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

The definitve steps for trying to find the cause of a failure I would take:-

1/ check design

2/ inspect components to drawing

3/ check material conformity.

What is aconcern is what control on quality and material conformity you have if you can't obtain material data.

regards

desertfox

 

[preload]

Mike, I want to see if the inconsistency of applying the lube oil is the main reason? So planning to do some non lube oil tests and see the variation. If that improves then we gonna look for applying the oil consistently.

If this dosent work, then I am gonna do torqe-angle control strategy. As I have no chance to get the material properties and dimensions of the joint the best bet would be doing a torque-angle signature tests on the actual joint and deduce the threshold and angle from several rundowns.

What do u guys think?

Cory,

I asked my boss again, and he said there is no way engg dept can help me because they have to search for 20 yrs back records for the material properties. What all I can get is the material properties and dimensions of the bolt. If I want to do any improvement to the joint then I have to do using the existing numbers.
When u say dimensions of the joint, u mean the grip length? Or u need the area of the joint?

 

[preload]

MintJulep,

I will upload the yielded fastener pic today

Desertfox,

I AGREE with you . I am new to the job/company. My boss said this is not a problem joint in the recent past.This is not a problem joint in the recent past. We have no problems from customers or auditors on tha joint. But recently in the last 3-4 weeks we saw some problems, so he says that the joint strength and material strength is good. There is some other variable causing the problem.

Our main aim is to reduce the scatter. They beilieve the joint is good and torque is good(as we never had problems with the joint before).So they asked me to do some studies on the line and try couple of other strategies and see if we can reduce the scatter.
Our req preload for this joint is 9300 lbs max. we are getting to 12000 on some bolts (we believe those bolts yielding). So goal is to keep the clampload below 9300 lbs for every fastener and make the joint process capable.

 

[preload]

Observed the yielded bolt and as Mike said, I found 3-4 striations across the bolt (along the whole minor diameter) in the yielded area. I will try to take a picture of the bolt with full magnification.

 

[desertfox]

Hi Preload

Well maybe as others have said its a material quality
problem or a thread geometry problem just keep us posted.

regards

desertfox

 

[jetmaker]

preload,

Sounds like the torque applied is inconsistent. Have you tried using a torque wrench to verify the applied torque? Nothing like using old-fashioned technology.

Next, are all bolts hand tightened before they are torqued? Aluminum is relatively soft, and what could be happening is that the high turn angle results because of the fastener not being initially tightened. My thoughts is that this nee be done with each fastener prior to torquing it. When I installed the skins on my son's drums set, all bolts were initially finger tightened, and as I when in star pattern tightening sequence, some were found to have a large initial gap again, requiring finger tightening.

If all that is not causing the problem, I would purchase a fastener load cell, or load indicating washers, and setup a test piece and install several fasteners using the turn of angle and torque wrench methods to verify the bolts are good.

hope this gives you some ideas.

jetmaker

 

[preload]

here are the pics of yielded bolts

http://img337.imageshack.us/img337/5097/dscf1675vw7.jpg

http://img390.imageshack.us/img390/5242/dscf1679ge3.jpg

you can clearly see the crack/striations formed around the minor dia of the bolt. Between 5th to 9th thread from the top.

And from 10th to the end of the bolt is the engagement length of the bolt thread and the tapped hole threads.

Does this say anything?

 

[desertfox]

Hi preload

I can't see the images on this comp have to wait till friday.

regards

desertfox

 

[MintJulep]

Your plating seems to be junk and the thread form doesn't look very good to me.

I suspect you have garbage fasteners.

It is also possbile that whatever the plating is, it is reducing the coefficient of friction below what you are assuming.

The one picture looks like a crack, although I can't tell if it only the plating that is cracked, or the base material.

 

[preload]

MintJulep

If the plating is junk and coeff of friction is less then I should get high preload in every bolt right? But its not the case here, I am getting preloads anywehere between 6000lbs to 12000 lbs , and 9300lbs being the proof load. Do u think this big variation is due to "inconsistency" in the plating and thread form?

Jetmaker,

We don’t do hand start of the bolts. May be I can ask them to include this in the process. Thnaks for the suggestion. I wil definetly look into your ideas too

 

[TVP]

I agree with MintJulep, this does not appear to be a high quality bolt suitable for engine applications. It looks like it was picked off the shelf from Home Depot. A high quality bolt for main bearing or cylinder head applications requires excellent dimensional control (not just the thread shape, but also cylindricity/runout), careful control of the coating/lubricant (usually phosphate + lubricant not zinc plating + yellow chromate), and proper microstructure (quenched and tempered martensite, no decarburization, etc.). You need to at least verify that the fasteners meet the design requirements as stated on the appropriate engineering drawing. I would also recommend sending these images, data, etc. to the design group responsible for this joint. I know if this were my responsibility I would want to see the fasteners, both unused and cracked.

 

[strokersix]

This smells like it might be the unintended result of a Supply Management Department cost reduction effort.

Have you checked the hardness of the bolts yet?

