Resonant tone at speed and red triangle, ((!)), and VSC warning lights

I'm not sure if this is a general FWD car problem or a Prius problem but a week ago I noticed a loud, and reasonably pure, tone that seems to be coming from the front of my '05 Prius at highway speed. I don't believe it is tire, alignment, or road noise. The car knows that something is wrong because it occationally throws some error lights but refuses to reveal the specific code. My current thought is drive shaft, axle, CV joint, or similar but I don't know if that can turn on the error lights. Any thoughts? Could I have two independent issues?

The car has 157k miles and was in two “minor†accidents before I got it. One required a rear bumper replacement, the other they had to pop out a dent from the front driver's side quarter panel. The records do not indicate that any other parts were damaged. I have owned the car for 25k miles.

Error lights
A few days after the noise started and 30 minutes into my morning commute while driving at highway speed up a hill (~20kW from the ICE) I got the big red triangle of doom, ((!)), VSC, and the ! on the red car on the MFD. Aside from the lights the car appeared to drive normally. I checked the voltage on the MFD less than a minute after the lights came on and it showed 14.1V. The MIL light was not lit, and the Torque app on my android was not able to pull a DTC but was able to clear the error and consequently turn off the warning lights.


Click on the picture to see the pretty lights

The next day the same warning lights popped up again and at the end of the commute I used a paperclip to connect pin 4 and 13 on the OBD port in an attempt to read the code. When I turned the car on with the two pins shorted the red triangle lit but the ABS, ((!)), and VSC lights all blinked at 2 Hz indicating their systems were normal. After turning the car on and off four times the red triangle went away.

The error lights have now popped up four times, each time after more than 30 minutes of driving. I have also had several trips longer than an hour where the lights did not come on.

Noise
The noise is closer to a pure tone than a hum because it is reasonably high in pitch and, while not sounding exactly like a 440 Hz tuning fork, it is surprisingly clean. The peak is around 457 Hz (A#) and you can see how narrow it is in the attached graph. It is loudest at 58 mph indicated but is audible from 57 to ~64 mph. The volume seems to depend on the torque to the wheels as well as the speed. At 60 mph the tone will be quiet if you are feathering the accelerator and the power to the wheels is between -3 and 3 kW. If you are applying or withdrawing more than about 5 kW the tone is reasonably loud at 60 mph. While the tone depends mostly on speed and torque, it doesn't care if the power is coming from the gas engine or going either direction through the electric motor.


Frequency spectrum histogram of cabin noise at ~60 mph

When I replaced my tires this week I went from well-worn FR710 185/65R15s to new Hankook 727 195/65R15s. This change corrected my speedometer error but did not change the indicated speed that the noise starts. In other words the noise depends on how fast the wheels are rotating, not the actual highway speed.

According to the priuschat forums the car assumes about 845 wheel revolutions per mile. At 60 mph indicated that works out to 845/60= 14.1 RPS. 457Hz is about 32.5 times larger. Stepping back through the 4.113 reduction you still end up with the tone being ~7.9 times faster than the rotation speed of MG2. (If you use the MG2= 59.1*MPH equation you get ~7.7) I consider this a good but puzzling thing; I didn't want the noise to be an even multiple of the rotation speed of an expensive part, but I would have liked to have pinned the source. I'm left assuming that it is the resonant frequency of some part or system.

Attempted diagnostics
Since this started I have gotten new (badly needed) tires, checked the alignment, and had the transaxle fluid changed but the problem persists. I looked though the forums for problems that threw these specific error lights and tried the tests that I could but did not find any red flags. My 12V battery is older but still holds enough charge (12.6 after overnight rest, 11.9 under load test, 14.1V typical when READY). The 3.5 quarts of 5W30 engine oil was changed less than 2000 miles ago and continues to read right in the middle of its range on the dipstick. The inverter coolant pump was replaced under the recall and is happily chugging along. According to the Torque app the inverters were running at 20-30C when the error lights came on.

All this leads me towards thinking the noise is coming from somewhere on the drive train between MG2 and the wheels but I'm not sure that explains the error lights. I'd love to hear to any other thoughts, suggestions, or diagnostic tests I can run before I have to take it to a real mechanic.

 

All,

Thanks for the replies. I had another hour drive this morning without any error lights but the noise seems to have worsened. To date the CEL or MIL never lit, I only got big red triangle, ((!)), and VSC last week. I’ve driven about 3 hours since I got my new (and delightfully quiet) tires put on and the lights have not re-lit since the change. I won’t declare a negative result for a while, but could a tire-imbalance, alignment problem, or something like that have set off those lights through a wheel speed sensor or other means? Assuming I get errors again I will take it to a dealer to get diagnosed.

