Engine Main bearing bolt yielding help 9

[preload]

Hi

I am new to this forum, but used to be in ROM (Read only mode). I am new to the fastening field but have quite a bit knowledge. My problem is we are yielding some bearing bolts right after installation on the cylinder block and crank case joint. I will provide you guys the info on bolt and joint parts and environment.

Bolt – carbon and low alloy steel flanged head 3/8-16 grade 8 bolt.
Joint – both cyclinder head and crankcase are made of Aluminum – copper alloy casting (soft joint compare to the bolt)

We also use a bead of gel seal in between the mating parts (cylinder head and crank case), when in contact, the gel seal uniformaly gets distributed on the mating surfaces. (making the joint even more soft?)

We make two engines V4 and V6. for v4 we have 6 bearing bolts and for v6 we have 8 bearing bolts. The tool used is two spindle dc electric torque angle monitor. For v6 We do bolts 1-2 then 3-4 then 5-6 then 7-8 and again do 1-2 (re torque due to elastic interaction) (I will try to upload a picture for better understanding)

Torque used is 31-15 lbs-ft on each bolt.Min proof load of the bolt is 9300 lbs.

Do you guys know why we are yielding bolts on a soft joint (if it is?)

 

[desertfox]

Hi preload

Its not a good idea to just reduce torque specification!
thats just a knee jerk reaction without proper technical thinking, all your going on so far is people saying that the materials okay, the joint design is ok, just look at the fasteners, you need to convince yourself these things are okay.
So having said that, if you have got to look at the bolt start by checking the material properties of the bolt have them confirmed by test, in addition you could torque some bolts to failure and compare them with that picture of a bolt failure due to hydrogen embrittlement on that site I posted previously.
Hopefully I will get to see the pics you have posted on friday.

regards

desertfox

 

[preload]

Desertfox,

I believe in you guys but my management has limited resources. I was arguing with them yesterday regarding my concerns on bolt and joint (material,geometry,strength) rather than angle control or torque change. I don’t know, for some reason they feel the preload scatter is due to inconsistent application of lube oil

What is the simple method to prove them wrong? My plan is to do some sample studies with no lube or consistent lube and see if we still have scatter. If we still have scatter then that means the preload scatter is not really due to lube oil. What do u guys think? Or gimme some ideas to prove them wrong.

 

[MintJulep]

Ok, once again. Once you yield, there is no longer a linear relation between bolt length and preload. If your 12,000 lb figure is based on ultrasonic length measurement, and calculation of load based on elastic deformation then it is WRONG. No simple. WRONG

At 261 deg of turn we almost see 31.xxx of torque but at 262 deg we always see a big drop in torque. Whats goin on in that 1 deg of difference?????????

Answer: The bolt yeilding.

 

[sreid]

As many have said, the most likely problem is bad bolts. The simplist way (and probably the least costly) is to buy new bolts from a known good source. If, say, 100 of these work OK, you've got your answer.

 

[unclesyd]

Looking at the pictures of the failures I don't like the cracking that you are having. I concur that something is going on with the fastener itself, as mention H2 embrittlement, heat treating cracks, etc.
If dimensions, clearances, or friction are involved you will normally get a shear or torsion failure. This type failure is normally across on thread and break is characteristically flat/round.
Failure from tension is normally a jagged break across sevral threads, as in you case.

My first inclination would be to toss this lot of fasteners.

If possible I would get a failure analysis of one of the failed fasteners.

The grade 5's that I've worked with all have had considerably more necking prior to failure with a clean break.

Grade 5's will crack from H2 embrittlement contrary to what some vendors will tell you.

 

[MikeHalloran]

For fifty bucks or less, you can buy a Hamilton gas-tight syringe that will allow you to apply very small controlled volumes of oil to the bolts. I would suggest buying the 100ul size and dispensing about 10ul for a start. That requires about 6mm of stroke on the plunger. Get the luer or luer-lok tip. You don't need a needle.

