FI Code 26

MidLife said:

If falling back to open-loop mode, power and mileage drop should be noticeable, right?

Click to expand...

Probably not. If conditions are right for it, the fuel will explode early and there won't be a thing the ECM can do about it. Otherwise, the engine should run like it normally does.

There are multiple loops in the system; the one you're probably used to hearing about is the one with the oxygen sensors. Any other sensor that isn't critical to making the engine run also has a set of pre-programmed defaults that get used if there's a failure.

The knock sensors are different creatures. Because of they way they work, the ECM doesn't have an on-demand way to detect that they've failed and has to rely on inference. Above a certain speed, the engine makes enough noise that the sensors report that they're hearing something, which is one way to detect a failure. Depending on the design, pinging in one of the left cylinders might be audible by the right sensor, in which case the ECM could infer that one isn't working. Again, that's an educated guess. With no working sensors (and the ECM maybe not knowing it yet), the best it can do is time the ignition at whatever was determined to be ideal during R&D. If you ping, you ping, and like in the bad old days of fixed ignition timing, there's nothing you can do.

Lately MIL hasn't blinked 25 on start up anymore. It goes off normally on restart and doesn't come back on until sustained highway speed.

Click to expand...

The magic number for that is 4,000+ RPM sustained for more then 10 seconds. At any rate, the two usual culprits are misrouted-and-cooked wiring, which was common before Honda figured out they had a problem. There have also been some failed sensors, and the easy way to check for that is to swap them and see if the code changes.

Some 2003s report the wrong sensor, or maybe they report the correct one and the manual was wrong. We never did run that one down solidly.

--Mark

Blrfl said:

The magic number for that is 4,000+ RPM sustained for more then 10 seconds.

Click to expand...

Last time MIL came back, it was after over one hour of 4000+.

Blrfl said:

.....the two usual culprits are misrouted-and-cooked wiring, which was common before Honda figured out they had a problem. There have also been some failed sensors, and the easy way to check for that is to swap them and see if the code changes.

Some 2003s report the wrong sensor, or maybe they report the correct one and the manual was wrong. We never did run that one down solidly.

--Mark

Click to expand...

Thanks. Will be good starting point.....after fixing brakes and windshield!

Blrfl said:

......If you ping, you ping, and like in the bad old days of fixed ignition timing, there's nothing you can do.

Click to expand...

In the bad old days....ping is what we were going by to adjust timing!

Thanks again.

Good thread, with lots of good information. I'm experiencing the same issue with my 2003 ST; it has about 112,000 miles on the odometer. For the past month or so I've gotten the FI light when above 4k rpm for ~10 seconds: really crimps my riding style. Nothing over 70 mph, which rules out the freeway.

I found the same issues on my bike: wiring color mix-up and the opposite side FI error code (left code/right failure). While inspecting the right side knock sensor, the plastic terminal and wire insulation crumbled in my fingers: problem found. It shouldn't be too hard to reinsulate the wire and reconnect to the knock sensor but I have two questions:

1. The signal wire seemed to have a braided shield around it up to the last few inches. Is this crucial for sensor operation? Is it a ground path that has to be intact or can I cut the shielding short and just heat shrink the signal wire? Also, any tips about heat shielding the wire so this doesn't happen again?

2. I took left side knock sensor signal line voltage readings just prior to the ECM. I measured ~10 mVAC when off, ~21 mVAC at idle, and 50+ mVAC when above 4000 rpm. All reading in respect to ground. Are these reasonable?

Again, thanks for the help with this. Nothing I like better on New Year's Day than fixing a broken motorcycle.

PS I hope this solves the problem; I really don't want to drop $700 on a new ECM.

David P said:

...really crimps my riding style. Nothing over 70 mph, which rules out the freeway.

Click to expand...

As long as the weather's not excessively hot, you're not using low-octane fuel and not asking the engine to make torque where it doesn't, riding at high speeds shouldn't be a problem. The FI light doesn't mean you have to stop riding. There are a few conditions where it does, and the ECM will let you know one of them has happened by stopping the engine.

1. The signal wire seemed to have a braided shield around it up to the last few inches. Is this crucial for sensor operation? Is it a ground path that has to be intact or can I cut the shielding short and just heat shrink the signal wire? Also, any tips about heat shielding the wire so this doesn't happen again?

