speedengineer

Yes...Front hub failure, not bearing failure. Just wanted to share...I know lots of people have had good luck with the Timken hubs, but I'll be going back to OEM for my next set.

Happened at Autobahn country club this last weekend, coming through T3 the car suddenly felt weird, like the wheel was loose and about to fall off, so I pulled off track in a safe spot and got towed in, lots of clunking and brake dragging the whole way. Back in the paddock I jacked the car up, had what felt like lots of play in the wheel bearing. I went to swap it out and when I pulled the rotor off got a surprise - the hub was cracked nearly all the way around! I'm sure the only thing that kept it from coming completely off was the brake rotor, a couple more corners and I might have lost the whole wheel...

This hub assembly has only been on the car since April (8-ish weekends maybe, running 205 Hoosiers at around 2400lbs comp weight). I had replaced wheel bearings on all 4 corners for preventative maintenance...lot of good that did. Looking at the back side, there is clearly a stress concentration/corner where the flats were machined for the wheel studs, just like others have noticed on some rear hubs that have failed. Another set of front hubs I own do not have this corner. Additionally this Timken hub was significantly thinner.

The parts was from RockAuto, Timken part number 513155. I will say that the Timken photo on rockauto's website does not have these machined flats. If yours do...might want to consider a change.

midpack

iTrader: (3)

Happens with Mazdaspeed hubs too
(not my car)
Glad you caught it before complete failure

speedengineer

Interesting! Thanks for the link. I hadn't heard of this happening before, but apparently there have been other failures. All the more reason to change these things frequently I suppose. :(

99Racer

I note you have Long Studs in those hubs... You may want to search "stud hub failure"… Or “When was the last time you changed your rear hubs?”

It appears some replacement hubs have stud holes that are smaller than ideal for the press in long studs... So, pressing in the long stud expands the flange of the hub which results in a fracture splitting the flange from the center. Its possible the bearing & flange were fine prior to pressing in the long studs… No manufacturing defect, just a parts mismatch. Studs vs stud hole size in hub.

Gee Emm

That is part of the answer, maybe. I can see that a hole/stud size mismatch will result in higher stress in the hub, but those cracks are nowhere near radiating from the holes. That suggests a different cause of this failure.

Looks to me like the hub was supported close to the bearing/axle, and the pressure from the press bent the hub flange (ie where the stud holes are) away from the centre body of the hub. This would create a bending effect, and those cracks are in the right place. This might be caused by a press inserting matched studs/holes, but if it had to be cranked up to force in studs appreciably bigger than the hole was made for, then the likelihood of damage would be greater.

Just my 2c ...

mx5-kiwi

you need to read the links supplied above...

Gee Emm

You're right. I read the first page, didn't see the rest. My bad.

Seefo

iTrader: (5)

Kinda concerning for me. Was planning to run extended studs in a new hub since I am still on the original OEM, but from reading that thread it seems the only good hub (ie hub without failures) is the OEM miata hub. Also planning to run 15x9 with a 5 or 10mm spacer, so extended studs are recommended here...

The OEM hub is different from the one offered by mazdaspeed now (this one is apparently a "chinese" hub). Anyone care to comment? I know a few of the guys here running turbo miatas with more grip than 205 RRs or SM7s. Does anyone replace the OEM hubs or just repack and replace bearings?

Is it safer to hand install the ARP studs (ie clamp method or a hand arbor, not stud method) or no difference between that and hydraulic press?

NiklasFalk

iTrader: (1)

Have anyone measured the studs in the Chinese hubs to see if there is any ARP that are identical?
Most probably, but I'm incapable to find a discussion fast.

They might be cheaper than the Miata ARP, just for kicks.
They have most probably picked a larger volume M12x1.5 stud. Using something custom for all M12x1.5 can also be an option, but sounds unlikely to me.

Seefo

iTrader: (5)

the forum discussion linked above has a measurement comparison of the ARP studs vs. the Mazdaspeed stud holes vs. the Stock studs (around pages 5-8). There is a difference, but it seems to be within ARP's spec after the math was worked out (and Mazda's factory spec. too).

