274 replies to this topic

[Brandon]

There's another variable I'm being at least somewhat cognizant of: I'm going faster than I ever have before in my car.

From my perspective, this may very well be a resultant 'feature' of being fast in SM - in addition to now tires & pads being wear items (again, I could go nearly a full season on XP10s, now I may be lucky to get 2-3 weekends), you also need to start looking at hubs & ball joints as well.  Things I'd never before thought necessary to check after each session up until this year.  

 

Additionally this influences what I'll bring to the track for any weekend but if that's a necessity, I'll shift my procedures to accommodate what the car needs and go from there.

 

One last item: admittedly we're not hearing from 100% of the front-runners in the class but we are, however, hearing from a few who not only run near the front but also support others that are experiencing a similar (albeit not 100% exact) situation.

 

To repeat: I installed a NIB (new in box) MSP enduro hub (with zero alterations) on the front axle, torqued to 140 ft/# (within FSM specs), sealed the cap with silicone, and the LF hub did not last a test day plus double-race weekend at NJMP-Thunderbolt.  

 

Until I hear from MSP or ProParts about what failed on mine and/or a torque spec to tighten future installations to, I'll continue to run an OEM set that I repack myself and torque to the same value as before.

[Jamz14]

Point A. Would you like a front hub with  a two row ball cartridge bearing or tapered roller bearings? A mature enterprising person.

 

Point B. A few years ago there was similar issues with the front hubs on the Mazda RX7 in E production, redesinged hub, issues solved.

 

Point C. Move from the Chick car class up the ladder to a class that provides your feel goods, Spec Miata can't be everything for everyone.

No dissrespect towards Kyle, unless he has insider info, don't hold your breath.

Point A: I want a hub that can withstand the loads and is easily serviced and don't really care if you do that with two row ball cartridge or berries and pixie dust. Whether it is done by an old enterprising person or young.

 

Point B: Sounds good. Write a letter to those with influence and the affliction, and I think you will have your problem solved.

 

Point C: I am not suffering from a tattered ego and enjoy working on and fielding chick cars that talented drivers use to show that they are not.

 

And lastly, I am not holding my breath........or experiencing the problem. Doesn't change the fact that Kyle is right. I have total faith that when those with influence experience a problem that they set their minds to a fixing, it will be so without my letter being needed.

[Caveman-kwebb99]

I have no insdier info, I just know that one particular guy will not be wanting to put a new hub on everytime he turns around, and after seeing all the hubs that had studs knocked out and hammered back in this weekend, if that keeps up imo that could motivate change as its a pita and there is cost involved.

[Keith Novak]

One item I haven't seen mentioned about a potential failure mode also has to do with the inner races.  (It may have been mentioned but these threads get long.) When repacking the hub, there is a gap between the outboard and inboard inner races where you can stick the washers to knock out the race.  I was rather curious while rebuilding a non Mazda direct set whether that gap goes to zero when the hub is torqued down.  It's rather difficult to see with the hub installed.  If it does bottom out, tolerances in the race dimensions could prevent the torque from applying the proper pressure to the bearings.  Loose balls are of course more prone to damage in more ways than one.

 

The evidence of damage could be seen in the outer races like Saul has seen, but caused by issues with the inner race, plus the hub flexure itself that generates higher stresses in the outer race.

[Brandon]

One item I haven't seen mentioned about a potential failure mode also has to do with the inner races.  (It may have been mentioned but these threads get long.) When repacking the hub, there is a gap between the outboard and inboard inner races where you can stick the washers to knock out the race.  I was rather curious while rebuilding a non Mazda direct set whether that gap goes to zero when the hub is torqued down.  It's rather difficult to see with the hub installed.  If it does bottom out, tolerances in the race dimensions could prevent the torque from applying the proper pressure to the bearings.  Loose balls are of course more prone to damage in more ways than one.

 

The evidence of damage could be seen in the outer races like Saul has seen, but caused by issues with the inner race, plus the hub flexure itself that generates higher stresses in the outer race.