 

[dimjim]

I wonder how much control you have on the
tapped drilled hole before tapping?
Looks like your length of engagement is at
least 2 times the diameter. What is the
strength of the castings? I know that was
asked before, but I did not find it as I
went down the replies. Has the casting
material been changed?

 

[Screwman]

Something is not adding up here. I think your torque calibration is way off on your guns.
I went back to your torque tension results and when you are getting 12,000lbs. of clamp at 35ft.lbs. it would give you a nut factor (K factor) of .093. That is lower than you are going to see with any type of commercial lube, fastener and nut material combination. Since the two variables are torque and clamp in that equation and you appear to have visual yielding, my guess is that your torque sensors in your drive gun are whacked out. my guess would be the one of the strain gages is becoming unglued and giving you crazy readings from hole to hole.
Do a torque study using a calibration standard and I think you will find your problem.

 

[preload]

Strategy : Torque control – Angle "monitor" from threshold

Torque installed : 31-35 ft-lbs and 32 ft-lbs is the target

Angle filter: 260 deg (device is set up for any bolt that sees more than 260 deg after the threshold will show a red light and operator will back off the bolt and scrap it and use new one)

Problem: when a operator sees 260 angle and red light,we examine the bolt , some of them shows signs of yield and some of them dosent (may be jusssst started to yield?)
I am uploading the data of the bolts which saw more than 260 angle and corresponding torque we saw in the controller. The bolts which saw less than 260 deg all attained the target torque of 31/32 ft-lbs. If u example the data at 261 deg of turn u get a torque of around 31 ft-lb but when u see the data for 262 deg the torque significantly drops (whats happening there?)

http://img244.imageshack.us/img244/8646/highangle1yp3.png

http://img467.imageshack.us/img467/1277/highangle2do1.png

Proposal from management: they asked me to develop a angle strategy instead of angle monitor to reduce the scatter.

I asked them about the calibration and bolt strength concern (as many here suggested) and they replied the guns are very well calibrated (checked last week) and all the bolts are to the print and design. Material properties are upto the mark.

Machine is a 2 spindle machine and I have the data of 5000 rundowns per spindle
Spindle 1 – 144 rundows showed angle more than 260 deg out of 5000 rundowns
Spindle 2 - 89 rundows showed angle more than 260 deg out of 5000 rundowns

I just attached the data for 89 rundowns of spindle 2 which are red lighted due to high angle.

Please look at the data and give some suggestions.

 

[preload]

Please magnify the image if u guys dont see then numbers clearly. Thanks

 

[MintJulep]

I'm still not clear on how you come up with 12,000 pounds load. Early on you wrote:

After the installation we are measuring the length of the bolt using ultrasonics. Then calculating the clampload using T=KDF

which doesn't make much sense. More sensible would be to calculate load based on extension using DL = F((Lb/EAb) + (Ls/EAs)).

However, this only works in the ELASTIC region. Once you get into plastic deformation calculating load based on change in length gets way more complicated.

 

[desertfox]

Hi preload

Just a quick thought your fasteners are plated could they be suffering Hydrogen embrittlement during plating.
This would cause the random failing of some bolts and not of others.
Have a look at this site:-

http://www.fera.org.uk/pdf/Fera activity listing - hydrogen - march 06.pdf

regards

desertfox

 

[MikeHalloran]

I'm not sure I'd want to 'reduce the scatter'.

The bolts in the photo appear to have cracks at the thread roots. The presence of cracks suggests that it would be possible to fracture the bolts by continued twisting, at a torque level below the desired range, so when the angle monitor stops the line, it's preventing continued use of a bolt that's basically already failed, and stopping the operation soon enough that you don't have to split the cases to exract the pieces of the bolt.

So the wrenching system appears to be working in a desirable way, perhaps even as someone intended.

It goes without saying, so I'll say it, that bolts flagged by the wrenching system should never be re-used, i.e. they should be cut in half immediately, and discarded. You might want to hold onto a few until you find out why a substantial fraction are failing at assembly.

Something about the topology of the cracks reminds me of grade 5 bolts torqued to failure, not grade 8 bolts.



Mike Halloran
Pembroke Pines, FL, USA

 

[preload]

Mintjulep,

We are using the same formula. I meant to say T=KD(EA*deltaL/L)
DeltaL being the elongation. When I say 12000 I agree we are really not seeing that clampload but according to alongation we get that number. But those reached 12000lbs are yielded. Because 12000lbs is more than the yield strength of the bolt. We are trying to keep it simple.

Desertfox,

I will confirm with my supplier quality guy reg the hydrogen embrittlement

Mike,

Exactly. When we are seeing 260 deg they are visually necked. So our angle filter is doing good job but at the same time we need to lower the angle filter to 200 deg and see if any fastener hit that mark are yielding??????? By doing this we are scrapping more fasteners and my management don’t like it. So they want to reduce the scatter by using angle monitor and revising torque specs (reduce the target torque). What do u think Mike?

But I dint understand one thing in my data. 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?????????