35 replies to this topic

[Brandon]

I will be starting a separate thread capturing my thoughts on this past weekend's activities and related failures of equipment - specifically the "SDMH" (Super-Duper Mazda Hub).

I'll try and include as many details as possible like Jim notes - curbs, strapping-down for towing...

 

Suffice it to say, there were two brand-new LF SDMH's that failed at a track with some significantly long right-hand turns of both high & low speeds but definite full-loading of the suspension at that corner and neither car owner/operator/maintainer is lackadaisical about the details in the care & feeding of a SM.

[LarryKing]

 

one of the biggest is under torqued hubs for those trying to get the car to roll easier. If the hub is under torqued, it will fail VERY quickly.

 

Ah ha! Now we might be getting closer to an actual cause.

[Brandon]

While I don't have the FSM handy, I do have the inter-t00b!  So grain of salt and all that....but consensus appears to be 123-159 ft#s and I know I torqued my SDMHs to 140 ft#s so a bit below 'middle of the spec' in my case.

 

Anyone else have an opinion on the number to torque to?  I had no thoughts about the level of 'free spin' of the hub since the SDMHs have the revised inner seal with the metal flange covering the rubber bits so there would have been additional drag there regardless...

[Bench Racer]

While I don't have the FSM handy, I do have the inter-t00b!  So grain of salt and all that....but consensus appears to be 123-159 ft#s and I know I torqued my SDMHs to 140 ft#s so a bit below 'middle of the spec' in my case.

Your consensus grain of salt has value. 1990 FSM front axle torque, 123-159 ft#'s.

[Jim Drago]

We started at 125.. we torque all to 150-155 now ( yes even mine)

[KW78]

Understanding the hubs that don't fail I believe is the key here.

 

Hub bearing failures are an issue, hub flange failures I have never personally seen with the Miata.  The rx7 failures in EP were the steel hats failing, and the replacement kit used the original bearing setup.

 

Random thoughts on the subject:

 

Tapered bearings are better for what we do, they have a line to transmit forces instead of a "point" that the ball bearing style has.  You have keep up with and set preload tho, and that is done dimensionally with the ball bearing style, therefore a wide torque range.

 

The flange failure pictures I have seen here I believe are absolutely from installing studs wrong.   Studs should have MAX .005 interfenence.  Own the right ream for the brand of stud you are using.  If you have to use a press to put in studs, that is too tight.  You should be able to hold the hub in your hand and drive the stud in with a hammer, and hear the tone change when it seats.  Drive the stock studs out the same way, and see how tight they are.  2 hits or maybe 3. 

 

We all need to realize the parts supply chain when talking about this.  There is no consistency.  We are racing mass produced parts for an econobox. Price point is not a relation to quality.  Steel supply price variance, shipping industry price flux, and quarterly growth or not have more bearing (ha) on the bearing quality IMO.

 

I have had expensive hubs fail almost immediately.  We tried expensive hubs for the 25hour 2 years ago, and went thru 4 of them.  The last month or so I have done 6 hubs on 2 cars, including 3 rears.  All have the best grease made.  Texas 8 hr enduro is a rough place.

 

I have opened boxes on an order of a dozen hubs, and got the same hub out of two different brands with 100% price variance.  I have gotten hubs from the same brand and had one made in china and one made in japan.  I now seek out a "look" of the induction hardening of the outer races that seems to correlate with long life.  Hardly scientific I know. 

 

Last year I went back to cheap hubs for the 25, and we won the race in E2, no issues.  Hubs were on the car, no issues thru last weeks majors (it ran TWS 8 hour), until a renter spun the car over curbs and pot holes.  2 sessions later they were audible.

 

If your car is pushy, then everything on the track you encounter is harder on the hubs.

The lower you run your car, the more shock the hubs get IMO and the faster they fail.

I torque front hubs to 165.  Rear hubs are trickier.  Replace the drive flange more often then you think, and torque the cv correct.  Rear hubs do take a good feel with the press, and the press fit of the inner drive flange is critical.

Hubs that get "fast loose"  (A tick of play) and stay that way, often last the season.

Re-torquing the nuts are an indicator you are losing one.  If the nut moves, it is a matter of a few sessions.

If you want to set up a loose bearing, do it with shims between races, not loose torque.   But, this is a red herring IMO.

 

Having said all that I do think a solution to the problem would be very doable.  Talking about 25K research and patents is not helpful.  We have a single dimension spindle that with a little research, probably already has a mass produced hub with a larger bearing off some heavier car like a nissan or toyota or kia.  Worst case I imagine is making a spindle collar to a slightly larger bearing ID. 

 

But, everyone, allowing a hat and rotor combo to get back to the stock mounted caliper will be a neccessity here I would imagine.

 

Since STL rules have the lattitude to do this, that is my plan if and when I ever get enough time to make progress on that car. 

 

Well thats my alloted forum time for awhile.

 

My .02

Kyle

[AW33COM]

In Ft-lbs

 

Front: 160

Rear:  185  (depending on production year, bearing type?, don't know)

 

As long as you don't mess with your bearings, and everything else is normal, you will run thousands of miles on a race track with those bearings.  Anything lower by 10/20ft-lbs is unreliable, as one curb hit to a bearing with play can be lethal.  Just remove the brake caliper, torque the hub to 140, 160, and 180.  Spin it at each value and you'll see for yourself what happens with your hub. 

 

Just to put something into perspective, Tony Ave recommends to repack Trans Am (880 HP) bearings every 750 miles. 

 

** One more thing, people often told me the "click, click" tool is rarely accurate.  Hope it helps. 

[Bench Racer]

  The rx7 failures in EP were the steel hats failing, and the replacement kit used the original bearing setup.