Here's a 250ul that will work okay for $17:

http://cgi.ebay.com/Hamilton-Gastig...4675100QQihZ009QQcategoryZ26414QQcmdZViewItem

That will require ~2.4mm of stroke to dispense 10ul, but you don't need extreme precision for screwing around with the idea.

Buy two; the barrels are made of glass. They're not delicate, but they can be broken.



Mike Halloran
Pembroke Pines, FL, USA

 

[preload]

Thanks Unclesyd and Mike.

What is the procedure of doing a failure analysis of one of the failed fasteners.

 

[unclesyd]

Here are some things you can accomplish with little or no equipment.
If you can get your management to agree I would send one or two yielded fasteners to a Metallurgical Laboratory for failure analysis.

I would by getting a hardness reading on the failed studs. The sample for the hardness test should be cut from one end of the stud. The faces of cut stud should be carefully ground parallel with no heat input from the grinding. Take hardness readings, discard the the first reading.

I would also take a small longitudinal sample, just split a short section of the bolt. Grind and polish the flat surface to get as smooth as possible. I would lightly swab this surface with a 2% HNO3 in Alcohol, laboratory ethanol, after say 15-20 seconds rinse in cold water dry, dipping in some pure ethanol, and look at the edges of the threads to see if they are lighter or darker than the body of the stud.

To take a look at the fracture surface I would take one of the elongated samples with cracks and tighten it until it breaks. Maintain the integrity of the fracture surface and if possible take a picture of the entire surface and post it on Image Shack for all to see.

 

[preload]

Unclesyd

Hey, we have a lab and I can get this metallurgical testing done.Do I have to do all the three testings you specified?

Hardeness
Longitudinal sample test
Fracture surface

Or they are all one test?

 

[Screwman]

Do a tensile test on the bolts using a lairly large sample size and I bet you will find that you have parts that are under strength. In my experience you are much more likely to have a component quality problem or a drive gun issue than a need to change the torque setting.
One quick thing that you could do would be to take some EP grease and coat the threads on a few bolts and see what results you get. You should yeild 100% of the bolts if this is a lube related issue and break some if it is a QC issue.
For a high load critical joint you should switch to angle control and torque monitoring, which will help to reduce your clamp load scatter (once you get the rest of your parts and process under control). Using a snug fit torque threshold and then using angle control will take much of the clamp variation caused by frictional changes out of the picture and you can still monitor torque at the angle shut-off to see if it is within limits. (low torqeu at angle means something is stretching or compressing unusally and high torque at angle means excessive friction (thread nicks or cross threading ususally)). Angle control is a very useful tool and I highly reccoment it for critical, rigid joints.

 

[preload]

Screwman,

The lube oil we are using is applied inconsistently. Some times we get some lube oil to the threads and also under the head and some time only to the threads. My management thinks this is the reason for some bolts having high clamploads up to yield point and some normal clamploads or low clamp loads.

The tensile load test studies done before on the same bolt has an very good avg of proof load. they did a 24 sample tensile test and found out that the proof load for this bolt is 9300 lbs and which is exactly right to the print. This study got confidence in my management to believe that bolts are doing good.

But my plan is instead of using EP grease, can I use the same lube oil we are using now but will make sure to apply consistently only to the thread area of the bolt and see the results? And if I still see this yielding problem and large scatter then I can argue with my management that the bolt or drive gun is the issue. And if I get comfortable results then I will agree with my management and develop an angle control stategy.

What do u think?

 

[Screwman]

That sounds like resonable approach. You may also want to apply oil under the head since almost half of the friction is under the head and if you lube both it would truly give you the worst case situation.
Were the bolts that the tensile pulls done on from the same lot as the suspect parts? I just have a feeling that the bolts are playing a role in this problem.....

 

[preload]

Screwman,

I can get the tensile load test studies on the recent (the one we are using on the line now) on Monday. It seems our supplier has done the tests and the bolts are pretty good regarding the tensile strength and yield strength. But I will get the data from them to make sure again.

