uh-oh - it hums....

At about 160,000 miles, my Classic now has a hum. The pitch of the hum is strictly determined by road speed. It is unaffected by engine speed or by shifting into neutral. It *sounds* loudest in the right rear.

From past threads it looks like a hum with exactly that behavior has turned out in the past to be a wheel bearing (ok, no sweat) or a shorted MG2 winding (yipes!).

I'm currently piling on about 600 miles a week helping out my parents for a while so I've already had about 1500 miles of hum without much time to think about it. It is not getting rapidly worse, but it is definitely not going away. Naturally I'm anxious as to whether it's a wheel bearing (or CV joint or other relatively simple problem) or an imminent transaxle failure that I'm looking at. I still don't have my hands on a scangauge capable of reading HVECU data like MG2 temp.

Using a tuner app on my PDA, if I tweak the reference from a=440 to a=433 then at 60 MPH the hum is the e below middle c. That makes it about 162 Hz. As you'd expect, 45 MPH gives the b below that, and 67-ish is the f# just above.

The tires (Harmonies ... who knew that would be so apt?) are spec'd at 908 revs per mile, or (conveniently) 908 RPM at 60 MPH. So the wheels (and everything that turns with them) are doing about 15 revs per second at 60 MPH. The hum is about 10.8 times that frequency.

The transaxle's reduction ratio is (75/26)(44/30)(36/39)=3.905, so MG2 does about 59 revs/sec at 60 MPH. The hum's about 2.75 times that.

Does anybody happen to know the number of poles in MG2's rotor, or balls in the wheel bearings? (For that matter, is there a nice formula for noise pitch from a ball bearing? I assume there is, but I get dizzy trying to work it out since the balls are moving too.)

Not to leave any clues out, I did hit the mother of all potholes a few weeks ago (it's hard to say exactly when I first started noticing the hum, but that's roughly in the ballpark). That left a big sidewall bulge in the left front tire. I replaced that and one other tire, rotating the new pair to the rear. When the bad tire was demounted, we saw the wheel also got a flat spot. I haven't had time to buy another wheel yet. The unround wheel (with new tire mounted) wound up in the right rear, which *is* where the hum seems loudest. But there's only one flat spot on the wheel, so from that I'd expect 15 or maybe 30 Hz, not 162.

What would be your next diagnostic ideas?

Thanks,
-Chap

 

I think we're agreeing ... I had wheel speed of 15 revs per second, giving MG2 speed of 59 revs per second (15 * 3.905) - is that what you get?

So the sound (162 Hz) is either produced by something vibrating about 10.8 cycles per wheel rev, or something vibrating about 2.75 times per MG2 rev. Since my first post, I found a photo of a classic MG2 rotor that makes me think it has 8 magnets, so I would naively guess that MG2 noise would be 4 or 8 times the MG2 speed (say, 236 or 472 Hz at 60 MPH) rather than what I am hearing. But also since my first post, I also found a thread where Bob Wilson was seemed able to analyze a recorded hum waveform to see if MG2 is the culprit, so he's probably well beyond my guesswork as to what the sound should be.

You may not believe this, but I'm still carrying around an ancient Handspring Visor onto which I loaded Ondrej Palkovsky's "TuneIt" app years ago. It doesn't quite act like a real tuner (using the microphone and identifying your pitch); it just plays any pitch you select so you can match it by ear. It's ok if it doesn't drive you too crazy trying to match the pitch of a square wave. It has a metronome feature too.

-Chap

 

While I know the stator is wound for 3-phase, I've been assuming (optimistically?) that in the case of the sort of shorted-winding defect you've described elsewhere, the fault current we care about is just in the unintended loop created by the defect (since the sound isn't affected much by shifting to N and open-circuiting the intended paths) and so the vibration would come from the interaction of the rotor magnets with one stationary defect in the stator, at a frequency like (MG RPM)(n rotor magnets/2).

You know, I haven't been around these forums for a while, and last I knew, it was hard to find a miniscanner and I couldn't do much with a ScanGauge in an NHW11, but I've been catching up a little and it looks like vincent1449p with you and w2co have rocked that reality. Wow, that's great. I bought a usb-programmable Elm OBD interface a couple years ago with great intentions of trying to figure the same stuff out, but right now I'm tempted to just buy an SGII and use vincent's tables. Maybe later I can play with a laptop and the Elm box using the same tables. Thanks for working all that out!

It would be hard for me to say, since my driving patterns are all different now dealing with my folks (short in-town trips replaced with 300 mile jaunts). But it doesn't seem obviously impaired.

That may be one of the next things I try. The math doesn't really drive me nuts, it's kind of fun and one of the few things I can actually do while driving six hours one way or the other, which is the most time I've got that isn't dealing with other things.

Thanks! I guess the 8 magnets are arranged like NS SN NS SN NS SN NS SN to get the 4xRPM frequency? (I think there's an illustration in the NHW20 features manual showing the NHW11 magnets arranged sort of tangentially to the rotor rather than radially.)