 

Kyle

You make many great points on the Miata front hubs. Per the RX7 front hubs, the 79/83 had front hats which sliped over the wheel studs. The 84/85 GSL had front brake rotors cast to the inside length of the hub and the wheel flange was cast to the outer length of the hub. It was the outer cast wheel flange that broke free of the hub (tire/wheel went for a ride without the car) and if one was lucky it didn't tear out the brake rotor and caliper. From memory and confirmed by ksraceware Charlie Clark.

David

[davew]

Trying to compare the hub off of an RX7 or a Trans-Am car to a Miata is like comparing an Oak Tree or a space shuttle to an apple. It is all info, just does not pertain to what we are doing.

 

I can visualise in my head how to do this. The hub would have to fit on an unmodified spindle, with a correct bearing inside dimension. then would need to be properly spaced with high tolerance spacers to get an accurate preload (like a TA car) and still work with stock rotors (no need to drill the stud holes) and calipers. Probably cost $300-$500 per front wheel. But it could be done.

 

Dave

[steveracer]

Not trying to negative, did you miss the post that Mazda's new suoer duper hubs failed out east this past weekend? I have not had a hub failure either. Maybe I have the most supreme hubs made or maybe I don't put an equal load on my hubs that failed hub drivers do. What about the hubs that have had the hub flanges fracture.

 

I did see that, and from here I'm inclined to attribute it to press operator error or too much interference fit with the shoulder. Strictly my WAG opinion and worth what you paid for it. Also, as stated earlier, that's a seperate issue from bearing failures. The failed flange I saw spun smooth as silk with no play or noise.

 

Yes, I wiggle the hubs to check for play every time the car is off the ground, but only repack during the winter break. Even then, they look fine but I figure since I'm there...hopefully I didn't just jinx myself.

[Bench Racer]

Guys, if I was a genius, I wouldn't post..... In my KISS principle view, the E production car hub/flange failure and the Spec Miata hub flange/bearing ball/outer race failure are both because of load. Two options, decrease the many different loads or beef up the hubs and bearings to support the many different loads which will find the next weakest link. Visualizing picking apples from an oak tree becomes real when this hub/bearing issue becomes a safety issue which was the reason for the E production car redesign. The E production basket cost is $400.00 per stem and they no longer loose wheels.

[davew]

Dew, compare the number of broken hubs (not loose or noisey bearings) per mile on  SM's versus the number of broken hubs found in EP Rx7's. I have heard of 2 hub failures in SM in the last 10 years. I don't think that is an epidemic.

 

FWIW; I have tried every brand of bearings, Timken, NTN, SKF, Mazda, white box ebay, and am now using Raybestos. My conclusion is that we are going far beyond the design criteria for these bearings. We need to expect them to fail, and be happy when the do not. I am using Raybestos because I can consistently get better life out of them, than anything else I have tried. BUT, they are not perfect. Some last a long time and some don't. If I had the perfect bearing, I would sell each of you 2 and retire!!!

 

Lets face it, the guy who does 3 races a year at the back of the pack can get much better life from almost every part on the car, than the Steyns, Dragos and Voyteks of the world.

 

dave

[DionJ]

I can buy trick hubs for my 40 year old Datsun 510 for $450 a pair. We as a class can award one single supplier the monopoly to defray design costs (just like Fat Cat), Mazda would be the obvious choice.. As far as chasing the failure, yes, something else will eventually be the weak link, but not nearly at the rate of hub failures. Few of us want to slow the cars down now that we're here. Mixed class racing is frustrating enough, I can't imagine driving a B-spec (sorry to those who do).  510 Hubs.jpg   113.86KB   1 downloads

 

[Bench Racer]

 

 

 

I have heard of 2 hub failures in SM in the last 10 years. I don't think that is an epidemic.

 

 

 

Lets face it, the guy who does 3 races a year at the back of the pack can get much better life from almost every part on the car, than the Steyns, Dragos and Voyteks of the world.

 

dave

 

 I have not had a hub failure either. Maybe I have the most supreme hubs made or maybe I don't put an equal load on my hubs that failed hub drivers do.

It would be a waste of my time back tracking for the number of SM hub flange failures posted in the past three months, more than two, maybe you didn't hear or read about them. 

Maybe had you read this post from 04 Jun 2014 you would have kept your sarcasm within your vision.

Sarcasm mode off.

[davew]

You would think that in the past 13 years of racing, building, prepping Spec Miatas In literally thousands of races, that I would have seen a hub flange failure. I have NEVER had one on either front or rear.

 

If I am that lucky I need to start playing the lottery.

 

dave

[Brandon]

To be clear and fair to all commentators on this thread, there have been two distinct types of failures identified in the last 6-8 weeks which had previously not been regularly observed or noted.

 

1. The new Mazdaspeed hub (their 'optimized ball' item) experiencing a perceived 'faster than normal failure rate' - a generalized failure of it as opposed to #2's catastrophic one.  

1a. Not previously observed or noted due to it being a new item early 2014.

2. The catastrophic flange failures of (provenance uncertain) hubs causing lost wheels & other ancillary damage to that corner of the car.

2a. The first I had heard of this happening was earlier this year on this forum.

 

My personal equipment failure (no comments from the peanut gallery) was #1.  Speculation has run rampant on both failures but until I remove & outwardly inspect my own (I will not disassemble it as I've been told to send it directly to the 3rd party) I will not add to the speculation.  I plan on documenting as many data points from my experience leading up to the failure for everyone here and more importantly the 3rd party to review.