What's new
What's new

Best compromise for less than ideal main bearing bolt design?

We discussed the matter with an engineer at ARP who said the best way to deal with this is to turn down the shank and the unwanted threads to the minor diameter of the coarse threads on a lathe, use a .4" fillet radius where the top of the stud transitions to this diameter (after the fine threads). We need to make sure the bolt doesn't get hot and that it ends up with a good surface finish after (probably chuck it up in a drill use sand paper). This way we get an even tension through the bolt with no notch effects (vs grinding unwanted threads off) and we tension the block in a place where it can better handle it without removing more material from the hole.

Going to take this to a few machinists and see if it can be done in one or two billable hours (the most we could afford).

With regards to some of the questions: the engine is being built, as best it can, to reliably tolerate higher than factory cylinder pressures and have a wide powerband. Putting the power higher in the RPM isn't not optimal given the engine utilizes a HEUI injection system which is relatively slow such that 3k is the limit; the powerband will narrow quite a bit if this is done since there's no headroom. The bedplate has the effect of distributing the load better across all mains fasteners and also ties in to the rails to increase rigidity of the crankcase as a whole and further distribute load.
I'm curious. Did the ARP engineer have an explanation of why his product wasn't correct to begin with? ARP seems to be a religious icon among the hot rod crowd. Everything can be solved with ARP hardware. The minor diameter is a standard thing for critical applications and yet they missed it.
 
I'm curious. Did the ARP engineer have an explanation of why his product wasn't correct to begin with? ARP seems to be a religious icon among the hot rod crowd. Everything can be solved with ARP hardware. The minor diameter is a standard thing for critical applications and yet they missed it.
ARP has their place but proper application of them is the responsibility of company or person buying them. I think in this case the studs are properly designed, rather the company that created this kit didn't put too much thought in to the particulars of the bolted joint when they ordered the studs from ARP, hence us having to come up with a way to correct for the deficiencies it created.

screenshot_20240311_110717_2_86f84f76642c9784ac07d771055e64e0d95964f9.png

The studs have been machined for reduced shank to engage the block's web below the bearing cap register surface without requiring a counterbore. Per my math we should be able to get the same clamping pressure at around 80% preload. The ideal way would be to setup a DI on the end of the stud to measure stretch as it's tightened, issue is the stud will rotate some during tightening which could throw off the stretch measurement. Even flexing of the bedplate or the block where the DI holder is mounted could throw it off.

The standard refrain is a potential 25% variance on fastener preload using torque... not sure if this is the same with ARP fasteners since they make the hardened & ground washers, nuts, lube, and generally can make things more consistent. Might have to bug the engineer more concerning this.
 
Last edited:
ARP has their place but proper application of them is the responsibility of company or person buying them. I think in this case the studs are properly designed, rather the company that created this kit didn't put too much thought in to the particulars of the bolted joint when they ordered the studs from ARP, hence us having to come up with a way to correct for the deficiencies it created.

View attachment 433171

The studs have been machined for reduced shank to engage the block's web below the bearing cap register surface without requiring a counterbore. Per my math we should be able to get the same clamping pressure at around 80% preload. The ideal way would be to setup a DI on the end of the stud to measure stretch as it's tightened, issue is the stud will rotate some during tightening which could throw off the stretch measurement. Even flexing of the bedplate or the block where the DI holder is mounted could throw it off.

The standard refrain is a potential 25% variance on fastener preload using torque... not sure if this is the same with ARP fasteners since they make the hardened & ground washers, nuts, lube, and generally can make things more consistent. Might have to bug the engineer more concerning this.
Sorry I haven't been back to you sooner but I took a time out for cataract surgery. I thought I remembered you giving a diameter for the shank but now I don't see it, only that the ARP engineer said to turn the shank to the minor ID. Your stud has 2 minor IDs, an unusual situation. My guess is they took a stock stud with the 9/16 thread on both ends and altered it with the 12mm thread on one end. this left it with the 9/16 minor diameter of .472". The 12mm thread has a minor ID of .388. These may not be exact as I used internet charts, too lazy to get my handbook. The good news is these will not likely break in competition. That is because at 200lb/ft you will probably break them with the wrench on them. This why the ARP engineer suggested turning down the shank to the minor diameter. Better to get new studs at 9/16 both ends, even if you have to open up the holes in the plate. This is a real bastard so good luck. Please let us know how this comes out.
 
Somewhere... some ????? years back ... torquing head bolts ... or studs ...

The stud stretches as you increase the tension on it, making the thread pitch longer ... while the tapped hole is in compression, making the thread finer.

Only the first few threads take the brunt of this ...

Some tweaking of the pitch looked to be beneficial... which of course required oddball custom studs or bolts to be made to work properly in tapped holes.

A total nightmare.... until CNC came along and you could program whatever pitch you wanted.

I think the 'ideal' required the stud, or hole, to be tapered .... or that was a 'solution' to altering the pitch ...