 

This was one item I observed after removing the failed MSP hub from my own car - the gap between the inner races was, to my uncalibrated eyes, significantly narrower than previously seen in non-MSP hubs.  It's not clear whether this is by design or indicative of a need for more torque (perhaps to make both inner races "solid" on the spindle) to be able to take advantage of the 'more round' balls....

[Bench Racer]

One item I haven't seen mentioned about a potential failure mode also has to do with the inner races.  (It may have been mentioned but these threads get long.) When repacking the hub, there is a gap between the outboard and inboard inner races where you can stick the washers to knock out the race.  I was rather curious while rebuilding a non Mazda direct set whether that gap goes to zero when the hub is torqued down.  It's rather difficult to see with the hub installed.  If it does bottom out, tolerances in the race dimensions could prevent the torque from applying the proper pressure to the bearings.  Loose balls are of course more prone to damage in more ways than one.

 

The evidence of damage could be seen in the outer races like Saul has seen, but caused by issues with the inner race, plus the hub flexure itself that generates higher stresses in the outer race.

Ok, went to my stock pile of OEM front hubs to check the bearing part mfg., number and washer fit between inner races. The inner race has this info, Japan YL, NTN, HUB055-6, a .060 inch washer fits between the assembled inner races. Several times I've used the .060 washers between the inner races of different hubs. This is a hub that was un-torqued a few moments ago. IMHJ .060 plus would be one hell of a gap to close when torqued at 100 and whatever ft lb. One could put some of that measuring squishy bearing plastic material in place of the washers and torque the deal, take apart and carefully remove the plastic and measure.

 

 Only the info messenger

 

EDIT:

One could torque to 10 foot pounds, measure axle nut location, torque to 100 and whatever foot pounds and measure axle nut location.

[Michael Novak]

I think it is simple.  Up until late last year my failure rate was quite low. I think the repacking for some people could be part of the problem, but we have some failing within a hour or so.  The outside hub is not being heat treated the same as it had been... The greases are not the main problem.....  I think I have now failed one of every decent brand out there..   OLD hubs that are repacked are working fine.....

 

 

I know zero about heat treating?   Can you re heat treat a part?   Cyro?   

[FTodaro]

If i understand the question about the gap between the inner races when the hub is tight on the spindle, I am pretty sure that there is no gap between the two races when they are seated together when you tq it on the spindle.
If I am following.

[Parity]

 

 

I know zero about heat treating?   Can you re heat treat a part?   Cyro?   

Not without deforming it. If it's induction treated then it can't be done at all. 

[Keith Novak]

If i understand the question about the gap between the inner races when the hub is tight on the spindle, I am gritty sure that there is no gap between the two races when they are seated together when you tq it on the spindle.

If I am following.

That is what I was getting at.  If that's true, very small differences in the total length of those inside races can be an issue.  Torquing the hub may not provide adequate preload on the bearings. 

 

I just did one where after rebuilding the hub, I found a little bit of play in the bearings both at the low and the high end of the torque range. The play did not change.  The play is uniform top to bottom and side to side.  This made me think it wasn't a pit in a specific part of the race but rather the bearings weren't fitting tight all the way around the races.

[Keith Novak]

Ok, went to my stock pile of OEM front hubs to check the bearing part mfg., number and washer fit between inner races. The inner race has this info, Japan YL, NTN, HUB055-6, a .060 inch washer fits between the assembled inner races. Several times I've used the .060 washers between the inner races of different hubs. This is a hub that was un-torqued a few moments ago. IMHJ .060 plus would be one hell of a gap to close when torqued at 100 and whatever ft lb. One could put some of that measuring squishy bearing plastic material in place of the washers and torque the deal, take apart and carefully remove the plastic and measure.

When they're on the bench I agree but there is play in the races at that point.  If you put the assembled hub in a clamp pressing the outer races together, can you still fit the washer in the gap?

