ECM FAILURE POLL AND DISCUSSION

[SMSW]

The megasquirt 3 fully assembled costs between 660 and 700 USD + a wiring harness for another 100. Not sure if this particular model would be required or if there are cheaper options.

New owner of a non ABS model 2003. There was a mention of accessories being a potential cause or contributing factor of ECU failures in another thread. I’d also like to know if not adding lights, cruise control, etc could be a preventative measure.

I don't think anyone has the final answer to your question. However, logically, adding electrical loads and properly fusing them is unlikely to affect the ECU, in either a positive or negative way. @Igofar wondered if improper tinkering could damage the ECU and the answer to this is of course! Any electrician or electronics tech knows you can damage other components by shorting something out or by sending electricity on a path for which the component was not designed.

I will add that in a line voltage situation (household wiring) some dimmers will add electrical noise to the circuit, and some fluorescent ballasts make so much RF noise that using a radio with an antenna is impossible. This noise sometimes interferes with the proper operation of other electronic devices, but AFIK, no such issues have been found with our bikes. Enough of us have added LED lights and there have been relatively few ECU failures so we can assume the two are not related.

 

[Bmacleod]

I get you... hmmm... seems like a maybe... I think the ABS module also changed... I think.. I don't know if individual sensors are different and ergo would or could send different data or voltage or whatever the sensor does...

Yes, I thought about the ABS module as well and that might have to be changed also. That's why this method would probably be considered last resort because you're basically updating the entire bike to late model electronics.

 

[Bmacleod]

Of course,

But what v8-7 suggested was to REPLACE the ECU with one of the standalone computers available today for that, like a microsquirt or a megasquirt. He was NOT talking about repairing the ECU.

His suggestion is for when there are no more ECUs to be found anywhere, to replace our defective, non-rpairable ones, so we have to come up with an alternate control unit.

If we have to go this route, we will still have to set up the replacement control unit, or even program it, and I only mentioned Yvan as somebody who might be able to help us emulate the ST ECU with such a computer, due to his extensive knowledge of the ST ECU derived from his work on a FCE solution.

Then again, there could be enough knowlege on this forum to tackle this without further help.

Okay, I see. I misunderstood

 

[W0QNX]

Joe those two above, which I think are for for the After 07 Standard (non ABS) and the 03 - 07 Deluxe (ABS) look the same.

An 03 deluxe vs 03 to 07 deluxe and After 07 deluxe are not the same pattern, They have different connectors. 03 deluxe has two WPC 26 P (26 pin) connectors and the newer 03 to 07 standard and deluxe have WPC 33 (pin) 025 connectors.

The non ABS diagrams are wired different starting with the first wire. On a non ABS the BR wire, first hole goes to the fan, ABS the BU/Y wire, first hole goes to the coils so there are differences in both the connectors and the wiring patterns.

I'm not sure what the plain 03 would be vs the 03 - 07 designation.

Just for clarity, I have no idea what will interchange. I'm only comparing wiring diagrams.

 

[Mellow]

Joe those two above, which I think are for for the After 07 Standard (non ABS) and the 03 - 07 Deluxe (ABS) look the same.

An 03 deluxe vs 03 to 07 deluxe and After 07 deluxe are not the same pattern, They have different connectors. 03 deluxe has two WPC 26 P (26 pin) connectors and the newer 03 to 07 standard and deluxe have WPC 33 (pin) 025 connectors.

The non ABS diagrams are wired different starting with the first wire. On a non ABS the BR wire, first hole goes to the fan, ABS the BU/Y wire, first hole goes to the coils so there are differences in both the connectors and the wiring patterns.

I'm not sure what the plain 03 would be vs the 03 - 07 designation.

I used what was in the post I referenced so could be those are incorrect. I'll kill the pictures.

 

[W0QNX]

I used what was in the post I referenced so could be those are incorrect. I'll kill the pictures.

There are 6 diagrams in the post you linked.

 

[Ryan_B]

What I find odd about this whole thing, is the ECU is only looking for a knock signal after 4000rpm? It seems like if you had a knock sensor (or signal) issue it would present immediately. But possibly the software only looks for a signal in higher rpms.

I wonder if you could "trick" the ECU by adding the appropriate size resistor, or the appropriate voltage for the knock sensor that is functioning correctly. I know, I know, I probably just gave Larry a heart attack - but if it stopped these bikes with a completely faulty ECU from going into limp mode, it might be worth the very small risk of burning up the engine with a knock that it couldn't detect.

Ryan

 

[Larry Fine]

It would mean always, always, always use only premium fuel.

 

[ASPC]

What I find odd about this whole thing, is the ECU is only looking for a knock signal after 4000rpm? It seems like if you had a knock sensor (or signal) issue it would present immediately. But possibly the software only looks for a signal in higher rpms.

I wonder if you could "trick" the ECU by adding the appropriate size resistor, or the appropriate voltage for the knock sensor that is functioning correctly. I know, I know, I probably just gave Larry a heart attack - but if it stopped these bikes with a completely faulty ECU from going into limp mode, it might be worth the very small risk of burning up the engine with a knock that it couldn't detect.

Ryan

The power profile changes around 4K. Different ignition timing and FI curve. I’m guessing the ECM drops to the pre 4K map if it detects a knock. Or perhaps an altogether different map.. get you home limp mode, as you say.

 

[ASPC]

Anyone remember when Dell had a major recall due to cheap capacitors failing in the warranty period? It was a big deal at the time.

 

[ASPC]

It would mean always, always, always use only premium fuel.

The user manual says to only use premium fuel.

 

[ASPC]

If I remember correctly? You can refit a bad ECM after it has been stored a while and it will work correctly for a short while. Which points to a bad component. It may not be the knock sensor circuit at all, a bad component could be very well cause a real knock.