Seems even harder than programming a 16TPI @ .0613, or whatever ...
 
Somewhere... some ????? years back ... torquing head bolts ... or studs ...

The stud stretches as you increase the tension on it, making the thread pitch longer ... while the tapped hole is in compression, making the thread finer.

Only the first few threads take the brunt of this ...

Some tweaking of the pitch looked to be beneficial... which of course required oddball custom studs or bolts to be made to work properly in tapped holes.

A total nightmare.... until CNC came along and you could program whatever pitch you wanted.

I think the 'ideal' required the stud, or hole, to be tapered .... or that was a 'solution' to altering the pitch ...

Seems even harder than programming a 16TPI @ .0613, or whatever ...
And a lotta years back it was not uncommon to leave the shank at the major diameter. This puts ALL of the stretch in the threaded area and causes problems. Modern practice has the shank at the minor or sometimes less to put ,as near as possible, all stretch in the shank. This reduces the deformation of the threaded area. Also gives the maximum stretch to the fastener which helps compensate for gasket crush and temperature change.
 
I'm thinking ... this came from someone who'd been working on nuclear stuff. Or high pressure test chambers? Huge studs, stretched with hydraulic cylinders, and the nuts run down hand tight. So you really know what you've got.
On a related note, I have seen pin shaped fasteners with a groove in the end. You pound in a tapered wedge of an exact thickness, and then you know exactly how much torque you have. I want to say I saw this on old flat head engines, where there was a slot in the edge of the head and the block for an I shaped pin, and a lip on the head so that once the wedge was installed the pin couldn't slide out the side. Not really practical with traditional engine designs.
 
The old Leyland diesels had exact length big end and main bolts ,and you tightened them to a certain extra length ,then enough more to get the split pin in.............The big Hercules diesels had all the main cap bolts extended up through the heads ,no threads in the casting at all .............Gardners had the main bolts held down the cylinder blocks ...again no threads in the castings.
 
On a related note, I have seen pin shaped fasteners with a groove in the end. You pound in a tapered wedge of an exact thickness, and then you know exactly how much torque you have. I want to say I saw this on old flat head engines, where there was a slot in the edge of the head and the block for an I shaped pin, and a lip on the head so that once the wedge was installed the pin couldn't slide out the side. Not really practical with traditional engine designs.
How would "you know exactly how much torque you have"? You could drive the wedge in with it loose enough to rattle.
 
Somewhere... some ????? years back ... torquing head bolts ... or studs ...

The stud stretches as you increase the tension on it, making the thread pitch longer ... while the tapped hole is in compression, making the thread finer.

Only the first few threads take the brunt of this ...

Some tweaking of the pitch looked to be beneficial... which of course required oddball custom studs or bolts to be made to work properly in tapped holes.
I called ARP and someone explained to me there is nothing weird going on with their threads to solve this problem

He was able to tell me that they sell studs for certain aluminum bmw blocks and they have 30mm of thread engagement. the studs sold for the iron blocks have 25 or 20 i can't recall.. if you don't buy the right ones you'll pull the threads out of the aluminum block.

So anyhow what i was able to figure out that the reason why the thread force distribution is a whole lot better than steel on steel is because cast iron is much less stiff than steel is. aluminum even more so, and is much more ductile, so you can have 30mm of thread engagement on a 11 or 12mm bolt.. i literally cannot find a single published FEA simulation showing thread force engagement in dissimilar materials.

aluminum fatigues, certain toyota camry engines had a problem of the aluminum block threads failing, but only on the 3 head bolts that get extra warm from the exhaust manifold. the bolt can handle it but the aluminum threads couldn't handle the extra 20% (?) change in tension every heat cycle.
 
i found a comment here https://www.eng-tips.com/viewthread.cfm?qid=375921

inline6 (Mechanical)1 Dec 14 17:03
I did some testing of gr12.9 metric m12 screws into various aluminum alloys to look at thread stripping and some elasto plastic axisymmetric FEA as correlation some time ago.
What I found was that due to the lower elasticity of the aluminum the loading of threads was highly nonlinear. At small loads the first couple threads took the vast majority of load but once the bolt tension approached the strength of the bolt then the threads become very uniformly loaded before subsequent shearing of the threads. This effect is more pronounced than in steel. The thread stripping resistance was basically linear with engagement length in the typical range of engagement lengths used.
On a fine thread I could not get the bolt to snap in any of the alloys the threads always stripped no matter the length of engagement, I did get quite close though. However with the coarse thread with 6061-t6 and 2011-t3 extruded bars it was easily possible. My tests were done in uniaxial tension so no torsion in the bolts or possibility of galling was accounted for so one would need to investigate additional safety factor if the threaded joint need to be reassembled several times. The tolerance class of the threads is important but requires further study
For stripping resistance you either need to look at UTS or even better correlation comes from the true ultimate strength, yield strength is not relevant to stripping.
 
