I'm thinking about the VERY beginning appearance of a cavitation pinhole.

Hopefully, you have got the correct diagnosis in hand.

Edit:

On second thought, I think my logic is fuzzy. Even a VERY-small pinhole will develop cooling system pressure with the engine just running, not only at elevated boost.

In the words of Emily Latilla, "Never mind....."

Pop

 

Well the laws of "Physics" say the FORCE which is required to rip a stud from a threaded hole is proportional to the number of threads that are engaged between the threaded stud and the threaded hole and this is exactly analogous to the yield strength of a bolt being proportional to the diameter of the bolt!

Keep in mind that studs should only be threaded into the block finger tight until they just bottom out using the minimum torque possible! This allows the full strength of the cast iron threads to be used for stretching the stud when the stud nut is tightened to its maximum torque!

 

Gene, I didn't mean to "challenge your knowledge" or imply that you're wrong in any sense, so please don't take offense at my questioning. I'm just trying to understand what you said and why it seems different from what I was hearing from industrial gasket techs the other day.

I understand that part of the physics, but the rest of 'bolt science" - from what I've been instructed - has also shown that once the bolt (or stud) is placed under the clamping load, it is typically only the top three to four threads which take the majority of the load. What I am talking about is the whole issue of "engaged threads". Why is the stud different than the bolt in this case? Why the stud was "fully engaged" and the bolt was not? Is there a difference in the accuracy of the stud's threading (in comparison to that of a bolt) that makes it have a more uniform "contact presence" within the block? Or is there something else at play here?

 

Pop the only experience I have with that is baby-ing my Pete with a 435 Cat home from South Carolina I stopped every 200 miles and put 2 gals of water in. That was a long trip. It didn't pressurize the cooling system it sucked it in.

 

I'll use a fictional numerical example here to hopefully clarify our discussion. Say you've got two identical threaded holes in a cast iron block that have N=12 full threads meaning it takes 12 turns (360 degree rotations) of a stud or of a bolt before hitting the bottom of a hole.

Now you install a stud finger tight for 12 turns until it just bottoms out in one hole and you install a bolt finger tight for 8 turns in the other hole. Now you measure the vertical force Fs that's required to pull the stud from the block and you measure the vertical force Fb that's required to pull the bolt from the block. You seem to be claiming that since more than 4 threads are engaged for both the stud and the bolt that Fs=Fb and I'm claiming that Fs>Fb.

The forces Fs and Fb are countered by an equal and opposite force that's provided by the total shearing force on the threads. If the threads are perfectly cut the total shearing force on the threads is uniformly distributed so that each thread provides 1/N of the total. This means that Fs=(12/8)Fb=(1.5)Fb and it takes a 50% larger force to pull the stud from the block compared to the bolt.

Here's an analogy. Assume you've got some rope that can just hold a 10 lb force before breaking. I claim you can use 12 parallel strands of rope to support a 120 lb weight which is 50% more than the 80 lb weight that can be supported by only 8 parallel strands of rope. You seem to be claiming that in both cases all the load will be carried by only 4 parallel strands of rope and that no matter how many parallel strands of rope you use you can only support a 40 lb weight?

Note that in the above example torque never enters into the picture because both the stud and bolt are threaded into the block finger tight. When you clamp a head to a block using a bolt you need to leave a safety margin of say 4 turns to make sure the bolt doesn't bottom out in the hole before reaching its full torque value and that's why in my example I used only 8 turns for the bolt versus 12 turns for the stud.

When installing a head, after the head bolt has engaged the first thread in the block and has been turned several additional turns using the fingers all of the slack is removed and now torque needs to be applied to the bolt to provide a clamping force. As torque is applied only the few threads that are engaged can take up the load and as they do these threads stretch a little and then more torque is applied and as additional threads are engaged they stretch by a different amount, etc... until the final torque value is reached at which point the load on the say 8 threads that are engaged is definitely not uniformly distributed as it was in the finger tight example.

So to recap when you install a stud finger tight into a threaded hole in a block and then apply a vertical load to the stud all of the engaged threads share the load equally. When you torque a head bolt into a threaded hole in a block the threads stretch by various amounts as they are engaged under load and the resulting vertical load is not shared equally among all the engaged threads.

When you thread a nut onto the head end of a stud you should be able to engage all of the nut threads using only your fingers and then as torque is applied to the nut the engaged threads on the stud will tend to uniformly share the load with the nut threads but even if after the final torque on the nut is reached the load on the stud and nut threads isn't quite uniformly distributed this is of no consequence because the stud and nut are made from high strength steel and the cast iron threads in the block are the weakest link.

 

OK. I'm understanding your perspective with this info above, and you are correct in regards to where you and I are taking different positions on the equality or inequality of forces based on number of engaged threads. It's easy to understand that the shearing forces imposed on the bolt and/or stud increases as the number of engaged threads increase - simple friction based on contacted surface area between the male and female thread surfaces.