General consensus I can gather from the discussion are hub failures are due to fatigue which implies a definite manufacturing changes between the OEM & OEM-like replacements (ie TImkin, duralast, Mazdaspeed's "chinese", and whatever else). There is one guy who has a reported failure with EVERY hub except the OEM one to boot.

Its hard to sort the truth from the bullshit, but the other issues mentioned (other than the "main" one of manufacturing ****):
1) hub repacking... (ie. we should not repack hubs and just use the factory grease)
2) ARP studs... (ie. ARP stud size is mismatched and putting excess stress on the hub)
3) under-torquing the hubs on install. (ie. too much play between the inner/outer races?)
4) tolerance mismatch due to manufacturing. (ie. the races or bearings are touching?).
5) increased grip + offset changes (I think we can debunk this since no one/very few had failures on OEM hubs...dunno though).

Last 2 are iffy for me. I got a little lost on some of the theories there, but they seem to be "opposites". One thing it seems is bearing faillure is usually indicated by bearing play increasing over a weekend/session, while hub failures tend to be quick and have no sign until a few laps before catastrophe.

Btw, the mazdaspeed hubs I think have an optional ARP stud installed. So it seems to me the hubs are intended for Stock miata/ARP studs. ie its unlikely you will find ARP studs in .020" increments I don't think or whatever minor difference may exist between the OEM hub holes and "chinese" hub holes.

NiklasFalk

iTrader: (1)

With my current hub consumption (1 every five year) it's no problem to stay OEM. I even have CV-2 in there now. Amsoil Dominator on the shelf though so repacking this winter, maybe...

Procrastinator, nah...

OGRacing

iTrader: (1)

Unfortunately hub failures are quite normal in our applications. i try to avoid aggressive rumble strips because of this.

Seefo

iTrader: (5)

Most of the spec miata failures are new hubs and fail within 1-8 (?) race weekends.

It seems to me if your hub is lasting past the first year, its probably going to last a long while. If yours are lasting 5 years, maybe I shouldn't worry about mine then.

NiklasFalk

iTrader: (1)

But my five years is no more than about 50 track hours, and few of them are at 80+F ambient.

shlbygt

The Ford Focus also experiences hub failures on track. On the Focus the front hub is similar to the rear hub on a Miata. On track, the hub face shears from the bearing area without warning. Thankfully the caliper and caliper mounting bracket keeps the wheel from exiting the vehicle.

On track failures occurred early on with original Ford hubs but they were limited to folks who were racing frequently with larger than stock sizes sticky tires. When bearings were replaced, most people also replaced the hubs, even though they looked fine. The reason being, why reuse the hub when new ones are cheap enough?

I ran original hubs without issue for a number of years until it was time to replace the bearings. I went with a Timken kit and pitched the original hubs. The next day I sheared the new hub at Mosport. A friend with another SVT Focus was also there and showed me a new Motorcraft hub he sheared at GingerMan a few weeks earlier.
We did some research and eventually obtained a set of original engineering drawings for the hubs.
It turned out there were material and heat treating spec changes made after 2004. Hub manufacturing also moved from Europe to India.

It would not surprise me that something similar happened with the Mazda hubs. Original hubs are probably fatiguing after years of use and being subjected to stickier tires and more horsepower than the Miata was originally intended. The 1990 model had 185/60-14 tires and 116Hp and now folks often run 225/45-15 Hoosiers with 200+Hp.

The aftermarket hubs are probably being made without regard for the folks tracking their Miatas and are adhering to a normal use specification.

I fixed the Focus issue by having another friend, who runs a machine shop, use the prints to make hubs from 4340 steel with better heat treating. The end result were hubs that were 1.5 to 2 times the strength of the Ford ones. None of the ten or so cars running them on track have had any hub issues in the five years since.

Now the bearings are the weakest link. The inner set of bearings tend to fail on my SVT after 20 to 25 days on track and ~5000 street miles. Folks on the forums think it is a lack of torque killing the bearings.

The only thing is, I check the torque each morning before going on track and it is always on spec. Every time I have found a loose hub nut the inner bearing is toast.