 

[desertfox]

Hi preload

I looked a the fastener pics you posted showing the cracks and came to the conclusion I would have liked to have seen
a pic of a bolt that had been completely destroyed ie:-
torqued till it broke in two.
The reason being if you look at the two fracture surfaces it can tell you where the crack started and also whether it was a complete tensile failure or tensile/shear stress combined.
This brings me to my next point the bolts are failing during tightening and therefore the yield stress of the material is being reached by a combination of tensile and shear stresses and not by tensile stress induced alone.
This tensile stress is known as a principle stress and would occur at some angle relative to the bolt central axis.
I surfed the net for some information and found that for a grade 8 material for your thread size the proof stress was
equal to:-

9300lb/0.0775"^2 = 120000 psi

the 0.0775" is the stress area for your bolt.
see this site:-

http://www.almabolt.com/pages/catalog/bolts/proofloadtensile.htm

Now with that info and a figure of 33lb-ft as the torque on
a plated bolt of your size from this link:-

http://www.almabolt.com/pages/catalog/bolts/tighteningtorque.htm

I used a Mohr stress circle and found that the principle stress acting at an angle of 14.24 degrees below the horizontal was about 136500 psi which is in excess of the proof stress of 120000.

Now it is interesting to note that the torque figure I got from the above site is the recommended one for joints that are not critical and (although it doesn't say so) the torque figure I believe is for an unlubricated joint.
It also shows an expected clamping force or pre-load for that joint to be 6975lbf
In your application you are using that torque figure with lubricant which is obviously reducing the friction greatly
and generating a high a pre-load.
Whilst I accept that the torque figure stated on the above site is for cap head screws and not flanged headed bolts which may well have some effect I now believe your main problem is one of friction coefficient.
If you really need a pre-load of 9300lb as you stated in an earlier post you need to get a better grade of fastener.
Finally it is still possible that the other points raised about material propeties and control, thread geometry etc
are still playing a part in your situation.
I will look forward to seeing the tensile results of the current fasteners next week and if you could torque a bolt in two and post pics of the fracture faces also that would be helpful.

regards

desertfox

 

[pennswoodsed]

1 independent(truly,anonymous)testing of bolts
2 source aftermarket(race)engine bolts and test
3 wait for screaming(pissing contests) to die down and
assemble realible engines
4 the whole thing seems like a dollar waiting on a dime
Regards,Ed

 

[desertfox]

Hi preload

I found a site that confirms that the 33lb-ft figure is without lubricant look here:-

http://www.morbark.com/Service/belttorque.pdf

Also look at the notes at the bottom of the page it talks about reducing the torque setting if oil lubricant is used
and a further reduction for new cap screws which are plated.
I haven'nt full access to the document they quote:-

IFI 5th Edition Technical Data N-12/N-16, using Equation (1)

as I am based in the uk but if you can get hold of it,it might help.

Now if I calculate the principle stress based on there torque and a 6975lbf clamp force then the principle stress
is:-
111102.56 lbf/in^2

which is below the proof stress of 120000 by about 7.5%

Reading back through earlier posts it is clear that this joint was not a problem untill recently so I wish to ask the following:-

1/ Was lubricant always used?

2/ Has the lubricant changed recently?

3/ Were the bolts always plated with the same coating?

4/ Were there any changes to pre-load or torque setting
methods?

The reason for my questions is that assuming everything else
is okay with thread geometry, material properties etc I cannot see how this joint would not give the problems your seeing now with the torque setting and lubricant unless something changed significantly.

regards

desertfox

 

[unclesyd]

The test I recommended are primarily use to quickly eliminate some factors that may be contributing to your problem.

The hardness test will indicate if the heat treatment of the studs has went awry. It also gives a tidbit of information for future reference.

The longitudinal section should show whether you have either carburization or decarburization which can throw you tightening data into a mess.

