jamesportman

It won't let me edit the post but just to say, go to the top level of that repo for updated explanation; https://github.com/james-portman/subaru-steering-angle-sensor/

It is basically the only eeprom chip, it's on the back of the PCB The pin numbers are on the board at the chip corners so you can get a clip on the right orientation, usually pink wire to pin 1

I've just solved the P1767 fault on mine (steering angle sensor circuit), it's quite complicated, I could check and fix sensors for people if posted to me Long version: These sensors don't actually seem easy to break, I think it's just the fact that they set permanent faults and pretend to be broken. The fault for mine was misleading, it wasn't a circuit/wiring fault, so don't take the fault code descriptions as gospel. It happened on mine because the steering knuckle had been off and the steering wheel had turned more than should have been possible. If you do this then the steering angle sensor sets a permanent fault, and supposedly the only way to fix is to get a new sensor. In this case if you use FreeSSM, the SAS will still give sensible readings. Good and bad sensors will always show 2048 when you first start up without having moved the wheel, then once you start moving the steering wheel it should start giving proper readings. There is a small eeprom (memory) chip on the underside of the SAS circuit board where the fault is stored. It took me a while testing but I figured out how to clear the fault on these with a USB eeprom programmer TL866-II with an 8 pin clip. Example of the data that was in mine: https://github.com/james-portman/subaru-steering-angle-sensor/blob/master/text Changing the a5a55a5a data to ffffffff fixed it, working fine, no faults on the DCCD module. There is more info in the github repo if anyone is interested. If you have a sensor that does not report sensible values in FreeSSM then it might actually be broken!