Most likely my bearing issue is heat related as there is no room for brake ducts and the grease always looks nasty on the dead bearings. So now I replace bearing before the first event each season.

Leafy

iTrader: (1)

Hondas have the same problem except it normally breaks the axle and then the hub. I've come down to it being that the factory torque spec is too low for racing applications and it needs to be higher.

I have the same problem on my car with front hubs. I stopped torquing them to 150ftlbs and started torquing them to hurrrrrrrrrrraarararagugh and I havent had one fail since. The only thing you have to fear from over torquing a modern hub bearing that uses ball bearings is exceeding the strength of the spindle or axle and with the size they're at the max recommended torque spec is more than most people are going to be able to produce without a cheater bar. The bearings are designed such that the inner races touch and that holds a constant clearance so long as the shaft the bearing is mounted in is strong enough and torque tight enough that the loads dont pull the races apart in bending.

I've got a whole bunch of nothing with regards to this failure. It looks like a fatigue failure from that stress concentration right there. I know the EF honda guys can no longer get front hubs and they're starting to just have them made and they're not heinously expensive.

shlbygt

On the Focus front hub torque is 233 ft-lbs and the rear 173 ft-lbs.

I came across the Honda issues when researching my hubs. The ones they were making were ~$800 a pair without studs. The first run we made for our own use were ~$700 a pair to make without studs. That is pretty pricey and a tough sell for most folks, especially Focus drivers.

We made a second batch to sell and most folks complained the price was too expensive.


On the SVT Focus hub failure can take out the half shaft (now obsolete from Ford), brake line, caliper, pads and caliper bracket. I was lucky in that I didn't hit anything when mine went but one driver I know hit the wall at Charlotte during a race and had significantly more damage.

I explained to folks that $700 for hubs was actually cheap considering what a hub failure could cost.
My example was what happens when a hub fails on a Sunday at the track.
First you need to source a new hub and bearing, then find a place to press it. That is, if you have a 32mm socket, breaker bar, torque wrench, tie rod and ball joint removal tools and actually know how to remove and replace the knuckle. Then you need a ride to get the parts and the place to press the bearings.


If that doesn't happen on Sunday, you get to spend an extra night at the hotel until the parts/press can be found the next day. You might miss a day of work or if all else fails a tow home will be needed.
It still didn't sway some people that $700 wasn't expensive.

I actually had a rear bearing fail at AMP two years ago and didn't have my spares with me. I had to find a bearing in Dawsonville GA at 5:00pm on Sunday. No one was open to press the bearing and I needed to be in Myrtle Beach for a flight home the next afternoon. Great!!

Lucky for me I had a friend I had not seen in ten years who had a diesel repair shop nearby. So I got to take him away from Sunday dinner to open up so I could be on my way. It is good to have friends you can abuse.

The photos of the Mazda failed front hubs definitely looks like metal fatigue or a weakness in design/material. The sharp edge near the stud should be radiused to eliminate a potential stress riser there.

Since the brake rotor and wheel are all held tightly in the same plane as the flange, the forces on the wheel could cause the hub to flex. Losing torque on the hub nut could also cause the hub to flex at the rear of the flange.

Once one of the areas fail by the stud the rest probably go very quickly.

compaddict

How did you install the studs?

Gee Emm

Coming back to this, after a thought occurred to me. My rear end is apart, and I was looking at my rear brakes, thinking maybe I should pull the hats off and have a look at the hubs.

Has there been any consideration of the role of the hats in the hub failures? My hats are heavily scalloped Wilwood (I think) alloy hats, and to a non-engineers eye they look pretty stiff. But maybe they aren't as stiff as they look?

Any thoughts on this, or any data?

freedomgli

My thoughts are that parallelism, flatness, concentricity and squareness of all wheel end components is critical, from the surface of the hub to the lug nuts and everything in between. I would think that geometric changes in brake rotor flatness and parallelism could induce bending moments, which could result in a fatigue fractures. But I'd think you'd most likely see any problems manifest with a broken wheel stud first. I suppose it all depends on the specifics of the components in question.