Corrupting Firmware

[jsdevel]

What are the top ways to corrupt firmware?

Bad cable?
Unstable voltage during flashing?
Loss of power during flashing?
Flashing without wiping the controller clean?

[jsmcortina]

What are the top ways to corrupt firmware?

Bad cable?

Possibly.

Unstable voltage during flashing?

Unlikely. The processor needs 5V. The voltage regulator should clean up most instability on the 12V.

Loss of power during flashing?

100% certainly.

Flashing without wiping the controller clean?

Not sure what you mean here. The reflash process includes a mass erase.

James

[jsdevel]

Thanks James!

[Matt Cramer]

Another way to brick an ECU - at least until you reload firmware with the boot jumper; this isn't an unrecoverable bricking - is to set some sort of impossible settings that had not been anticipated in firmware. One example we found a few years ago was that a user had set up a crank trigger where the number of missing teeth was equal to the number of base teeth, creating a zero tooth trigger wheel. Obviously, there's no way to decode this, but it didn't occur to anyone that someone would try it. We've tried to catch every settings conflict, but occasionally somebody gets one through that nobody anticipated.

[gboezio]

We had a 12V starter cable to come loose and spark on the headers, spark shower lasted for about 30 minutes with a constant relay flickering, huge misfires sequences, tranny gears hunting mostly default 3rd. We welded the header and changed the cable and it's like nothing happened.

[jsdevel]

Another way to brick an ECU - at least until you reload firmware with the boot jumper; this isn't an unrecoverable bricking - is to set some sort of impossible settings that had not been anticipated in firmware. One example we found a few years ago was that a user had set up a crank trigger where the number of missing teeth was equal to the number of base teeth, creating a zero tooth trigger wheel. Obviously, there's no way to decode this, but it didn't occur to anyone that someone would try it. We've tried to catch every settings conflict, but occasionally somebody gets one through that nobody anticipated.

What is the best way to recover from something like this? Boot jumper and a firmware re-flash? Would one be able to trust their current tune, or would it be better to manually re-create it from scratching using only the required settings?

[Matt Cramer]

Boot jumper and a re-flash. Whatever you do, don't load the same problem tune or you'd bring the problem back. Your best bet is to make a new tune starting with the default if you do not have a copy of the tune from before the problem settings got set.