DIY Gen 1 MG2 Replacement P3125 P3000 P3009 P3191

I never saw anything published by Toyota on how to do that; it seems they do it at the factory and then say "don't touch that".

But you can probably get some ideas out of this Ford patent just published three years ago.

It really does seem like Ford has a better idea. Instead of having to fuss with micro-positioning the resolver just so, they'll let it be somewhere in the ballpark, and program their ECU to try a range of offsets on either side, in software, and remember the best value, and just apply that to the actual resolver reading. The method in the patent would allow them to do that while driving. (Not normal driving, but driving in a special learning mode. Still, pretty neat.)

Column 9 of the text (page 12 in the pdf) is where the yummy math starts.

I wonder if there'd be a way to build a sort of resolver faker for a Prius, something you could plug in between the real resolver and the ECU, and advance or delay the apparent reading over a small range of degrees. I'm not sure how easy that would be as a pure analog circuit; maybe it would be easier to buy the same IC that's used inside the ECU to turn the analog sine/cosine signals into a 0 to 4095 digital rotor position. (It's an off-the-shelf IC; somebody posted its data sheet here years ago.) Then just add or subtract a small offset (wrapping around from 0 to 4095). The fun part would be the D/A conversion to turn that back into the analog signals the Prius ECU is expecting.

Then you might be able to run the Prius in a stationary way, say on a dyno, and sweep the fake resolver signal over a small range ahead and behind the real one. If you also had measurements of the MG2 U, V, and W phase voltages and at least two of the currents, presumably you could apply the same Clarke and Park coordinate transformations discussed in the patent, find the zero crossing in the transformed space, and use the faked resolver offset at that point to work out what the real one should be.

The last step, instead of having the ECU remember the offset as Ford wants to do, would just be to go back to the real resolver and loosen and nudge it by just that amount, then tighten it real good.

 

See the Weber Auto video on how to get it probably close enough:

 

Man I cannot imagine doing this with all the 1nz import engines avail from Asia. But uhhh good show. My Aus buddies like these long drawn out jobs. Some their cars down a year whilst 5hey work and tear down etc . I couldn't stand the wait. So I keep 3 cars and a few motorbikes around inexpensive. And never wait to go

SM-A715F ?

 

I just did this in my gen 1. The step that is not listed in the tear down...

Before removing the k frame bolts, the steering column needs to be detached from the rack. This is done inside the car under the floor mat just behind the gas pedal. The battery needs to be disconnected and the position needs to be marked and put back in the same poston. It is a splined shaft, but not keyed. Be sure and mark the splined shaft as the connection is being removed. It can be reassembled in numerous positions. If the steering column is not removed, a lot of pressure is put on the rack shaft and the steering wheel will pull to the right from it getting bent under the weight of the transmission being lowered.

 

Great point, this step is the same on a GEN2, I have done it twice without issues.