Look, studs are cooler than the flange:
https://www.miataturbo.net/forum/t38220-2/#post443324

 

A cross-post here from the mirror topic at M.net, but did we abandon the belleville washer idea?

Seems that if one can be found that will retain its springiness at the temperatures we are dealing with (and is strong enough that we can apply enough torque to properly seal the turbo-manifold flange without completely flattening the washer) then it ought to solve that bit the the problem, yes?

As an example, these fine folks have spring washers rated for use at up to 700°C. From their own description:(my emphasis)

 

Thermal simulation = good
There was much rejoicing.... Yay....

I'm not a big fan, but I'd be interested to see someone try it.

 

I'd say I'd try it, but I'm not running uber lap times. So even if my **** never failed it wouldn't be a good enough test. So far Sav and Hustler are going Tial so the potential pool of candidates has dwindled further. Looks to me like the issue will basically go unsolved.

 

Any reason in particular? Seems like William of Ockham would approve of the idea.

I'd be happy to put a few on my current system, though of course I'm not presently having problems with my bolts loosening as it is. (I'm also not doing trackdays.)

 

If my through-bolt setup does not work (due to yielded bolts) then I may try Bellevilles next, made from a high temp alloy like A286. I could probably fit them on the manifold side of the flange under the heads of the bolts. No way they will fit on the turbine side.

So far my through-bolt setup is holding just fine at 20 psi boost pressure, but this is just on the street.

 

In a most-extreme situation, it seems that the bellevilles, torqued to appx 50% yield, plus Travis' bolts (or studs) made from 100% pure ridiculoum, ought to hold fast no matter what you throw at it.

 

Yep, when unobtanium does not quite do it, then ridiculoum is the next step Seriously though I tend to agree with you. We'll see. Hopefully what I have now will do it.

Joe, check your PMs by the way

 

I'm just most concerned with bottoming out of the bellviews. You need a really strong bellview to make it work. If you can find that I think it would do ok.

 

17 ft-lbs to bottom an M8 Belleville; maybe back up 10 mils then Resbond to keep it in place. Read my posts:

turbo bolts wont stay tight!! - Page 12 - MX-5 Miata Forum

 

stack them until you reach the load point you want.

 

Well, Y8s kinda nailed it- two washers stacked in the same orientation doubles the yield strength, two washers stacked opposite doubles the available spring travel for a given bolt torque.

If it takes 17 ft.lbs of an M8 to flatten one, then I guess the question is how many ft.lbs of that same fastener does it actually take to hold a turbo on? Assuming the bolt is prevented from rotating by external means (serrating the edges of the washer, using safety wire, pinning the threads, etc) does it need to be torqued that tightly?

 

Truthfully 8-12 ft lbs will do it. The bolt preload must not exceed 10000psi, or it will overwhelm the bolt at load and temperature. You must use a quality stainless fastener, although the spring washer maybe any material that does not harden/weaken significantly with heat cycling.

 

Pats is saying 10ft.lbs at M.net, which I can believe. Even if we are only able to torque to 5ft.lbs, and the turbos leak a little at idle, once you get them warm they should seal up.

 

Here's my update after a weekend at AAA Auto Club Speedway with Sav. With going 140mph on the bowl at WOT for about 30-40 secs, my turbo brace and studs and nuts are still good. Chalk up 5 trackdays plus 1 auto-x on this setup. I think it's now a record.

 

No if you stack the Bellevilles the force to flatten them is the same, but if you back up to half-compressed say, there is more room for the flange to grow.

 

That depends on how you stack them Jason.

<<<<<<

or

><><><><><

 

Roughly speaking, if you stack two belleville washers together in the same orientation, the force to flatten them, or to deflect them by a given amount, doubles. (It's not a 100% linear relationship due to friction between them, but it's close enough for our purposes.)

Conversely, stacking them in opposite directions will cause each one to deflect by the same amount that a single washer would for a given clamping force (so the total deflection will be doubled) and you will have twice as much capacity to absorb expansion relative to a single washer.

edit: Neo beat me to it.

Agreed however that the correct technique is to set the fastener to a point that the washer is partially compressed, so that you are actually applying some clamping force, yet still have room for expansion. This seems intuitively obvious to me- was there a contrary opinion somewhere that I missed?

I wonder if there is a difference between torquing down the bolt to full flattening and then backing it off to achieve the desired deflection vs. simply torquing it down to achieve the desired deflection without flattening it first?

Reminds me of the procedure for half-masting a flag.

 

dont forget there is also >><<>><< if you want higher force AND deflection.

I had to email mcmaster.com and tell them their technical description was incorrect for alternating washers. ><><><>< is springs in series and will not equal the rate of one washer. I thought they fixed it but it is still wrong. effective rate is k/n where k is rate and n is number of washers.

 

Not from me. That is the way to do it IMHO.