Why are these extension housing bolts sheared?
 

This is my beloved Quaife box:

So, upon finding some oil at the bottom of the transmission - I quickly found some extension housing bolts missing.

Upon further inspection, most sheared and left 2/3" thread in the housing. I think only 2 are somewhat 'OK', 4 I've removed so far. The other two are impossible to reach, or are not fully through - making it hard (impossible?) to remove the thread remains.

Did they shear because
a) they were reused, which should not be done?
b) they were overtorqued
c) my car makes to much torque (PPF, and such)
d) bad carma?

A) Like most other bolts, you should be able to reuse them a couple times.
B) I see no necking, i doubt it's from over torquing.
C) We've seen a million other failures from power, I don't think I've ever seen this failure.
D) I vote this one.

You said they were reused, was the gearbox taken apart recently? I'd vote first for missing dowel pins and/or UNDERtorqued bolts. This let it twist under acceleration, and shear the bolts.

Hmmm....I can only speculate here. Does the housing have dowel pins that help align the trans properly? Are they possibly missing?

Does one housing twist enough that it could have hammered the bolts enough to shear them? Perhaps they were not tightened properly.

That is my best conjecture.

Edit: Curly is faster than me.

I'm not sure but I'm fairly certain that miserable, POS company Quaife is behind this.

Right... helpful :)

Are they the OE case bolts? A high grade bolt would fail like that from over torquing. Whereas say a grade 5 would stretch.

Don't these bolt thread into aluminum? I would finger the aluminum threads failing before a high grade bolt reached failure. I'm also having trouble placing these, do they hole the tail shaft housing on?

Yea, OEM grade 8 bolts. They go into the alu, which is still fine.
These are the bolts:
(you're looking only at the back half of the transmission, not the bell house)

Can you get a shot point down the length, like so we can see the face where the break was? From the last picture it almost looks like they were loose and sheared.

Like diz?

IMO it looks like over torque on high grade hardware. I realize that Grade 8 is used there as the clamping force is needed, but I suspect they've be torqued before (multiple times?) and gave up this time around. So I vote (a).

I'll say undertorqued, but that's a non-expert opinion.

I'd buy new bolts, lube & torque them per mazda specs. Make sure dowels are used too. You should be fine if the threads in the aluminum are not damaged.

I've now been able to replace 6 out of 8. The other 2 I cannot reach/remove the remains.

From the few that were still tight, I can say those were not undertorqued. I was afraid they would snap too.

The re-use multiple times - could have been the issue.
The torque spec is 34Nm.

I'll report back after the next run or two.

Those pictures make it look like the bolts failed by fatigue, high cycle fatigue specifically. But not normal bolt fatigue. If the bolts were torqued correctly they would generally fatigue at the first thread (first from the bolt shank). And the surface would look different.

The fatigue looks like it is an alternating bending pattern.

Blah blah blah, I could go on and on BUT

To me it looks like they were undertorqued, or they had lost torque. The undertorque condition led to the bolts being put into a cyclic bending load, one they are not meant to take. I'm guessing the other ones didn't fail because they are torqued correctly.

That is my professional failure analysis and there will be a $2000 consulting fee. I'll PM you how you can settle the fee...

Thanks for the analysis. There are two options
a) Torque maybe was OK initially, then they loosened (vibrations?) and fatigued
b) Overtorqued combined with re-use of original bolts

Either way, I should be fine with
a) Resbond in threads
b) Proper torque and new bolts
Downside: only 6 out of 8 bolts.

There is no more oil sweating from the transmission, so it's tight now.
I'll report back

Trust me, they did not fail because they were over-torqued. It is possible that they were torqued correctly, and then they lost torque for whatever reason (vibration, reused bolts that were stretched, etc.).

In my experience, that sort of failure surface would not be seen in a proper torque condition.

Regardless,

Yes, this should solve your problem. You might also want to add "check extension housing bolt torques" to your inter-race maintenance list. Resbond should prevent the bolts from losing torque, but it would be foolish NOT to check them periodically since you've had these failures.

Thanks for your input, it's on the list already :)

I'm replacing these bolts faster than changing oil.
I need to check my motor mounts. Something is killing them.

Is that part supposed to have any locating pins like the transmission bell housing?

Like maybe one here, and again on the other side?

https://www.miataturbo.net/attachmen...ine=1379610475

I had one of my fuel rail bolts separate in the middle of the threads like that. I swear I threaded it all the way down by hand. The second I put a little snugging torque on the wrench it felt funny. I pulled it out and it looked just like that. No idea what happened. Called FM and they said they had never heard of that happening. Bad bolts maybe? Maybe try a little strip of white nail polish where the bolt head meets the case and see if the thing is backing out a hair. Vibrating funny.

My vote is for missing dowel pins allowing the tranmission to rotate relative to the block under power. The bolts which hold the transmission to the block using compression are not designed to prevent the two rotating against each other under torque. In fact, the bolts fit so losely through the transmission side holes before engaging in the engine side threads that I wouldn't put it past the engineers to have designed the part such that the shank of the bolt doesn't touch the inside of the bolt holes on the bellhousing. The dowel pins are designed to be nearly press-fit which virtually eliminates torque hammer between the two parts.

I have seen similar failures on heavy machinery due to an under torqued condition. It allows a shock load to break a bolt in a way that hundreds of pounds of torquing could not accomplish. But I could be wrong in this instance. All I can do is speculate.

After a quick flip through your build page, it looks like you've got a very stiff chassis with very little flexibility. A system can be over-restrained to the point that it becomes 'infinitely rigid' compared to the components it connects to. When this happens, loads get redistributed to the least rigid components and it leads to failure. I'd like to say that you're chassis is so stiff that the lack of chassis/ppf flex put excessive tension on the bolts and failed them.

Any update to this #Laur3ns? I know this is an old tread but what did you end up finding out? and doing about it? I have the same issue

Dowel pins were not missing of course. Some 12.9 replacement bolts have taken care of this issue.

There is a guy in the UK who went through the same story and it worked for him too. Quaife apparently now includes better bolts in their gearset.