We were able to speak to the engineer at ARP again. From that conversation we determined that the best way to hit desired preload on these fasteners would be to measure stretch.

Being blind hole studs that's tricky since we can only measure one side and if the stud rotates at all during tightening it will throw off the measurements. The idea to deal with this is to apply loctite to the studs, torque to snug, let them cure. This should keep the studs from rotating while tightening the nuts, letting us measure stretch.

The only concerns about doing this would be other factors contributing to what the the DI is reading that isn't bolt stretch: the castings moving where the DI is mounted, the block pulling up slightly where the stud is threaded in... the geometry of the joint changing slightly. I don't have a good intuition for how much these factors would contribute; if it's small enough to be safely ignored.

I assume that, when doing the stretch calculation, I should use three different stress areas and respective lengths since there's shank diameter, 9/16 coarse stress area, and 9/16 fine stress area?
 
I pulled down a V6/71 that had low oil pressure ....the cause was a broken main cap bolt ......but worse ,the bolt had broken about 4 threads into the block,and pulled a piece out of the block behind the bearing shell..........fortunately for the boss,the dealer who sold the compressor came good with a another motor ,which I rebuilt and fitted to the compressor.
 
The only concerns about doing this would be other factors contributing to what the the DI is reading that isn't bolt stretch: the castings moving where the DI is mounted, the block pulling up slightly where the stud is threaded in... the geometry of the joint changing slightly.
I can't comment on whether or not the above will happen or how important it is, but you could quantify it on the first bolt you tighten by mounting a DI to read off where the other mag base is mounted and another mounted elsewhere on the casting to read off the joint itself. You might end up mentally tying yourself in knots but it would offer more information.
 
We were able to speak to the engineer at ARP again. From that conversation we determined that the best way to hit desired preload on these fasteners would be to measure stretch.

Being blind hole studs that's tricky since we can only measure one side and if the stud rotates at all during tightening it will throw off the measurements. The idea to deal with this is to apply loctite to the studs, torque to snug, let them cure. This should keep the studs from rotating while tightening the nuts, letting us measure stretch.

The only concerns about doing this would be other factors contributing to what the the DI is reading that isn't bolt stretch: the castings moving where the DI is mounted, the block pulling up slightly where the stud is threaded in... the geometry of the joint changing slightly. I don't have a good intuition for how much these factors would contribute; if it's small enough to be safely ignored.

I assume that, when doing the stretch calculation, I should use three different stress areas and respective lengths since there's shank diameter, 9/16 coarse stress area, and 9/16 fine stress area?
We were able to speak to the engineer at ARP again. From that conversation we determined that the best way to hit desired preload on these fasteners would be to measure stretch.

Being blind hole studs that's tricky since we can only measure one side and if the stud rotates at all during tightening it will throw off the measurements. The idea to deal with this is to apply loctite to the studs, torque to snug, let them cure. This should keep the studs from rotating while tightening the nuts, letting us measure stretch.

The only concerns about doing this would be other factors contributing to what the the DI is reading that isn't bolt stretch: the castings moving where the DI is mounted, the block pulling up slightly where the stud is threaded in... the geometry of the joint changing slightly. I don't have a good intuition for how much these factors would contribute; if it's small enough to be safely ignored.

I assume that, when doing the stretch calculation, I should use three different stress areas and respective lengths since there's shank diameter, 9/16 coarse stress area, and 9/16 fine stress area?
No. use the diameter and length of the shank. The stretch will occur here. The 9/16 will not stretch enough to be measured and the end with nut will not be in tension for enough off it's length to matter. To be picky use the distance from the block surface to the nut mating surface. What formula are you using to calculate stretch?
 
No. use the diameter and length of the shank. The stretch will occur here. The 9/16 will not stretch enough to be measured and the end with nut will not be in tension for enough off it's length to matter. To be picky use the distance from the block surface to the nut mating surface. What formula are you using to calculate stretch?
Was going to be using these formulas: https://www.engineeringtoolbox.com/bolt-stretching-d_1164.html but using the area of the shank for the tensile stress area variable.

Shank diameter is .455-.460"... will calculate based on the lower number. Shank lengths are ~4-9/16" and ~4-1/16" for the long and short ones respectively:PXL_20240330_035559138.jpg

On an aside, speaking with the ARP engineer last week there was frustration with how difficult it is to cut through certain myths and unfounded knowledge regarding the customers of their products. Basically all performance oriented fasteners in tension joints seeing cyclic loads should be reduced shank as they are much more fatigue resistant. The problem is they cost more to produce--ARP already has to compete with no-name brands offering "high performance" fasteners from China--and customers think the bolts are weaker because the shank is smaller in diameter vs a traditional style bolt. Hence these style bolts aren't seen much unless they are special ordered by customers who actually know their stuff.
 








 
Back
Top