ATS manifold cap screws keep backing out

[rweis]

Very Interesting,



I can get the Belleville info from the distributor easily.



I will need to experiment to get the manifold growth rate vs bolt growth rate. That may take a while to get. I will need to take it all apart (turbo off, manifold off) put the ATS bolts back in, set them, do a heat cycle (how long should the heat cycle be?, 1/2 hour or so?), measure the rotation to retorque on a couple of the bolts to get an average rotation.



Questions:



If I torqued the steel stud to 32ft # which seems is a torque to yield limit. Found the nuts on the stud loose, then retorqued, do this a couple of times, you would eventually break the stud because you are forcing it to stretch longer and longer. Is that correct? Not a good idea!



If you put loctite on the stud the first time and torque it to 32 ft # since the loctite probably will act as a lubricant then the stud is probably very close to its mechanical limit. What is the best way to get the stud out without breaking it? It has been in place for about 15 thermal cycles, but has not been retorqued beyond the orgional 32ft # with loctite cold.



Extremely interesting concepts. Finally making sense.



Bob Weis

 

[RustyJC]

Originally posted by rweis

I will need to experiment to get the manifold growth rate vs bolt growth rate. That may take a while to get. I will need to take it all apart (turbo off, manifold off) put the ATS bolts back in, set them, do a heat cycle (how long should the heat cycle be?, 1/2 hour or so?), measure the rotation to retorque on a couple of the bolts to get an average rotation.

Don't go to all that trouble. If I know the material of the manifold (e. g. , nodular iron, etc. ) and the flange thickness as well as the material of the stud and its loaded length (e. g. , from the flange face of the cylinder head to the nut face), I can thumbnail some calculations.

If I torqued the steel stud to 32ft # which seems is a torque to yield limit. Found the nuts on the stud loose, then retorqued, do this a couple of times, you would eventually break the stud because you are forcing it to stretch longer and longer. Is that correct?

Yes, if it is in its plastic deformation range (i. e. , higher than the 2% yield prestress but below its ultimate tensile strength), it will continue to stretch at each retorque and subsequent thermal cycle until it eventually fails in tensile overload (snaps).

If you put loctite on the stud the first time and torque it to 32 ft # since the loctite probably will act as a lubricant then the stud is probably very close to its mechanical limit. What is the best way to get the stud out without breaking it? It has been in place for about 15 thermal cycles, but has not been retorqued beyond the orgional 32ft # with loctite cold.

Remove the manifold. Double-nut the stud at the cylinder head face. Put some heat on the stud to soften the Loctite and remove the stud.



Rusty

 

[rweis]

Rusty,



I gave the requirements of 32 ft# of holding to the distributor's engineering dept. I am getting the feeling that is the incorrect value.



If the studs are torqued to limit (70,000 psi) at 32 ft # then the correct torque must be someting less than that. Something like the range of 16 - 17 ft # (35,000 psi) dry as you suggested. Would that allow the stud material remain elastic and yet work with the Belleville's to accomodate the changes in length. I am going to use stainless Drake locknuts as the fastener.



The origional information I gave the distributor calculated a requirement of 4 Inconell Belleville's for 32 ft #. If the torque is less than 32 ft # then the number of Belleville's changes. And what configuration for the Belleville's?, >, <, <>, >>, <<, <<>>, etc?



Since I have torqued the studs to 32ft # and stressed them I will change them out.



Will a regular H hardness machine stud work best, or should it be stainless? Would a steel machine stud grow and shrink and work in concert with the Belleville's better than the stainless? If stainless then what grade so it is stong enough and elastic enough at the same time?



Fascinating, absolutely fascinating!



The effective spring rate and maximum allowable travel:



100% deflection is 0. 015 deflection and generates 1625 pounds of force.

The maximum recommended deflection is 80% or 0. 012 maximum travel.



I will get the flange thickness of the manifold this afternoon. I have a meeting I have to prep for right now.



