steering gear jitters

I'm not sure I understand the tin-whisker phenomenon well enough to connect it to a resistance that decreases while on an ohmmeter. Back in 2008 I was imagining more an oxidized contact surface (which is why I spent time on those ideas about wetting current). On the other hand, my 'before' trace showing VT2 apparently glitching toward zero does seem easier to explain with a VT2-GND whisker than with any kind of wiper surface deterioration. In any case, I can imagine a wall-wart trick might be helpful with either condition, or both.

The warranty extension letter I received said the condition is covered by Toyota through 31 December 2013 (15 months from now) with no limitation on mileage. The wall-wart approach will be good to have in our bag of tricks for any problems appearing thereafter.

-Chap

 

Glad it's getting fixed!

For anyone else worried about that, remember that the steering ECU is already designed to have the highest amplification at low speeds, like parking, when steering takes the most effort, tapering off to almost nothing at highway speeds. Although a bad torque sensor will be sending noisy signal all the time, the effect is pretty much imperceptible except at very low speeds, getting in and out of parking spaces, etc.

The one thing that does eventually happen is that the ECU's threshold for noisy signals gets exceeded and the ECU logs a code and quits, giving you a manual-steering Prius. It will reset on the next IG-ON. So depending on how bad your torque sensor is, this will only happen on long drives. The first time it happened to me was on a 14-hour trip back from my sister's (and just within a few miles of my driveway). By a year later it was happening on any drive exceeding two hours. Again, because the assist gain is tapered off at speed, if the ECU happens to quit at highway speed you won't even really notice until you slow down or try to turn from a stop. Then it is harder to steer, but you can IG-cycle if you want your assist back. To be honest, at driving speeds I kind of liked the slightly tighter steering feel with the assist off.

At least that was my experience, with the particular way my torque sensor failed. It was an annoyance but I never felt unsafe driving the car while I documented the problem before taking it in.

-Chap

 

At 2½ years after replacement:

	VT1-VT2	VT1-GND	VT2-GND	VT1-VCC	VT2-VCC	VCC-GND
1	1231	1425	1427	615	617	810

Still solid steady readings, not moving around on the meter as my old bad one did. I'm thinking about this again because Behrens may have found that whisker zapping as proposed by Bob might not be as effective as we all hoped, so now that the warranty extension is past, steering jitters may one day await us all....

Interesting that over 2½ years nothing seems to have changed at all except the resistance to ⏚ gaining a few ohms. Not what I'd have expected exactly.

I've always kind of wanted to solve this system to quantitatively draw the parallel-potentiometer picture I think is down there, only setting up the system with the measurements above makes my head hurt. So this time I also took this set:

	VT1 to (Vcc ⏚)	VT2 to (Vcc ⏚)	Vcc to (VT1 VT2)	⏚ to (VT1 VT2)	(⏚ VT1) to (Vcc VT2)	(⏚ VT2) to (Vcc VT1)
1	615	617	308	1118	350	350

Weirder measurements to take, but at least they lead to a system where each equation involves just two unknown resistances. Now to see if I can make any sense of it.

-Chap

 

The alert reader has probably already noticed what bugs me about this: this does not look like a voltage divider that would give you 2.5 V at VT1 and VT2 when fed with 5 ... but plug it into the ECU and that's exactly what you get.

Either there is something more down there than the pictures in the manuals suggest ... or maybe the ECU's VT1 and VT2 inputs have much lower impedance than I would have guessed, and pull the signals down. Maybe they even did that on purpose to draw a constant wetting current, as I was speculating about in #25 ... which would as a side effect make the voltage to torque relationship nonlinear. I've still never built any jig to make voltage to torque graphs....

Or am I missing something really simple?

-Chap

 

Update: the secret is to pry the wedge out.

-Chap

 

Exactly ... before Toyota replaced mine, I had it cut out on me a couple of times. Apparently the ECU has a threshold number of glitches it will tolerate (within one IG-ON session) and then it gives up, so it never happened to me except on long-enough trips (> 10 hours the first time) on the highway. Generally I never noticed the exact moment it cut out (on the highway, at highway speeds) ... I noticed when finally arriving at my exit and trying to turn a corner.

-Chap