So who wants their ECU's "Chromed"?

[Greg E]

Greg do you still recommend flashing the car while it is running and just letting it die?

You can do that but the best way to flash is with a charger/tender on the battery.

[Chris@Rvengeperformance]

You can do that but the best way to flash is with a charger/tender on the battery.

Does anyone know what actually goes wrong with the ECU is damaged from a bad flash? I assume there is software on the CPU necessary to enter writing mode and it gets wiped out, but there must be a way to fix this without replacing hardware, right?

I mean I suppose there would be 2 different types of failure. One that is just corrupted rom data and another where the flash has "worn" out and can't be written to anymore.

If you want to take this to Emails, let me know.

[Jimvr4]

I don't know our ECU architecture but the Cable Modems I designed used flash for program storage so a corrupted flash would fail to boot properly. We could recover by loading the flash through a serial connection to the processor or by soldering a new flash on board. Both methods were beyond the user. We had to investigate and eliminate flash corruption rather than deal with failures.

[Greg E]

Does anyone know what actually goes wrong with the ECU is damaged from a bad flash? I assume there is software on the CPU necessary to enter writing mode and it gets wiped out, but there must be a way to fix this without replacing hardware, right?

I mean I suppose there would be 2 different types of failure. One that is just corrupted rom data and another where the flash has "worn" out and can't be written to anymore.

If you want to take this to Emails, let me know.

The data becomes corrupted when there isn't proper voltage to ensure a complete flash. The hardware got better on the later Mitsu ECUs as they started designing the boards with their own voltage booster instead of needing external ones like we are forced to use with the tactrix 2.0 cable.

I'm not convinced the board can "wear out" from too many flashes. I have 300+ flashes on my ECU. The 99 had about the same and I know Jesters Deadd is up there with me. The H8 manual only guarantees the CPU is good for 10 flashes but I'm guessing it's like a light bulb. They're only rated for so many hours but almost all of them last much much longer. They only under rate them for warranty reasons.

It's also happened with me a few times where a CPU became corrupted but after multiple attempts to reflash somehow it would come back to life and be just fine...

People need to be aware of the risks when flashing as bricked ECUs from low battery voltage is starting to become a common enough of an issue where the importance needs to be stressed. Guys should turn their stereos, headlights and AC OFF when flashing to ensure as little draw on the battery as possible. Weak batteries from dying alternators need to be replaced BEFORE attempting to flash tune your ECU. EVOScan has the ability to read battery voltage at the ECU. If it's under 11volts, you're in danger of bricking during a long enough flash!

[Greg E]

by soldering a new flash on board.

This is how Brett and Adam fix bricked ECUs.

[Jesters Deadd]

I don't know our ECU architecture but the Cable Modems I designed used flash for program storage so a corrupted flash would fail to boot properly. We could recover by loading the flash through a serial connection to the processor or by soldering a new flash on board. Both methods were beyond the user. We had to investigate and eliminate flash corruption rather than deal with failures.

So far with all the bricks we have seen it has only been with a low battery. These are pretty robust ECUs as long as your battery is good. I regularly start a flash with the car running. I have not seen any problem with this method. Also it doesn't always brick with a low battery. Some times 50% or so its recoverable.

[Jesters Deadd]

The tell tale sign it is bricked is it will hang during the flashing process. If you look at the code it always hangs on the same address then after a couple seconds fails. If the ecu is recoverable it will unhang after two or three trys at reflashing it with a fully charged battery.

[ChargerX3]

Who else is bricking ECU's? Its also nice to have one you can test with for initial version changes. I always use my Diamante just to be sure. So far so good.

[Jesters Deadd]

Who else is bricking ECU's? Its also nice to have one you can test with for initial version changes. I always use my Diamante just to be sure. So far so good.

There is no need for testing initial version changes. Thats what the testers are doing. The only people I know that have bricked ECUs are ECU builders and testers. Thats how we know that its a low battery that causes it.

[Jesters Deadd]

There's already a fuel cutoff in the stock code - hold the throttle to the floor while cranking and it won't inject fuel.

BTW tested this the other night... FAIL.
BUt its a great idea and GReg has agreed to eventually add this feature.