Bob Weis

 

[RustyJC]

Just for drill, I've run some "what if" calculations based on the following assumptions:



Desired stud prestress (cold) - 35000 psi

Belleville washer spring rate - 108. 333 lbs per 0. 001 inch travel (single washer)

Belleville washer max allowable travel - 0. 012 inch (single washer)

Manifold material - cast iron

Manifold coefficient of thermal expansion - 0. 00000655 in/in/degF

Manifold flange thickness - 1. 000 inch

Manifold flange temp (cold) - 72 degF

Manifold flange temp (hot) - 800 degF

Stud material - carbon steel

Stud coefficient of thermal expansion - 0. 00000633 in/in/degF

Stud loaded length - 1. 300 inch

Stud temp (cold) - 72 degF

Stud temp (hot) - 350 degF



Using these assumptions, 3 Belleville washers in series (i. e. , <<<) could be used on each stud with the primary nut torqued to 17 ft-lb using EaseOff 990 as a lubricant/anti-seize to get the following results:



Cold conditions:

Belleville washer stack compression - 0. 008496 inch

Clamping force - 2761 lbs

Fastener prestress - 35000 psi



Hot conditions:

Belleville washer stack compression - 0. 010976 inch

Clamping force - 3567 lbs

Fastener prestress - 45220 psi



In this example, we have not exceeded the travel limitation of the Belleville washers, and we have kept all deformation of the stud within its elastic deformation range well below the 2% yield strength. Therefore, upon cooldown, everything should return to the cold conditions shown above.



Does this help explain how the fastener system should function?



Rusty

 

[rweis]

Rusty,



The manifold boss is 15mm thick.



Bob Weis

 

[rweis]

I'm sort of speechless,



Questions:



The stud will be anti-sieze lubed and then torque on the stud when torqueing into the block is 17 ft #



Then the manifold gasket



Then the manifold



Then a stainless steel washer as a bearing surface for the Beleville washer to ride on so it does not cut into the manifold boss



Then 3 Belleville washers with the skirt toward the stainless steel washer (and manifold boss)



The crown of the Belleville washer stack toward the Drake nut. Does a stainless steel washer need to be between the Bellevilles and the Drake as a bearing surface?



Then antiseize the stud where the Drake will go. Then torque the Drake to 17 ft # and lock it down.



Correct assembly procedures? Especially in the torque of the plain stud into the block part.



Bob Weis

 

[RustyJC]

Bob,



The process you describe should work fine, except I would go ahead and Loctite the stud into the head using 17 ft-lb torque. A hardened washer between the inner Drake nut and the outer Belleville washer is a good idea to protect the Drake nut - I'd use a hardened washer between the inner Belleville washer and the manifold flange as well.



The 15mm manifold flange thickness measurement changes the following assumptions:



Manifold flange thickness - 0. 590551 inch

Stud loaded length - 1. 200 inch (estimate only)



The cold and hot conditions would then be as follows:



Cold conditions:

Belleville washer stack compression - 0. 008496 inch

Clamping force - 2761 lbs

Fastener prestress - 35000 psi



Hot conditions:

Belleville washer stack compression - 0. 0092 inch

Clamping force - 2990 lbs

Fastener prestress - 37902 psi



The outer (lock) nut on the Drake locknut should be torqued to a 20000 psi stud prestress, or something like 10 ft-lb, again using EaseOff 990 as a thread lubricant/anti-seize.



Is this enough information for you to give it a try?



Rusty

 

[rweis]

Yep, YAHOO



An unbelieval THANKS, words do not express the gratitude.



I have the Drake's on the way, new studs come tommorrow, I have the invoice for the Belleville's and they are packed ready for shipment.



By next weekend, it should all be REDONE, IF the hurricane leaves us alone.



Thanks, thanks, thanks,



Sounds logical, and understandable, and makes sense



Bob Weis

 

[rweis]

Belleville disk washer source.



I ordered the ones made from Inconell to ensure it would handle the temperature ranges.