The axial loading strength also increases with the number of engaged threads, but the total axial load is NOT equal for each thread. The thread geometry has a lot to do with this as well. The only way to combat that reality is by altering the geometry of how threads are cut into the block (or nut). Finite analysis science has already revealed that conventional 60° thread cuts allow the first thread alone to take as much as 34% of the total axial loading.

I've provided several links to articles which explain and demonstrate that my proposition of unequal loading is clearly understood in the field of finite element anaylsis, and is not a new or wild theory. Even with improved threading geometry, the loads still do not carry equal loads for every thread engaged by the bolt or stud - higher loads are still experienced at the entry point and less at the farthest end of the stud/bolt's last engaged thread.

Ramp breaks common thread of fastener failures | Machine Design

Notes on Nuts and Bolts

Simulating bolts in finite-element assemblies | Machine Design

All of that being said, I can still agree with you that a bottomed out stud can probably carry more load than a partially threaded bolt that does not bottom out for the same number of engaged threads. However, using an improved thread geometry can increase that stud's clamping potential even more by allowing more threads to carry higher loads while simultanesouly reducing the liklihood of stripping out the threads in the block.

 

I'll add the following just to make sure we're not getting "cross threaded" here! Yes it's true that threads in general are more complicated than in my idealized example but the question being addressed in my original post was do head studs have an advantage over head bolts? And the answer is yes they do but it's not because of their "220,000 psi tensile strength" which is the reason stated in the ARP sales pitch!

A head bolt provides less clamping force than a head stud does because when you torque a head bolt into a threaded hole in a block the threads in the block stretch by various amounts as they're progressively engaged under load by the bolt threads and this stretching action doesn't occur when you finger tighten a stud into the block.

Non uniformly pre-stretched threads are always weaker than the alternative of finger tightening a stud into the block and then applying a load with the stud nut. Of course engaging all the available threads in the block with a stud instead of only some of the threads with a bolt is also stronger.

When I installed studs in my racing engines I used red high strength "Locktite" thread locker so that in addition to being threaded into the block unstressed the studs were also glued into the block as well and only after the Locktite cured for at least 24 hours did install the head and apply a load to the stud threads. You can't do this using head bolts!

The stud that ripped loose and stuck in the garage roof pulled all the threads out as clean as a whistle and left a near perfect hole in their place. Doesn't this prove that all of the threads in the block are involved in holding a stud in the block?

The reason the stud ripped loose was because I was applying 50% more torque than the OEM spec called for because I was trying to solve my own problem with blowing head gaskets. The OEM gasket was a layered steel, asbestos, copper and I was blowing combustion pressure into several of the coolant passages.

The solution to my problem was to look at the equation... MSP={clamping FORCE}/{material AREA} psi and to realize that I'd reached the limit for clamping FORCE so I had to reduce the material AREA to achieve a higher psi MSP. So out came the tin snips and I cut away as much of the superfluous material AREA as possible around the outer perimeter of the gasket and this gave more clamping psi on the remaining area.

EDIT ...A useful experiment would be to have someone with a trashed PSD block install some studs using my above procedure and then measure the torque required to pull the studs out of the block. If you used at least 10 studs spaced around the block you should get a statistically significant result.

 

Scott I have seen these same issues in the 6.0 that is known for lifting heads. I did gaskets and studs in one not too long. No leak with pressure tester. No coolant lose in easy driving but hammer down and it would puke. Heavy power braking in the parking lot and I could make the cooling system cap whistle dixie. We removed the cab and tore down the engine. No sign in the gaskets (multilayer SS) and no sign of a coolant in the cylinder ( no clean piston tops) and there was never any smoke from the engine. New gaskets and studs fixed the issue and my customer was able to make his appointment with BTS for his new trans.

Also on the Cometics I wouldn't go there. The head and Deck need a near mirror finish to work withe the MLS gaskets. So unless you have the good machine shop who knows what they are doing resurface the heads and block I would stick with the stock gaskets and put some studs in there.

 

Wow... thanks for the input on with your experience, Tim.

This threads ranks as an excellent educational experience... at least for me, anyway.

 

I'm going to go get the trailer hooked up, and go get Big Blue & Cookie hopefully we will have some answers today.We'll let ya know later today.

 

best of luck to you guy's hope its a cheap and easy fix..

 

I am anxious to hear about what ya'll find.

 

you guys pullin the heads this morning?

I've got daddy daycare duty so I'll be watching to see what you find .

 

Ditto!

 

a friend of mine bought a 03 6L at the local Ford dealership. He had the same thing happen to him. At WOT he would loose all the coolant out the degas bottle. Took it back to Ford and it had a cracked head. Dont know if that helps ya but just my 2 cents. Hope ya get it figured out.

Edit.... He didnt loose any at idle because according to the dealership the boost made the crack in the head open up and when it was idling didnt make enough pressure to leak.