There is a new “wind-up†noise when accelerating between 15-30mph and a “wind-down†noise when decelerating through that same range. The noise is louder than the inverter whine but slightly quieter than the engine. The pitch is clearly linked to the speed. It reminds me of the reverse noise from a ‘90s Honda transmission but it isn’t anywhere near as loud. I’ll try your side-to-side wheel-bearing test in a large abandoned parking lot after work.

The noise at highway speed has changed a little bit as well; it is louder across the whole range. It now becomes audible around 55 mph when under power and fades out around 67. The pitch rises slightly over that range as well so that makes me think it is less likely to be a resonance and more likely to be a rotating part with some slop or a MG2 problem. The Gen II’s MG2 is an 8 pole motor so the ~460 Hz highway noise could be generated by a bad winding. If you want to see some neat pictures comparing the MG2 on Gen I vs II take a look at this .pdf: www.ornl.gov/~webworks/cppr/y2001/rpt/120761.pdf

I don’t have a multimeter at the moment; my 15 year old one died two weeks ago and my next one should get here tomorrow. As a typical cheap DMM it won’t have the resolution to see a difference between the MG2 windings so I’ll rig up a power source and measure DC voltage and current tomorrow evening.

I just added MG temperature gauges to my torque display and will watch them on the drive home though even with a partial short I wouldn’t expect to see anything anomalous.

I had a dealer change the transaxle oil and the only thing they reported back was that it looked “normalâ€. It was last changed about 37k miles ago. I don’t know if they meant normal for 37k, normal for a 120k interval, or if they didn’t really look at it and just said that to placate me.

I plan to get a new 12V Optima battery with whatever money remains after solving this problem.

 

Somewhat mixed results

I did the tight-turn-in-a-parking-lot test and discovered that, once all the change in my car finished moving to the outside of the turn, I could hear that the right front bearing sounded a little “raspyâ€.

I also monitored my commuting MG temperatures as reported by Torque and found that one of the MGs was 50-70C hotter than the other but there are two problems. First the app recently changed how it reports temperatures from extended PIDs so I was only able to get a relative temperature difference. The second problem is that I don't know which was hotter. If there isn't a typo in the spreadsheet MG2 is hotter; if the spreadsheet was wrong MG1 was hotter.

I do not know how my '05 prius was driven for the first 133k miles of its life but I've tried to be pretty gentle on the powertrain for the most recent 25k. I have one on-ramp that takes 30-35 kW but the vast majority of my commute is spent requesting between 12 and 20 kW from the engine. I try to minimize power into and out of the battery because of the low round-trip efficiency. In the fall I was able to sustain 57-58 mpg but with the cold weather my mileage has dropped to about 50.

The winding failures I have seen were all on less than 1kW DC air-cooled electric motors. As far as I can tell those tend to happen suddenly and with gusto. Can winding failures creep up over 88 million revolutions (25000 * 855 * 4.113)?

 

I wasn't able to ohm out MG2 so I drove the car to my old mechanic, dropping it off after he closed for the day. The tone continued to get louder and wider; it can now be heard starting at 52 mph. After 15 miles at 55-60 mph MG2 was 85C hotter than MG1.

The next day the mechanic hooked it up to his Mastertech scanner and was able to pull VSC codes that pointed to the steering position sensor calibration. Satisfied, he called me to let me know the good news:
Mechanic: Car fixed, bill <$150!
Me: Drive the car at highway speed and then get back to me. It will sound like a 440 Hz tuning fork.
Mechanic: I don't really remember what that sounds like.
Me: You will.
Mechanic: [eyes roll]
...20 minutes later...
Mechanic: I don't recall tuning forks being that loud. How did you diagnose it?
Me: Priuschat, OBDII scanner into Torque app...
Mechanic: What is your occupation?
Me: Engineer
Mechanic: I knew it! You can always tell engineers apart.
Mechanic: Do you want a new or a used transaxle?

Quoted $926 for transaxle labor, $110-210 for fluid and misc hardware. I'm buying a used transaxle from Autobeyours for about half the price of a locally-sourced unit.

It appears that I had two separate problems:
In the VSC system the steering calibration was way off and the yaw sensor was somewhat out of specification. A quick recalibration fixed the issue.
In the transaxle a shorted MG2 winding caused lots of heating and speed-dependent ringing noises. Transaxle replacement should solve the problem.

Thanks for all the help!

 

Car is all fixed up and driving beautifully with a new transaxle. After 30 minutes on the highway MG2 is about 5C warmer than MG1, a huge improvement over the 80C difference before the repair.
Costs:

	Column 1
0	Labor 1029.50
1	Transaxle 650
2	Shipping 425
3	Toyota Fluid 71.90
4	Misc Supplies 59.95

Transaxle replacement total:
$2236.35

I also had both rear wheels shimmed to reduce total toe from 0.6 to 0.2 degrees. That cost:

	Column 1
0	Labor 205.9
1	Alignment 89.95
2	Shims 47.72
3	Discount for giving them all that other work -100

Total: $243.57