I got 36 (3 for each stud) from Key Belleville, 100 Key Lane, Leechburg PA 15656.



1-800-245-3600



They will sell in small one time lots. I paid by check. Item for a 10mm stud was the A10-X @6. 98. The shipping was ~$3. 00



I called them, they faxed a proforma invoice, and I paid off that.



They were very nice to deal with. Transaction went very smoothly.



Bob Weis

 

[rweis]

Rusty,



Do all the calculations hold the same if the studs are stainless steel? And what kind of stainless steel should it be?



ie 3 Bellevilles, 17 ft #, the elasticticity of the stud, etc



Just wondering for anyone that wanted to use stainless studs that has been following this thread.



Bob Weis

 

[RustyJC]

No, coefficients of expansion, 2% yield and ultimate tensile strengths will all differ between "stainless" steel and standard carbon steel. Unless money is absolutely no object, true cobalt or nickel-based superalloys such as Hastelloy X (major constituents are typically 47Ni-22Cr-18Fe-9Mo-1. 5Co), A-286, Waspaloy, etc. are not really needed and are far too costly for this application.



Typically, a corrosion and heat resisting nickel-chromium steel (SAE family 303xx) or chromium steel (SAE family 514xx and 515xx) might be used for this application. I'd be glad to rerun my calculations if you can tell me what stainless material your machinist is proposing for the studs (assuming I can find the physical properties for that particular material. )



Rusty

 

[rweis]

Rusty,



Not necessary to rerun calculations. I re-read back through the thread to see if anything was left uncovered or unanswered that would be of use to someone else. There were a couple of reply's that mentioned use of stainless steel as a stud material. A "common" stainless steel stud material.



I am going to check on the different choices of steels and the makeup of the steels. I will ask about the heat resisting nickel-chromium steel or chromium steel availability. When I get the specific makeup I'll pm you.



The thread seems to be complete as no additional post have been added for a week or so.



Bob Weis

 

[drrhoades]

Just a question. Are you running these calculations by hand or do you have program on a computer that is doing it for you and you just plug in the data. I'm sort of curious becuase I have done similiar calculations by hand for school and they took forever even after I figured everything out and understood it.



Doug

 

[RustyJC]

No, I have some Lotus 1-2-3 (that tells something about my age ) and Excel spreadsheets that I've made up for fastener prestress calculations here at work.



Rusty

 

[drrhoades]

I was hoping that you had something like because I know I definitly would not like doing that stuff by hand to much, without knowing exact specifications on the parts being used.



Doug

 

[rweis]

Well finally assembled the parts, took a day off work to do this and a couple of other things.



Fit:



Fit was good. There was BARELY enough room on the ATS to get the Bellevilles diameter wise on a couple of studs. Had to loosly put it all together and had to rock the ATS upward to put #4 lower package on the stud.



Stud length was excellant and all stud stacks have about 3 turns after the Drake lock nut is on.



Will check torque in a week, month, 3 months and report back.



Bob Weis

 

[rweis]

Reporting Back,



Torque is holding on all Drake nuts and on all lock nuts. No leaks, no evidence of soot around the manifold gaskets.



This solution seems to work.



Bob Weis

 

[RustyJC]

Bob,



Does that mean you'll pay my invoice for consulting engineering services now? :-laf



Rusty

 

[rweis]

Sure,



I'll buy you a hotdog and a drink at any Cleveland Indians game here in Winter Haven. I'll even buy your game ticket!



LOL



Bob Weis



ps. Without RustyJC's expertise, ideas, and calculations, I could not have done it. I DEFINITELY owe him one. I want to publicly thank him for his ideas that worked magnificantly!



"just LOVE it when a plan comes together"

 

[RustyJC]

Originally posted by rweis

Sure,



I'll buy you a hotdog and a drink at any Cleveland Indians game here in Winter Haven. I'll even buy your game ticket!

Sounds good. The last Cleveland Indians game I saw was in the old Lakefront Stadium in Cleveland just before they tore it down!



Glad I could be of help.



Rusty