S80 burning ignition coils [honda k20]

[stevieturbo]

wiring coil power via the fuel pump relay is fine.

It's just that split second delay from powering the ecu on until the FP triggers that removes the "problem"

If indeed you actually deem it a real enough problem to worry about.

Mine would do it with the DTA and now Syvecs.

But really, it takes some effort on my part to make it happen.

ie lots of short stop/starts, or it running very rich among the stop/starts.

Really, when the engine is tuned and running properly it isnt a scenario that should raise its head.

Although that's getting off topic to the coils burning out which is a real problem.

[MarcoV6T]

An ECU shouldn't do things like that, for me that's a flaw in the hardware design, there’re things one can do to avoid that.

Say if you have a leaky injector(or bad injector output) and after the right time(AFR) you switch on the ECU, BOOM, you could break things.

I’ve a Pectel ECU which has dedicated TTL outputs for amplified coils, they never fire when the ECU turns on, as it should be.

[TRN]

Everytime I switch on the relay my coil(s) fire once -> noticable because some remaining mixture ignites. Don't know if this is just one or all of them. This is something I never had with other ECU's. Could this be a clue as to what's wrong?

This is normal with most aftermarket ecu's, this is because the ignition outputs are grounded for a few msec until the processor boots up. The common way to correct this is to power the ignition coils from the fuel pump relay - then they cant charge until engine is cranking.

That really should not happen, but the ‘problem’ here is the fact he’s running amplified coils, and they fire on a positive edge most of the time.

When you check the ‘amplified coils’ in the DTA settings, the ECU inverts the outputs, thus at rest the coil output is ON and thus grounded, when the coil needs to charge the output goes OFF and a pull-up resistor brings it to +5V or +12V, the coil charges and when it needs to fire, the output goes ON(grounded) again and you have your spark.

The problem with this is that the pull-up resistor is hardwired to the positive rail, so the moment you turn on the ECU, the coil gets the positive edge to charge the coil, in the meantime the ECU checks the settings and turns ON the coil outputs which grounds the outputs, which then fires the coil(and that can be the reason of the failing coils, as the charge time here can be a lot more then 1,7 msec let alone 2,5msec).

Connecting the coil supply with the pump relay is a good idea(also for safety reasons), even though the pump primes for a few seconds when you power the ECU, but it then has grounded already the coil outputs.

Many thanks for all the replies. This sounds like a very plausible explanation of what's happening. I will wire the FP relay to power the coils, change the dwell time table and see if everything stays in shape.

[TRN]

Since I believe in giving feedback for future reference:

Powered the coils through the FP relay, haven't had any coil die on me as of that moment. The dwell table in the linked file was the appropiate one for the coils in use, however I would notice an occasional misfire when using the table (dwell time too short). I reverted to a fixed dwell time of 2.5ms and problem was gone.

Car has actually won a few of the stages since these mods, so many thanks to all the help .