yield point in the test was measured properly (10 samples tested). The question still is the torque value.

 

[sdra2]

Torque is not a good indicator of applied load for a bolt. With all other things being equal, you can expect a variation of +/- 25% for actual applied load using a wrench. Is this how you are applying your load?

Turn-of-nut is better, where you measure the angle turned by the nut. Measuring bolt extension is more accurate still but hardly practical for most field applications.

If you are going to use a torque setting for your field guys or customer, perhaps specify a nominal torque equal to 60% of yield. Plus 25% gives you 75% of yield - pretty much the most anyone would use for re-usable bolts. Minus 25% would give you 45% of yield. This is a big span, but that's the nature of using a torque wrench!

As for your test results vs book problem, as well as the torque variation, the thing to look out for is book vs real values for friction I'd guess.

 

[teejer]

Thanks for the reply and for the information on reusable bolts and torque. The ASME standard which does not technically cover this part, but covers similar parts, calls out for a torque rating. I want to start with a recommended torque and then ultimately look at a DTI (Direct Tension Indicating Washers)down the road as a possible solution. Everything I read about the turn of the nut method says that it is not accurate and is pretty subjective. Let me know if you have any thoughts on the critique of the turn of the nut method.

 

[unclesyd]

Would it be possible to use double-ended studs in the component to be lifted to where you just have to worry about tightening a few nuts on the lifting point device. Once the component is at it's final resting place remove the studs.

 

[Tmoose]

I believe fastener tightening reality varies considerably Depending on the industry. I have been, am, and in the future would be grateful if I can count on a lubricated (lube specified by me) fastener being tightened to a specified value with a torque wrench.

 

[Griz42]

Have you considered using a bolt that indicates the the clamping force? There are several different versions, but here is one:

http://www.smartbolts.com/index.htm