[FTodaro]

That is what I was getting at.  If that's true, very small differences in the total length of those inside races can be an issue.  Torquing the hub may not provide adequate preload on the bearings. 
 
I just did one where after rebuilding the hub, I found a little bit of play in the bearings both at the low and the high end of the torque range. The play did not change.  The play is uniform top to bottom and side to side.  This made me think it wasn't a pit in a specific part of the race but rather the bearings weren't fitting tight all the way around the races.

That could be it or the race tolerance is is to large or that it is expanding with use due ti shortcuts in the mfg process.

[Johnny D]

You can make jokes if you want after I say this.

 

You may have a problem with big balls and small (Dim) races crushing.

 

Or the other way around with slop.

 

Then the TQ you use may help or hurt.

J~

[Keith Novak]

I looked at a hub this evening which has not failed

 

Ok, went to my stock pile of OEM front hubs to check the bearing part mfg., number and washer fit between inner races. The inner race has this info, Japan YL, NTN, HUB055-6, a .060 inch washer fits between the assembled inner races. Several times I've used the .060 washers between the inner races of different hubs. This is a hub that was un-torqued a few moments ago. IMHJ .060 plus would be one hell of a gap to close when torqued at 100 and whatever ft lb. One could put some of that measuring squishy bearing plastic material in place of the washers and torque the deal, take apart and carefully remove the plastic and measure.

 

 Only the info messenger

 

EDIT:

One could torque to 10 foot pounds, measure axle nut location, torque to 100 and whatever foot pounds and measure axle nut location.

I found I have a hub with the same p/n races and the the gap when you press the races together is smaller than the thickness of a piece of plastic from a bag of sunflower seeds.  My calipers are MIA during some garage remodeling but I'd say the gap is 6-10 mil.  There is a tolerance stack of the cast hub, races, and balls.  The races and balls should be of higher tolerance than the cast hub.  If the races on the cast hub are off a bit, contact would not be that difficult and you lose the preload from torquing the hub down.

[Bench Racer]

I looked at a hub this evening which has not failed

 

I found I have a hub with the same p/n races and the the gap when you press the races together is smaller than the thickness of a piece of plastic from a bag of sunflower seeds.  My calipers are MIA during some garage remodeling but I'd say the gap is 6-10 mil.  There is a tolerance stack of the cast hub, races, and balls.  The races and balls should be of higher tolerance than the cast hub.  If the races on the cast hub are off a bit, contact would not be that difficult and you lose the preload from torquing the hub down.

Very interesting ^. Along with others thoughts/guess of inner races touching or zero gap.

My hub is a non failed OEM/junkyard/unique rectangular punch mark to set hub to axle slot.

I did some measuring of how far the axle nut moves when applying torque/hand rotating between torques from finger tight being the base line in 20 foot pound increments through 140 foot pounds with the last torque being 10 foot pounds to get to 150 foot pounds.

The nut moved .0030 inch with respect to the outside machined end of the hub. Could explain my measuring process if need be. Lets say the nut moved .0060 inches, twice as much as I measured the ends of the inner races will not touch per my one measuring of one hub. The .060 inch (Duh, as Keith posted the race gap opens for the .060 washer to slide between.) washer fit between the bearing inner races with some clearance. Tomorow, I'm going to stick some bearing clearance measuring plastic or something between the inner races and see how much gap remains after 150 foot pounds of torque. No question part tolerances play a part relative to the inner race gap. Being tolerances are involved in these cheap a$$ hubs, I would dought there was any design parameters set forth that when the specified torque was met the races would be touching each other. If the tolerances were such that at the specified torque the race gap was zero, they would not be cheap a$$ hubs. 

 

Saul, if you've been through this inner race gap measuring please post your dimensional results/findings.

EDIT:

This morning, measured from end of the axle to the outside end of the hub, .0040. On this hub/bearing assembly the inner race ends do not meet/touch.

