Turbo bolts keep breaking

Cat bolts? Proper torque? Mounting flange warped?

 

I had this problem a couple years ago with new manifold, two different new BW turbos, all Cat hardware, proper torque etc, etc. Happened about 4 times in 3 mos. Finally broke down and called a good old school Cat mechanic I know. Here is what he told me - when installing your bolts pick up on the turbo so that any slop in the holes is in the opposite direction from which gravity wants it before you start snugging and torquing things up. Use a small jack if necessary. (Was for me) . So I put a chunk of soft pine 2x4 on frame rail and raised it with a little bottle jack carefully after finger tightening everything. Then proceeded to torque to spec. Never had another issue in the next 3 mos before getting rid of truck.

 

From the two I see. Your exhaust manifold studs are to short, not even engaging lock nut. I would question a vibration. Do the manifold gaskets last, or leak fast.

I do agree with @w9l. Starting flush and not using bolts to pull the weight up is a reasonable idea. Maybe one of the holes are not drilled straight.

Do you have a pyrometer to know its not excessive heat problem?

 

That is what I thought too. I checked both surfaces for flatness with a piece of ground flat tool steel 1/2"x2"×4" with a piece of 400 wet or dry adhered to it. Spray paint both surfaces lightly, let dry then carefully sand off using even pressure and contact. The high and low spots will be evident. Stick shop paper towels in runner and se a little wd40 or the like to keep any dust contained. The corners where the bolte suck in will usually show to be the high spots. Take some cheap bolts and install turbo as above but without gasket .Evenly snug bolts and see what the largest feeler gauge is you can slide in the gap. If it isnt over 4 or 5 thou, I wouldn't worry about it. I also checked to see if holes in both flanges were reasonably straight. And I did use Cat bolts and nuts 3 of the 4 times. The aftermarket bolts were all I could get one time and they held as well as the Cat. The Cat bolts, nuts and gasket cost me between 90 and 120 bucks each time. And I would never recommend using anything else. Btw, you need a torque adapter for offset to properly torque those, or atleast I did.

 

Use Cat high temp bolts and nuts, plenty of anti-seize, and torque to 41 ft/lb. By Cat manual you are not supposed to final torque mounting bolts until turbine housing is correctly rotated for oil line installation and v band is torqued to 10 ft/lb. Are they shearing off?

 

Been there done that….. Genuine CAT bolts will solve your issue. Also, when I had the New Cat head replaced, I bought all new Cat studs and nuts. More money with good insurance.

 

This type of fastener failure is due to inadequate preload (stretch). It is extremely difficult (almost impossible) to use a click indicator type torque wrench within the engine compartment to get anywhere close to proper bolt preload. Torque is required to overcome both bolt stretch and friction. If the friction is too high, it will cause the applied torque to read much higher than the torque used to stretch the bolt. If all the metal surfaces are new, and all the surfaces are lubricated, the friction component of the torque will be <10%; but even clean, non-lubricated fasters may require 50% of the torque to overcome the friction.
I'll use some numbers as examples; these are not correct but they show the concept. If we need 6000lbs of preload on these bolts, the torque without lube would only apply 3300lbs of preload which will always fail if the load repeatedly exceeds 3300lbs.
So if normal turbo vibration applies 20,000 pound of force to the bolts, the 4 bolts with 6000lbs of preload will never change dimension as 4 x 6000 is 24000. The bolts with 3300lbs of preload x 4 is 13,200 so the 20000 lbs of force will stretch and release the bolts repeatedly until the bolt fails. It works just like a paperclip you bend back and forth until it breaks in half.
To make matters much worse, the click indicator torque wrench will definitely measure the torque, but torque is horrible at indicating bolt stretch. Torque is force required to stretch the bolt and friction, but friction is tricky. Static friction is about 4x greater than kinetic friction. This means that if you stop turning the torque wrench before it clicks, it'll probably click as soon as you start again. While that is definitely 41ftlbs of torque, it is not nearly enough bolt stretch. Most of that torque is actually friction. So the bolt wont come loose but instead will fail.
So here're my recommendations;
Use new nuts, bolts, and washers. Everyone will hate this but the brand of bolt is irrelevant, so long as it meets the spec. Fastenal can hook you up with a whole box of bolts for the price of 4 CAT bolts.
All surfaces that contact the part of the fastener you're going to turn, not just the threads, must be very, very clean; good as new.
Use liberally applied anti-seize as the lubricant everywhere friction may occur.
Hand snug each.
Torque each to 10 ftlbs.
Mark each bolt head and nut for an index. If possible use micrometer (probably not) to measure the length of the bolts.
Torque each fastener to 41 ftlbs or whatever spec is. Ensure you have clearance to continue rotation to achieve a continuous rotation until wrench clicks. If you stop or slip remove that fastener and start over.
Measure the angle between the 10 ftlbs index and the final torque location. Ensure the measured rotation between the highest and lowest 4 bolts is within 15 degrees. If it's not, something is not clean or not lubricated so bolt preload was not accomplished and it will fail. They make an indicator wheel for that but I don't know how you get it into the engine compartment.
If clearance is insufficient to go from 10ftlbs to 41flbls in one swing try 20 ftlbs. No matter what else happens, be absolutely certain the fastener was turning when the wrench clicked.

 

You're right, I don't like that you feel any brand of bolt will suffice. Why not just use grade 5, they're cheaper yet?