How to tell which wheel bearing?

Ok, the first wheel bearing I changed on this car, 50,000 miles ago, was the right rear, and by the time I changed it there was no problem telling which one was bad; it was noisy and felt notchy just turning the hub by hand.

I can tell by the sound (driving on pavement) I've got another one on the way out, but so far its not bad enough to pin it down by feel. (I've jacked up all four corners and spun/pushed/pulled/shook by hand, no difference I can feel.) It is getting gradually louder, and I'm sure I'll eventually be able to tell....

Here's what I know so far:

• sounds like a small propeller-driven Cessna or Piper, only not as loud
• louder when taking right curves or cutting wheel to the right
• softer when taking left curves or cutting wheel to the left
• hub temps after several minutes on highway: front 50.3 and 50.8 C, RR 39.5 C, LR 43.3 C

Anybody think you can call it from the info I've already got, or have another test I should try?

Thanks,
-Chap

 

Well, I finally knuckled down and repeated the human frequency counter trick from my earlier story. It just required finding a sufficiently cop-free stretch of road at midnight where I could drive at various wrong speeds swerving constantly left and right.

As before, with the standard tire sizes, tweaking a standard tuner from a=440 down to roughly a=433 (down 28 cents, if that's how your tuner's marked) seems to be a good futz factor to make tuner pitches match by ear with Gen 1 hub noises at "nice" speeds in US MPH (30, 40, 45, 60, ...).

In this case, the pitch at 30 MPH is g (second below middle c), with 40 and 45 naturally being c and d (below middle c). 60 of course makes g again, below middle c, but in that octave it's harder to pick out of road noise.

The car's now wearing Defenders (they retired the Harmony line Defenders are louder) rated 905 revs per mile, so at 40 MPH they're doing 603 rev/min. The c below middle c (a=433) is about 129 Hz, or 7740/min, so this noise frequency is about 12.8 × wheel speed. Back when my right rear went, the factor was 10.8.

I wonder whether a difference like that would mean it has to be a different size/configuration of bearing, or it could be accounted for even in an identical bearing depending on whether the inner or outer race has the chief damage.

Hmm....

-Chap

 

Well, I don't know - I honestly don't know. My line of work hasn't ever involved learning about bearing noises, or even learning what bodies of literature I'd even look in to learn about bearing noises. So I'm just experimenting by collecting the information about the noise, and at whatever time I finally manage to confirm where it's coming from, maybe I'll see in hindsight whether anything about the noise could have told me something. Maybe knowing that either way could save somebody else some time.

I have been thinking about bearings though. You've got an inner race (say radius R), a ball/roller cage (with n rollers of radius r), and an outer race (R+2r). In our application we've fixed the outer race at 0 RPM and the inner race is doing the wheel RPM. The cage is loafing along in the same direction as the inner race but a lower RPM, which I think might be ½(wheel RPM)R/(R+r). Don't quote me yet. My scribbles are hard to read and were done late at night, I should recheck them. But it's gotta be something like that.

Anyway, if there's an outer race defect the n balls are going to smack it with frequency n(cage revs/sec), and if there's an inner race defect, that'll be n(wheel revs/sec - cage revs/sec). Of course there could be multiple defects, and the sound I think I hear could be any harmonic of that, so there might be so many possible variations on the sound that it won't have any useful interpretation. But sometimes real life turns out to be more cooperative than it strictly has to be.

At least this picture satisfies me about why the sound frequency (in both my cases so far) has been a non-integer multiple of wheel speed (I kind of expected the ball count to be an integer, but the radius dependence of the cage speed solves the mystery). I can imagine how the sound might be useful to distinguish front from rear if they are of different radii [edit: or have different numbers of balls]. (Unfortunately, I can also imagine how it could be different even for identical bearings depending on which race is gronkled.)

If my problem turns out to be in front, maybe I'll be able to report the sound is possibly useful (two observations, one rear at 10.8x, one front at 12.8x, make of that what you will). If it turns out to be rear again, then maybe not as useful (but we will have learned that both 12.8x and 10.8x are possible sounds from the rear).

Good thought. I haven't, and it wouldn't be that hard. Over the last month I've already had wheels off three times trying to see if I can feel the damage yet (still can't); might as well put 'em back on different corners and see what happens.

I'm not putting my money on that though; I really do think a tire defect would sound at wheel speed or a harmonic of that, not some non-integer multiple.

I hear you. My last bearing was going during a crazy time when I was making weekly 600-mile road trips and didn't even have time to try anything until several trips had gone by; nothing bad happened (though I could clearly feel the damage by the time I was able to get to it).

Based on how I understand the construction, I think it would be very difficult for a front hub to leave the car, as long as the axle nut is secure; I think the stub shaft has a flange too wide to go through the knuckle. Things would certainly have to get spectacular before anything could part company.

As far as I know, that's not quite as true of the rear; let the balls and races wear down enough to clear each other and there's not anything left holding the hub on. I'm sure I'll be able to catch this well before anything like that happens, but one does hear of others experiencing it [edit: general stories, I don't mean Prius specifically].

Right now I'm not making regular long trips as I was then. I could still be called on one unexpectedly, though, and I'd prefer to have a new bearing first. Once I know which to order, I'd prefer to use one of the online dealerships for the discount, at least if it's the rear. (The front bearings are cheap, but the job's involved. The rear's a straightforward job but the hub/bearing assembly's a chunk of change that I'd rather not spend until I know that's the culprit.)