It is always nice to look at a fracture surface. Somethings it will point to the smoking gun while at other times it becomes a chore to pin down the mode of failure.

 

[preload]

Desertfox,

I greatly appreciate your help.

I will get the tensile test data soon. I will also get the picture of a broken bolt.Regarding your 4 questions in the recent post , I wil surely answer them on Monday.

regarding the math,

33 ft-lb torque - 6975 bf according to the document and he said take 10% less if the engine oil is used. so in our application :

29.7 ft-bs generates 6975 lbf right? and I am using 32 ft-b as the target torque which shoyld generate 7500lbf, which is way below the proff load (9300lb).Thne how am I yielding?Am I using way too much of lubricantion? thats generating more than 9300 lbf for the same 32 ft-lb torque than what the document says?

Could you please explain the following statement more
"I used a Mohr stress circle and found that the principle stress acting at an angle of 14.24 degrees below the horizontal was about 136500 psi which is in excess of the proof stress of 120000."

I am not that good at mohr circe.

 

[desertfox]

Hi preload

The Mohr stress circle can be used to find the resultant stresses acting on a plane, in your case the bolts are failing during tightening which means they are failing when they have both shear and tensile stresses present.
Only when you stop tightening the bolt are the shear stresses removed.

Have a look at these sites for Mohr's

http://www.sunybroome.edu/~biegen_j/strength/examples/chap7/ch7_cl1_mohr2.htm

http://physics.uwstout.edu/StatStr/statics/spec1/spec85.htm

http://en.wikipedia.org/wiki/Stress_(physics)#Principal_stresses_in_2-D

Now regarding your other points the 6975lbf at 33lb-ft torque is quoted using a friction factor of 0.15 without any lubricant,for a joint that is not critical and is subject to +/- 25% on pre-load.
Bear in mind that torque figures are only a rough estimate
and if your joint is critical then you need to monitor your pre-load by a more accurate method like measuring elastic stretch of bolt during assembly. I cannot put an accurate figure of pre-load against a torque because the friction factor which can vary considerably is dominating the equation.Having said that the point which I am trying to make is that you're using a torque figure which is recommended for a joint without lubrication,therefore in a joint without lubrication I would expect the friction factor to be higher and for a given torque you would achieve a smaller pre-load resulting in lower tensile stress in the bolt during and after tightening and less likely to fail during assembly.Because you are lubricating the bolt your friction factor is much lower causing the bolts to fail during tightening.
Now the only way to solve this problem is to reduce your torque figure for the lubricated joint or more importantly your pre-load, however this puts us back to square one because we need to know what the pre-load needs to be in order to comply with the original design which as yet has not been confirmed.
The other point I am making is that if in the past the assembly was being made then as it is now I beleive that the same problems would have occurred unless there has been a major change in material supply or lubrication, pre-load procedure of recent.
Not sure how you got from 32lb-ft to 7500lbf depends on your friction factor.
First thing to do is define your joint pre-load with a tolerance, then we can estimate a torque to achieve the pre-load, from there you need to make practical measurements with a consistant pre-load procedure to ensure that goal.

regards

desertfox

 

[preload]

Desertfox,

I am gonna call a meeting with the design engineer of the joint soon. He should give me a clamp load spec.

And regarding the 32ft-lb - 7500lbf figure, the math goes

from document,

without lubrication 33 ft-lb gives 6975 lbf for my bolt.

with lube oil (which is my situation) the document says to decrease 10% of the torque used to get the same 6975 lbs.So 10% of 33 is 29.7 lb-ft. so 6975 lbs clamp will be generated from a 29.7 lb-ft in a lubricated state. right?

so in lube state,
29.7 lb-ft. - 6975 lbs
then 32 lb-ft (my target torque) should produce 7500 lbf right?

but, my bolt has a capacity of 9300 proof load,then how am I exceeding the proof and yield strength?

May be I am over lubricating the bolt...