You caught me, I should have said that more carefully. I guess I'm seeing it as kind of analogous to DNA evidence, which is really useful for exonerating people (it's pretty obvious when it doesn't match), but not so much for building databases and hauling people in because a computer search matched them at 500 gene loci or whatever with some hair sample.

In this case, because my hum is not 4xMG2 rpm (not even any integer multiple or divisor of that), right now I'm feeling a lot better about continuing to make long trips in the car, and just trying to pin down a tire/wheel issue as I can. I'll probably try to get a reading of MG1/2 temps too, just for comfort. Maybe this hasn't completely ruled out the tranny either (reality could turn out to be more complicated than the model in my head, which happens a lot really) ... but if my hum had turned out to be exactly 4xMG2, I might still not be certain MG2 was the culprit but I might think seriously about renting a car for all this travel, because there's not enough slop in my schedule to be possibly looking at a red ! on the dash 200 miles from home some night at 2 am.

Maybe I'll see if there's time this weekend to move a tire around. (Looks for wood to knock on.)

Thanks for all the suggestions,
-Chap

 

Lady luck seems to be smiling on me this time. My sense that the hum was loudest in the right rear seems to be borne out. I put the spare on in that corner, as Patrick suggested, and noticed the hub felt a bit notchy when rotated by hand (schezbzflat!) and noisy when spinning the tire by hand.

I went for a test drive and the hum was still there. I had hoped that the spare tire's revs per mile would be different enough for me to notice the hum was still there but at a different pitch, which would have been *really* strong evidence it was that bearing. As it turns out, they matched the revs per mile so well that I could scarcely hear a difference in pitch. But I think I still have enough evidence to convict the bearing.

Looks like I'll be ordering a hub/bearing assembly (wow, a bit pricier than I expected). I think I'll look for a round wheel too, before changing the hub, in case the flattened wheel contributed to the bearing condition.

Thanks for all the suggestions!
-Chap

 

That did it, no more hum. Picked up a rounder wheel on Monday and took it to Discount Tire, where they didn't even charge me the usual $12 to demount the tire from the square wheel and mount it on the round one.

The hub/bearing assembly came in at my Toyota dealer on Wednesday for about $262. I had not seen the earlier thread on here about an under-$200 Timken alternative. Does anybody know, does the one from Timken include the ABS sensor, or does it have to be moved over from the old assembly? The Toyota assembly comes with the ABS sensor in place.

Replacement took me about 4 hours, but that includes a break for dinner and a short drive to retrieve my slide hammer.

The bulk of the time was spent on the second half of step 5(b) on page SA-45:

5(b): Remove the 4 bolts and axle hub assembly.

Sure sounds nice. And the bolts came out with no trouble at all. It's just that that the hub assembly has more than a flat flange for the bolts; its cylindrical body protrudes past the flange, right through a big center hole in the brake backplate, sitting right in the big central hole of the mounting bracket, and it did not want to come out.

Putting the bolts back in (just a few threads worth) and tapping them didn't budge it, maybe just because it was awkward to get a good angle with a hammer. That might be easier on a lift.

This is where I sprayed penetrant all around where the hub comes through the bracket, and took a dinner break.

What did the trick was to torque a slide hammer down onto the wheel studs and ring the chimes for about five minutes straight. That gets the hub to part from the mounting bracket, but not from the brake backplate, which still comes along with the hub and firmly believes it's permanently attached.

Having the bolts still in place by a few threads (from the earlier tap with a hammer attempt) is still helpful so when the hub/backplate finally pull loose from the bracket they don't just dangle by the brake line.

To separate the backplate from the hub, I then took the bolts (through the bracket, backplate and into the hub) out, and put them back only through the backplate and into the hub. For that I had to pull the hub/backplate gently away from the bracket about as far as I wanted to bend the brake line, to make room to sneak the bolts in. Then I sprayed more penetrant behind the hub/backplate where I couldn't reach the first time, pushed it all back so the bolt heads were against the bracket, and tapped the backplate in that direction using a hammer and a big blunt drift (end of a ratchet handle really) around the edges of the hub flange. I wound up having to remove the brake shoes to get good angles for that.

The new hub fit easily right through the backplate and into the bracket bore. It's not even a tight fit, minus the corrosion. For future convenience I used some anti-seize (not on the bolts, only where the hub goes through the backplate and bracket).

Installation step 6 (confirm the ABS sensor signal) is kind of fun. You ground Ts, watch for the ABS light to start rapid blinking, leave the brake unpressed for a sec, press the brake hard for a few secs, then make a straight-line run fast enough and far enough for the light to stop blinking. Then you stop, ground Tc, and watch for the light to blink a normal code (or a code saying what wasn't normal). It tested out just fine but I had to do the test twice; the first time I didn't accelerate hard enough, and didn't reach enough speed for enough distance to make the blinking stop before I ran out of street.

-Chap

 

Followup: three years later, the saga continues with a front bearing this time. The differences between the fronts and rears are interesting.

-Chap

 

If you don't follow the blue link, maybe you'll never know!