ATS manifold cap screws keep backing out

[rweis]

Well, I solved my saga by going to studs (10x1. 5x47mm), MANIFOLD washers (3/16"), lock washers (not totally convenced on these, time will tell), and nuts. Studs so only the nut is movable, not the whole fastener, and torque can easily be tactile monitored.



When I took the cap bolts out almost every one was loose to hardly torqued. EVERY gasket had soot blowby on it.



I put in the studs with loctite (torqued to 32'# with a double jamb nut for the wrench), manifold washer, lock washer and nut. The stud did not have ANY interference with the manifold and left about 3/16" of stud showing when all components were assembled.



The stud system also had the side benefit of securely hold the manifold gasket during assembly.



I can use the 3/16" exposed stud as a tactile check for the hard to see studs under #3, #4, #5.



A wrench was a lot easier to use than fighting the water hard lines with the allen socket, and you can't really get a perpendicular well seated drive into the hex caps on #3 upper and lower #4 upper and lower. You can get a well secure wrench contact with a crows foot (and you do have to watch your lever arm length as discussed in other threads) on these and a deep socket on the remaining.



I plan on ordering the Drake fastener to replace the lock washer nut combination, but the lock washer and nut should get me through until that is ordered.



Thanks to all who gave their ideas. I think this is a pretty sound fix.



Thanks Again,



Bob Weis



10mm split lock washers @ . 10

10mm nuts @ . 20

10mm manifold washers @ . 30

10x1. 5x47 stud @ 1. 84

12 sets @ $29. 28 + tax

 

[rweis]

Drake lock nuts



Lok-Mor

800. 843. 7230

10x1. 50mm

Class 10 / Plain recommended due to high temperature

$5. 62 @

Backordered for 2 - 5 weeks



Very pleasant to order from and will take an order from an individual as well as a company



Will bring this back up when they come



Heat observations of split ring lock washers



After running about an hour at 65 I stopped and checked the split ring lock washers visually to see how they were doing. The gap in the split had gone from 1/32" to about 4/32". When it cooled down overnight the split was back to 1/32"



There is DEFINITELY some movement going on with the split ring lock washers. The nuts did not noticably move (I marked a line of nut, lock washer, manifold washer, manifold). I will check them and bring the thread up if I see them move.



Bob Weis

 

[RustyJC]

The Drake locknuts should provide a fastener that won't back off in service. To maintain a controlled assembly preload with thermal expansion and contraction, however, you might want to use Belleville washers between the Drake locknuts and the manifold.



Rusty

 

[rweis]

What temperature do the holding surfaces of the exhaust manifold reach? I guessed 500* - 1000*. EGT can be 1200* for hours, there must be some disipation through the casting. I have seen pictures of cherry red manifolds and turbos from heat. Was I close?



Heat of over 500* pushes the material to in-ka-nell (major guess at the spelling of it). 10mm, 32'# and there is a drop in the torque as the belleville disk spring heats up during the cycle. I get the idea that you have to go over 32'# to like 40'# because there is a 40% reduction in the torque as the disk spring takes on the heat and the expansion. At 40'# that would leave 24'# at the worst case of holding.



Does anyone have any experience with these? I'm starting to feel more under water than above water. HELP!



Any idea of the $$ for these things. I wanted 25 to be able to double up on them with 1 spare. They are trying to see if they have some of a production run overage. Production runs are multiple of 1000 units, a few more than I need for this application.



Interesting roller coaster this one is.



Bob Weis



ps They thought the Drake lock nut was a good choice.

 

[RustyJC]

I think you're referring to Inconel. From your description of the split ring lock washers, it sounded like your stud/nut assemblies were losing preload under operating conditions. With properly sized Belleville washer assemblies, you could retain more clamping force when everything heated up than is possible with a nut/stud or conventional bolt-type arrangement.



Yes, we use Drake locknuts (crankshaft counterweight, main bearing and connecting rod cap fasteners) as well as Belleville washers in the reciprocating engines, reciprocating compressors, turbochargers, gas turbines and centrifugal compressors our company has built.



Is it overkill? Probably. Is it better than conventional fastener technology? Our experience would indicate that such is the case.



Rusty

 

[rweis]

Rusty,



Trying to do it right, once. Actually have come down this path finding that what seemed right was a reasonably far distance from it.



I will hear back form WestCoastLockWasher (WCL) tommorrow on what is available. Should be interesting.



Am I even close on the temps (500* - 1000*)? The engineer at their end thought anything over 500* = Inconel. He is going to try to find a suitable disk spring to do the job. Seems there is a equation based on thickness, amount of deflection at operating temperature, and diameter of the spring itself and it all comes out based on a % of orgional torque (I think).



If it runs in the "stuff" your company builds, then it certainly ought to hold a exhaust manifold on. I find some overkill prevents going back and doing it again and again and again and ...



The Drakes are ordered, we'll see how the Belleville's turn out tommorrow.



I have learned more about high temp fasteners in the last couple of days than I thought I ever would. Interesting though.