I've ordered a multi-microphone listener thingy, should have it by next week. They cost too much but there was exactly one for sale used the other day so I snagged it. We'll see how much I can learn from that. Who knows, it might be back on the used market soon....

-Chap

 

For this car, anyway. If you're in, say, (what appears to be) some rear-wheel-drive SUV, don't count on that:

The video is spent more on closeups than big picture, and the geometries aren't obvious at first: it's the right front, leaving the vehicle pointed opposite the direction of the pavement marks, so I infer it must have at least 180'd in travel, though maybe there's another interpretation I'm not seeing.

-Chap

 

Ssh! Hey, aren't you the same one who said

? Toys for geeks, pal.

-Chap

 

BTDT. It's not reached the point where I can hear anything that way. Part of the problem is having the bearing unloaded. If I could, say, spin it while exerting a quarter ton of pressure, we might get somewhere. (I've even thought about jacking up a wheel, then setting it down on some kind of roller contraption to carry the weight and spinning it that way, but I'm betting the 4-channel listener will do the trick with less trouble.)

-Chap

 

Yes and yes. The pickup elements are built into what look like cheap battery clamps so you can clamp them right onto something for direct mechanical sound conduction. I clamped them onto one of the caliper bolt heads on each front, and one of the hub-assembly-to-bracket bolts on each rear.

The pickups come with maybe half a meter of wire that you can route to the wireless transmitter; there's one for each pickup, about the size of a deck of cards. Each transmitter contains magnets (would hold to steel parts, but didn't seem strong enough to stay stuck over rough road), a spring-loaded thing like a belt-buckle clip, and a slot for velcro straps (for the win!). So, yes it's wireless, and yes, you still get to intricately route wires anyway.

I strapped each transmitter to some part fixed to body rather than suspension, so they wouldn't get bounced as much. I routed the wires from pickups to transmitters alongside each wheel speed sensor cable, which is already cleverly routed to flex with the suspension and not get in anything's way. The setup comes with a bunch of velcro straps and a bunch of ordinary one-time-use nylon wire ties, but instead I ran the wire along the speed sensor harness with a bunch of these thingies, which I already had a bag of and find much quicker to use.

The sound is very clear. You can listen to the sounds of brake linings against rotors, and hear exactly when the ECU's switch from regen to exclusively brake pressure just before a stop. You can hear shoes expand and retract. You can hear some bearing sounds, very faint on most of them. Not faint on my left front. No, sirree, not faint at all on the left front. I'm pretty sure I know where my problem bearing is.

Interestingly, I get really none at all of the low, Cessna-like sound that's the only symptom naked ears hear inside the cabin. The pickup works in a whole different frequency range where the sound is more like dragging an anvil down the street. Quite horrible sounding. If I plugged the receiver into the car's aux in and turned the bass up and everything else down, I thought maybe I was picking up some of the lower sound, but only barely. The pickup presents a whole different aspect of the bearing sound.

Most surprisingly of all, I get no expected correlation with IR temperature readings at the hub center (negative correlation, even). I popped the hub buttons out of the wheels so I could shoot straight at the axle shafts in front and hub centers in back, thinking that's probably where bearing heating would best show up. The left front, the bearing that sounds the worst by a country mile, in fact reads 2 to 3 K cooler than the right front, which sounds perfectly quiet and smooth.

In the morning I might go swap a couple of the pickups and transmitters just to rule out anything funny like transmitter #1 somehow having way higher gain than the others. But barring anything weird like that, I'm pretty sure I've got my culprit.

-Chap

 

Swapped and nope, nothing funny. LF bearing is definitely shot. Parts on order.

For those keeping track, so far I've lost one rear bearing (which rumbled at a pitch 10.8× wheel speed) and one front (12.8×) ... make of that what you will.

-Chap

 

Well, given that nothing else I had tried gave me a conclusive ID on which bearing was going, and this thing made it obvious about 3 seconds after being turned on, I'm gonna say it's pretty useful. (Perhaps the 12.8/10.8 pitch business will prove to be useful for distinguishing front from rear, if it turns out to hold for anybody else, but that still won't pin down one bearing.)

Now, cobbling together a few microphone elements and some spring clamps and velcro wouldn't be that hard for somebody with a bit of time or some parts lying around. A small amp and speaker, or just a plug to connect to the car's aux in and play through the car amp. This setup just saved me the time of cobbling anything together, and offered for a used price, that worked for me.

One of the applications mentioned in the instructions is to clamp the pickups onto fuel injectors. Apparently they distinctly go tinktinktinktinktink if they're clean, and thumpthumpthumpthumpthump if they need cleaning.

-Chap

 

I didn't, sorry. In front I clipped the pickups onto the big ol' heads of the upper caliper-to-knuckle mounting bolts, and routed the wires exactly along the wheel speed sensor wires with three or four clips right over to the inner fender wall, and hung the transmitters on velcro straps fed between the plastic engine undercover and the body at the rear bolt (which is pretty high up on the fender wall).

On the rears I clipped the pickups to one of the four hub/bearing assembly mounting bolts, again feeding the wires right along the wheel sensor wires. On reaching the body, the sensor harnesses go through a rigid guide that's held by a couple screws, and I hung the transmitters by velcro from those guides.

It was just a matter of looking closely to see where stuff wouldn't get in the way (remembering my Dad's oft-repeated "never get between something that moves and something that doesn't"). I spent a long time placing the stuff, removing each wheel in turn. It turned out when I was done that everything had ended up where I could reach it with the wheels on the ground, so it was much faster to take off (or put on again, should I ever need to).

-Chap