Gen1 2002 Prius Steering wheel spins to the right!

This might not be related to the MG2 work. Gen 1's electric power steering used a pair of potentiometer-style torque sensors to detect how hard you are turning the wheel, and calculate how much steering assist motor power to pitch in with to supplement your effort.



Like all the other potentiometer-style sensors in Gen 1, the steering torque sensor can get flaky.

The most common symptoms are: (1) occasionally, steering wheel shakes like a wet dog; (2) occasionally, often late in a long drive, the steering ECU decides it's had enough of that flaky sensor input, and takes itself offline. Then you have non-power steering, which you might not notice at speed because steering is so easy then anyway, but you notice the next time you have come to a stop and try to turn a corner, or park, and it feels like a truck. IG OFF and back ON gets your power assist back, until next time.

If the torque sensor gives a flaky signal that really looks as if you're turning the wheel hard in one direction, the power assist can follow that input and powerfully zip the wheel in that direction. This is really rare; everything about the design of the system is to make it rare. That's why the torque sensor is two separate potentiometers; they would really have to both agree, say, you were turning the wheel hard to the right, before the assist ECU would be fooled and spin the wheel that way.

But I do think I remember one other thread, anyway, where that did happen to somebody.

You could try unplugging the torque sensor from the steering ECU. The ECU is behind the glove box, next to the ECM.

You can recognize the steering ECU because one of its connectors will have two fat wires (yellow and black/yellow, that's for the assist motor) and one will have four skinny wires (white, red, green, black, that's for the torque sensor).

If you unplug the torque sensor, you'll get a steering trouble code, of course, but if the weird wheel spinning stops, you'll pretty much know what the issue is. You'll then have a manual-steering Prius, and you'll get nicely ripped from the workouts of turning low-speed corners and parking.

 

There was also a purely mechanical thing that could cause that. Nut at the bottom of the pinion shaft would get loose. Then, when you turn left, pinion rides up on the rack and puts a mean side load on the motor gear, hard to turn. When you turn right, pinion rides down again, easy. There was a recall for that.



Unplugging the torque sensor connector wouldn't change that, of course.

 

The pinion nut repair is done with the steering rack removed from the car. It is quite technically demanding (a lot more work, in fact, than just swapping the assembly), because the thing wasn't built to have those nuts ever changed, and when they realized there were problems requiring a recall, they had to come up with a 35-page procedure for how to successfully (most of the time) rebuild the rack with new nuts, and the procedure has several places where failure is possible. (For example, the replacement jam nut is made with such extreme locking deformations that it eats the threads on the pinion shaft; if you don't get it properly torqued and locked on try #1, there is no try #2.)

So if any of those things go wrong on a pinion-nut-recall job, Toyota is left with no choice but to eat the cost of a new entire assembly. No fun for them, good luck for you. There were some people who did go in just to have the nut recall done, and end up driving home with brand shiny new steering racks, because of that.

 

I would have tried removing just the top connector (the 4 skinny, torque sensor wires) and leaving the bottom connector (the motor power) attached. That would show whether getting rid of the torque sensor input is enough to make the ECU stop trying to assist your wheel to the right.

That would be right in later generations, where the sensor is on the steering column. In Gen 1, the torque sensor is inside the steering rack; the rack is what's replaced.

No, the nut is behind a sealant-locked threaded plug on the bottom of the rack. It is really two nuts (an adjusting and a jam nut), and his idea that you "could have tightened it" was fanciful, especially if the nut recall had been previously applied.

The new nuts supplied for the recall have such highly-deformed locking threads that they ruin the pinion threads, on purpose, as they are being installed. There is one chance for the mechanic doing the recall to get it right; if anything goes wrong then, or if the nuts are later disturbed for any reason, it's new-rack time.