 

[ASPC]

I don’t know anything about the alternators on these bikes and I would really like to know the output at specific RPMs. ???

I wonder if the problem is caused by a burned out varister in the ECM. It would be nice if it was.

 

[Bmacleod]

Anyone remember when Dell had a major recall due to cheap capacitors failing in the warranty period? It was a big deal at the time.

I was just thinking about that. They called it the great capacitor plague and it went from 1999 to 2007. Reportedly it cost Dell $300 million dollars. they said the majority of the capacitors were from Taiwan but some were also from Japan. If that were the case with these ECU's, it would be evident after removing the potting material. The capacitors would bulge and break open from faulty electrolytes. I built around 20 computers during that time and I had to RMA probably five motherboards that had the problem.interesting read:

https://en.m.wikipedia.org/wiki/Capacitor_plague

 

[Ryan_B]

The power profile changes around 4K. Different ignition timing and FI curve. I’m guessing the ECM drops to the pre 4K map if it detects a knock. Or perhaps an altogether different map.. get you home limp mode, as you say

Okay, that does make sense. It is not "looking" for a knock prior to 4000rpm, and when it does if it is not getting the correct voltage signal it throws a code.

Automotive (once again, I am putting my foot in my mouth by comparing), when a knock sensor fault is detected it will retard timing as much as possible and I believe dump more fuel on some cars. This is all from the ECU assuming there is a constant knock, and it is desperately trying to save your engine from pre-ignition and burning up. It is not a "limp" mode as I mentioned before, but it is a fault mode and it will not self clear.

Some vehicles with variable cam timing will not allow the lobes to slide on the cams when it is in this mode, limiting rpm/power, which sounds similar to what is happening on ours.

There has to be an internal short on the ECU board, now is that a component (resistor, capacitor or ???) Or is it simply just a circuit path/pin that needs to be re-soldered? I guess the hardest part of checking these out would be trying to get the sealant/encased resin stuff off without damaging the board and components.

 

[Larry Fine]

I still think you guys should eliminate the melting sensor wires issue first.

Look for discussions about the sensor melting against the exhaust shields.

 

[ASPC]

I still think you guys should eliminate the melting sensor wires issue first.

Look for discussions about the sensor melting against the exhaust shields.

I checked everything and then replaced the ECM. Problem solved. But it would be nice to be able to repair the ECM on these bikes.

 

[STumped]

Ok, I just chose my answers on the poll and I chose, "live with it" but that isn't wholly accurate. My particular situation resolved on its own. Not sure why, or even exactly when.

So, once upon a time my 2004 non-ABS lit up the dreaded FI Code 26. And it was just as described, it only lit up if RPMs were over 4k for more than just a brief time (which equates to 70mph, so extended freeway rides always lit up the FI light).

Then, one day I noticed that it didn't light up anymore. I am not exactly sure when it stopped lighting up but one day I noticed that it hadn't lit up in a looong time (kind of like the opposite of when you hear the fridge stop running...you notice that, but when did it start running? <shrug>).

This is what I know........


It threw the code for approximately a 1-2 year period.


Things that may, or may not, have contributed to it throwing the code and/or it not throwing code anymore:


• Prior to it throwing the code, it sat in my garage for a little over a year (I bought another bike and thought I would sell the ST....perish the thought! LOL).

• I moved across the country twice (for work) in a little over a year's time so, combined with the previously mentioned other bike, I didn't ride it often but somewhere in there was when I first noticed the dreaded code.

• When I bought the bike, I originally bought premium gas for the ST but then changed to regular unleaded, per the various discussion on the forum about how premium gas was "recommended" but not "required" in the owner's manual.

* At some point (and no idea when that was, really -- I didn't keep notes on this) I decided to go back to using premium gas since that was recommended and it didn't cost that much more than regular (pre-COVID, of course).

•Sometime after making the second move across the country, I started riding the ST more often (sold the other bike). It was somewhere around that time that I noticed that the bike wasn't throwing the code anymore.


Soooo, what "fixed" it? I have no idea but it hasn't thrown a code since. <shrug>


My $0.02

 

[ASPC]

Okay, that does make sense. It is not "looking" for a knock prior to 4000rpm, and when it does if it is not getting the correct voltage signal it throws a code.

Automotive (once again, I am putting my foot in my mouth by comparing), when a knock sensor fault is detected it will retard timing as much as possible and I believe dump more fuel on some cars. This is all from the ECU assuming there is a constant knock, and it is desperately trying to save your engine from pre-ignition and burning up. It is not a "limp" mode as I mentioned before, but it is a fault mode and it will not self clear.

Some vehicles with variable cam timing will not allow the lobes to slide on the cams when it is in this mode, limiting rpm/power, which sounds similar to what is happening on ours.

There has to be an internal short on the ECU board, now is that a component (resistor, capacitor or ???) Or is it simply just a circuit path/pin that needs to be re-soldered? I guess the hardest part of checking these out would be trying to get the sealant/encased resin stuff off without damaging the board and components.

I think it’s just a voltage issue caused by a faulty component. ECM’s are relatively easy to repair and many are made to be accessible for repair. The silicon based potting material in these Honda ECM’s makes it almost impossible, and I wonder if that is why it is used because it’s not really required.

 

[ASPC]

I was just thinking about that. They called it the great capacitor plague and it went from 1999 to 2007. Reportedly it cost Dell $300 million dollars. they said the majority of the capacitors were from Taiwan but some were also from Japan. If that were the case with these ECU's, it would be evident after removing the potting material. The capacitors would bulge and break open from faulty electrolytes. I built around 20 computers during that time and I had to RMA probably five motherboards that had the problem.interesting read:

https://en.m.wikipedia.org/wiki/Capacitor_plague

Wonder where Honda got their caps from ???