Thanks for the ideas and input,



Bob Weis

 

[RADdodge]

Originally posted by rweis

Drake lock nuts



... Class 10 / Plain recommended due to high temperature...



Bob Weis

Bob, plain as opposed to what - stainless? I noticed several posts that mentioned using stainless steel for the studs, nuts and washers. It would seem a good place to use stainless considering the cast iron's propensity to rust. It looks like the Drake Lock nuts come in stainless. I would like to try what you are doing, but would like to use stainless steel. Would a stainless stud be rated the same class 10? The Belleville washers look like the way to go. It will be interesting to know what you find on those. Nice work.

 

[rweis]

RADodge,



I do not have the answer on the Drake lock nuts. The Drake rep thought any plating would be destroyed by the heat. Therefore no plating. Stainless would probably be fine.



I struck out on the first Inconel attempted purchase. That company did not carry small amounts of stock, and a production run was out of the question $$$ wise.



I found a second company that does carry small amounts of stock and will sell to an individual. To get 32 ft lbs of holding takes 4 Inconel disk springs for each stud. The inconel is pricey at $7 per each disk or $28 per stud times 12 studs. The engineering group was very helpful, I told them I needed 32 ft lb of holding on a 10mm stud in an environment of 500*F - 1000*F and was going to use a Drake lock nut. They looked up the requirements, found the right disk spring, passed me off the sales, will fax the invoice in the morning, will take a check, let it clear, and ship the product. They will also do a credit card, but I much prefer to send cash and know the transaction is complete.



I will reveal the source when I get the transaction complete. I told them they might get several calls to solve the same problem.



Like RustyJC mentions, there are very few true solutions to this.



I am starting to see why Cummins puts a smallish manifold (so the large spacer will not create interfearance) on the CTD then the large spacer to absorbe the stretch and contracting and the bolt strap to keep the nut from turning.



I would rather the ATS manifold, but it also creates an expense (~$400 (48 Belleville disk springs, 12 Stainless Steel flat washers for the Belleville disk springs to "float" on so the Belleville's do not cut into your manifold, 12 Stainless Steel studs, 12 Drake lock nuts) to get it bolted up correctly.



I am tending to think the 10x1. 5x47mm stud MAY be a little short in length. The Belleville's will take up . 327" + a SS flat washer + the Drake nut. Will just have to install one to see. I put in steel studs and probably will change them to stainless steel studs if I have to replace them.



Looks like the solution is right around the corner.



Bob Weis

 

[illflem]

Wow Bob, 10,000 years from now when some archeologists dig up your truck the only thing left will be the manifold bolts!

 

[rweis]

Bill,



Nope, in 10,000 years I'll still be tweaking and changing oil and filters,



Bob Weis

 

[rweis]

Rusty,



Do y'all typically stack several of the Belleville's?



Is the concept that each Belleville will take xx torque and then you stack them to get the torque you want? If so what about if you wanted 80'# you would have to stack 10 of them?



Your applications surely has significant torque values on them. How do you use the Belleville's?



Bob Weis

 

[RustyJC]

Bob,



Pardon the delay - I just got back into town from a business trip.



I've been reading through this thread, and something's not making sense here! I need to get this clarified before I can make any recommendations.



According to my screw thread tables, a 10mmx1. 5 external screw thread has a minimum diameter in the thread root of 6. 063mm, or 0. 238701 inches. The thread pitch of 1. 5mm translates to 16. 93 threads per inch. I don't have any metric torque tables handy, but the closest SAE equivalent fastener I can find is a 5/16 - 18 fastener which has a minimum diameter in the thread root of 0. 2431 inches and 18 threads per inch - pretty darned close to a 10 x 1. 5 for discussion purposes.



Assuming you're using a high-strength alloy steel stud and nut material, we normally shoot for a prestress of 35,000 psi. This would generate a clamping force for each fastener of 1,566 lbs for the 10 x 1. 5 fastener (35,000 psi x 0. 04475 sq in area), or 1,625 lbs for the 5/16 - 18 fastener (35,000 psi x 0. 046415 sq in area).



My problem is this - my torque tables show that, using EaseOff 990 anti-seize as a thread and face lubricant to prevent high temperature galling, a 5/16 - 18 fastener only requires a torque of 7 ft-lbs to produce a 35,000 psi prestress. A torque of 10 ft-lbs would produce a 50,000 psi prestress. Where this analysis comes unhinged is when I see you guys talking about torques of 40-45 ft-lbs elsewhere in the thread.



In that 5/16 - 18 fastener, a torque of 42 ft-lbs would generate a prestress of 210,000 psi, far beyond the yield and ultimate tensile strength of the fastener material. For example, normalized 4140 steel has a 2% yield strength of 95,000 psi and an ultimate tensile strength of 148,000 psi (ref Machinery's Handbook, 20th Edition, page 2106).



Before I can calculate the clamping force (which should give you the information to determine the number of Belleville washers needed), I need to get this discrepancy reconciled. Something just isn't adding up!



