For those using EcuTek products, have you ever bricked an ECU using Flash99?

PM me if you'd prefer..

 

Once or twice yes. It happens from time to time, unfortunately the first thing the flash operation has to do is erase the entire firmware, including the bit it uses to flash the ECU. So it relies on what's in RAM, if it fails part way through, or the car battery gets low on voltage, or the laptop shuts down, you're screwed. Don't use a mains inverter while you're flashing, don't disturb the laptop, don't use a phone, don't flash a car with a questionable battery, generally don't mess about and all will be fine.

I've probably had 3 failed flashes out of around 500, and if it happens I replace the ECU at my cost.

Paul

 

Thanks Paul, that's some useful stats. Do you think there'd be any merit in backfeeding the ECU during reprogramming from a bench supply?

 

Yes, I always add a battery charger to the car when it's on the dyno now.
EcuTek tell me the minimum programming voltage is 11.5v.
EcuTek usually replace the ecu for us when one gets fried but there are only limited supply of replacement chips available.

 

You'll have to work out what's best for yourself. I assume you're trying to go down the route of reflashing yourself, in which case be prepared to lose a few ECUs just like Stephen Done did years ago!

 

I was considering Flash99 as a development platform but was made aware of the brick-potential aspect of it. Was just after some experiences

 

What you don't know at this point is where that "brick potential", as you put it, arises. That's something you'll need to explore yourself.

There is no stability/"brick potential" advantage in running an external power supply provided the car's battery is capable of providing a stable input between 11.4 and 12.6 volts throughout the programming process.

If you are going to run external power, it will only have any real value if you power the ECU, in its entirety, from it - as opposed to providing external flash power while leaving the main and standby supply lines coming from the battery.

 

Yes, I did consider this. I'm not aware of any spurious drains that may occur while the vehicle is in a switched-off state which could disrupt things. The safest measure, as you suggest, should be to bench mount the ECU and program it from there. Perhaps not the most flexible solution.

I bet Subaru don't have the same problems when applying a re-cal....

 

In practice that isn't necessary IME, in fact I'd suggest that using the car's own battery, (unless it's fecked), is probably more reliable than using a bench supply with the attendant possibility of someone stepping on a cable or accidentally switching it off, not to mention the increased possibility of longer cable runs introducing radio interference from phones etc.

The reflash routine in the ECU is robust and fundamentally fail-safe - provided it is used properly. As Paul said earlier there's a practical issue with the M32 CPU in the way in which the code is copied to and run from RAM while its underlying ROM is erased and rewritten, but in practice it is possible to recover from almost all error conditions, including, for example a PC crash/shutdown during program, provided you (and the programming interface) knows what to do.

The only one where you need to take the ECU out of the car to recover it is if you have a *total* power failure to it at a point in between the initiation of erase and the completion of reprogramming.

 

Thanks for the info chaps.