EDIT:

It took a while for Keith's several thousands of clearance info to sink in. Just did my last measuring of the OEM NTN hub, by inserting some material between the inner races, tourqued to 150 ft lb, took apart, measured and the clearance between the inner races of the one hub I measured is .005/.006 inches.

[SaulSpeedwell]

That is a good question because what is on the box does not always tell the truth about where the hub came from, or who manufactured it. An NTN hub could be NtN parts assembled by a Jobber.

 

I want to find some OEM hubs and repack that is my answer.

 

I'm trying to help with this by gathering data, but it isn't easy because, as you point out, it is a moving target of what is in the box, and not all "NTNs", etc., being created equal. 

 

I don't know what I would buy tomorrow.  I have found ONE continued OEM usage (in Japan, or at least not in the U.S.) of this bearing, but it is one of the ABS versions with a tone ring.  I don't have the part number in front of me, but ordering that part from Mazda and "seeing what is in the box" might be interesting. 

 

There is also a similar bearing that is "one size larger" and meets all the specs (I think?), except for two:  it has a 30mm ID instead of 28mm, and it would space the rotor outboard 2mm.  It IS a higher load-rated bearing, and it is still used OEM.  If we could find a reasonable and reliable way to sleeve the spindle up to 30mm, and determine that the 2mm rotor issue is negligible or easy compensated for, this could be an option as opposed to a "full custom" setup.

[Todd Green]

Finally remembered to take closer pics while I was at the track.  (Images are much larger than what the forum resizes them to.  Just click on them.)

 

[Andy Mitchell]

If I remember the stuff they taught me in engineering school, that's a fatigue failure. Giveaway is the scalloping on the left side in the second pic. 

[FTodaro]

I find myself looking behind the rotors on tire changes these days. Would like to know if anyone has had a failure like this with stock studs?

[Brandon]

Very interesting ^. Along with others thoughts/guess of inner races touching or zero gap.

My hub is a non failed OEM/junkyard/unique rectangular punch mark to set hub to axle slot.

I did some measuring of how far the axle nut moves when applying torque/hand rotating between torques from finger tight being the base line in 20 foot pound increments through 140 foot pounds with the last torque being 10 foot pounds to get to 150 foot pounds.

The nut moved .0030 inch with respect to the outside machined end of the hub. Could explain my measuring process if need be. Lets say the nut moved .0060 inches, twice as much as I measured the ends of the inner races will not touch per my one measuring of one hub. The .060 inch (Duh, as Keith posted the race gap opens for the .060 washer to slide between.) washer fit between the bearing inner races with some clearance. Tomorow, I'm going to stick some bearing clearance measuring plastic or something between the inner races and see how much gap remains after 150 foot pounds of torque. No question part tolerances play a part relative to the inner race gap. Being tolerances are involved in these cheap a$$ hubs, I would dought there was any design parameters set forth that when the specified torque was met the races would be touching each other. If the tolerances were such that at the specified torque the race gap was zero, they would not be cheap a$$ hubs. 

 

Saul, if you've been through this inner race gap measuring please post your dimensional results/findings.

EDIT:

This morning, measured from end of the axle to the outside end of the hub, .0040. On this hub/bearing assembly the inner race ends do not meet/touch.

EDIT:

It took a while for Keith's several thousands of clearance info to sink in. Just did my last measuring of the OEM NTN hub, by inserting some material between the inner races, tourqued to 150 ft lb, took apart, measured and the clearance between the inner races of the one hub I measured is .005/.006 inches.

Bench,

Your edits seem to imply there might be something between the size/sphericity of the balls and any possible pre-load applied by the torqued value.  If they've specified an even slightly smaller ball (regardless of relative "roundness") the loss of pre-load would spell an immediate end to the hub should the faces touch (even if they didn't rotate/move against each other).

 

As I said before, I'm still waiting to receive back my failed hub from MSP (calling again tomorrow) but I will definitely check the face of the inner races to determine if there was any contact between them.

 

Whee!

Brandon