Rusty

 

[rweis]

Well,



I did not antisieze the stud into the block, but used Loctite blue to hold the stud from migrating. I got the 32 ft # from the bolt torque orgionally used by DC. Which may or may not really have any real meaning. I torqued the stud into the block at 32 ft #, and torqued the current nut onto the stud at 32 ft #.



I think the 40 - 45 ft # they are talking about is using a stainless steel stud and tightening it "enough" so that it does not move, ie the fix up to now.



I also think what is happening is individuals are trying to see what "works" and go with that, rather than gathering the parts that go together as a system and installing tham as a system.



So, with that, which I hope is helpful the ball is back to you.



Bob Weis

 

[RustyJC]

Bob,



Can you confirm that the stud size is 10mm diameter x 1. 5mm thread pitch? The torque values you mention still look far too high for the data I posted previously. Either the stud size or torque almost have to be wrong. Does the stud look just a little bigger than a 1/4" - 20 stud?



Rusty

 

[rweis]

Yep, 10mmx1. 5mmx47mm long.



The torque value I used was what the OEM bolt had on it. Basically I guessed.



The 32 ft # I am sure of, used an accurate torque wrench.



Bob Weis

 

[rweis]

I am guessing as to procedures and values.



Just because the OEM torque is 32 ft # on a bolt does not mean that it is right for a stud.



Just because the OEM torque is 32 ft # on a bolt also does not mean it is right on the nut on the stud.



That's why I'm trying to figuer out how to do this right.



I hope I have not screwed something up yet that can not be undone.



Bob Weis

 

[RustyJC]

Bob,



If you have the original manifold bolts, can you mic the diameter of the unthreaded shank just below the bolt head? Inches or millimeters - either will be fine. I just need a diameter to match to my bolt tables. If you don't have an outside micrometer, measure the circumference as closely as you can, and I'll go from there.



Rusty

 

[rweis]

I will do that, takes about 45 min to get home, kiss the wife, mic the OEM bolts.



Will be right back (45min)



Bob Weis

 

[rweis]

The OEM bolts have a . 3888 inch shank.



The shank is as large as the OD of the threads.



Does this help?



Bob Weis

 

[RustyJC]

Bob,



0. 388 inches equals 9. 856 mm which is within the major diameter tolerance for a 10 x 1. 5 fastener (10. 000 mm max, 9. 850 mm min). According to my thread tables, the acceptable minor diameters at the thread root of a 10 x 1. 5 fastener are 8. 160 mm max, 7. 967 mm min, or 0. 3216 in max, 0. 3137 in min.



Correlating this minor diameter data to UNC fasteners, it falls midway between a 3/8"-16 (0. 2983 in) and a 7/16"-14 (0. 3499 in). From my torque tables (lubricated threads), the torques required to generate a 35,000psi prestress are as follows:



3/8"-16 - 13 ft-lb



7/16"-14 - 20 ft-lb



Therefore, by extrapolation, I'd expect a torque of around 16-17 ft-lb to generate a 35,000 psi prestress on a 10 x 1. 5 fastener. I don't know what the material of the OEM manifold bolts is, but for sake of argument, let's assume that it's SAE 4820 normalized steel - a recommended material for heavy duty bolts (ref Machinery's Handbook, 20th Edition, page 2100).



SAE 4820 normalized steel has a 2% yield strength of 70,250 psi and an ultimate tensile strength of 109,500 psi. A factory torque specification of 34 ft-lb (lubricated) would, therefore, produce a prestress of approximately 70,000 psi, which in this case would make this a "torque to yield" fastener!



This correlates with your complaint that the bolts are loose after the manifold thermal cycles. If the manifold grows more than the bolts, since the bolt is already torqued to yield, it will stretch the bolt more, and since the bolt is already at its 2% yield point, it will not act like a spring and return to its original dimension when it cools - it will remain at the "longer" hot dimension (it was stretched in its plastic versus its elastic deformation range. ) Therefore, the bolt is loose when everything cools off.



That's why the stock manifold uses the long spacers. They give the fastener a longer loaded length, which means it stretches less per inch of loaded length than would a shorter fastener when the manifold is thermal cycled. For this to work, however, the stock bolt has to be below its yield point when cold; therefore, I can only surmise that the stock torque spec of 34 ft-lb must be for a dry, not a lubricated, fastener. Otherwise, if the bolt were torqued to a 70,000 psi prestress (the 2% yield strength), it would not return to its original length when the stock manifold cools off.



By using the Belleville washers, we can let the spring function of the washers provide the clamping force while allowing for expansion and contraction. To do this, we need the Belleville washer(s) on each fastener to provide adequate clamping force when cold, but also adequate travel to allow for hot expansion of the manifold. If your vendor can tell you the effective spring rate and maximum allowable travel of the Belleville washer you're going to use, I'll try to run some quick calculations to identify the quantity and configuration of the Belleville washer(s) to be used on each fastener.



Another piece of information that would help is to know how many degrees of rotation were required to retorque the ATS bolts after they had stretched and you found them loose. With this information, we can do some rough calculations of how much the manifold is growing over and above bolt growth when it thermal cycles.



Rusty