Click to expand...

It's heat shielding. The sensor itself is grounded to the engine block. There's no reason you couldn't add some additional shielding of your own. There's a certain way the wire on one side of the bike has to be routed to prevent heat damage, and I can't remember off the top of my head which side it is. There were a number of early bikes where the wire wasn't routed properly during assembly, although yours probably isn't one of them if it got to 112K before failing. After almost 14 years and that many miles, you could just have cruddy connections, too. I'd explore that before a failed ECM.

2. I took left side knock sensor signal line voltage readings just prior to the ECM. I measured ~10 mVAC when off, ~21 mVAC at idle, and 50+ mVAC when above 4000 rpm. All reading in respect to ground. Are these reasonable?

Click to expand...

Honda doesn't spec anything for that, but that at least says the sensor and the wiring is good. Since you've got everything apart, I'd repeat that experiment with the right side for comparison and to make sure you're not barking up the wrong tree. Things were a little odd when we discovered the mistake in the manual, and to this day I'm not positive which editions of the manual are wrong.

--Mark

I reinsulated the right side knock sensor wiring and plugged it back into the sensor. Measured both left and right knock sensor signals immediately before they entered the ECM: AC voltage levels pretty much matched from idle to 5k rpm. This make me believe that the sensors are OK. Also checked the left side engine harness plug, it looked OK. Bike still exhibits the same problem above 4k rpm.
Is there any way to check the ECM or is this just a cross-your-fingers replacement? It would irk the hell out of me to plunk down $700 and not find the problem.

FYI, here's my voltage readings:
Left side
Off - 48 mVDC, 2 mVAC
Idle - 14 mVDC, 8 mVAC
2k - 14 mVDC, 11 mVAC
4k - 20 mVDC, 32 mVAC

Right side
Off - 50 mVDC, 2 mVAC
Idle - 16 mVDC, 7 mVAC
2k - 10 mVDC, 8 mVAC
4k - 12 mVDC, 25 mVAC

David P said:

Measured both left and right knock sensor signals immediately before they entered the ECM: AC voltage levels pretty much matched from idle to 5k rpm. This make me believe that the sensors are OK.

Click to expand...

The engine-off voltages probably don't mean much, but the fact that the AC measurements pretty much track each other lead me to think that's a good conclusion.

I put a copy of your measurements on on STWiki for posterity.

Is there any way to check the ECM or is this just a cross-your-fingers replacement?

Click to expand...

Any fault is going to be on the board inside the ECM, and it's already told you all it's going to tell you. Based on the number of code 25/26 failures where the ECM has been at fault, it's a decent bet that yours

You could probably find one at a decent price through one of the used parts networks or on eBay. If you order one, ask them to send you a photo that shows the connectors and the Keihin label. The right unit will have two connectors and the part number 38770-MCS-L01 printed on the label. Any other part number is for a European, police or 2008-and-later bike and isn't going to work. (I've been tempted to snag one as a tester that I could loan out but haven't yet.)

There are repair houses, but I have no idea if any of them do motorcycle ECMs and don't figure the price to fix yours would be less than that of a used, working unit.

--Mark

I purchased and installed a new ECM that I got from Partzilla - cost ~$700. Bike works fine now: no issues at 4,000 rpm anymore. Thanks for all of the legwork that's been done by those who dealt with this before me.

David P said:

I purchased and installed a new ECM that I got from Partzilla - cost ~$700. Bike works fine now: no issues at 4,000 rpm anymore. Thanks for all of the legwork that's been done by those who dealt with this before me.

Click to expand...

That is good you repaired it, stinks that they cost so much but that is the price or technology.

I'm coming round to suspecting a bad ECM again in my case of 4200 RPM misfiring. I would be extremely interested in more extensively trying out an alternate ECM before plunking down the $$$ for one.


Sent from my iPhone using Tapatalk

Don B said:

I haven't yet determined positively that the ECM is my problem, and don't want to buy a replacement "just in case".

Click to expand...

Don B,

Couldn't find the rest of your story in this thread, on chasing after your code 25/26 issue. Did you post it somewhere else? Did you end up having to replace the ECM?

I'm trying to help out a local rider that just purchased a 2004 ST1300 with this same issue. I know there are 2 different ECUs, one for 2003 -2007 (38770-MCS-L01) and one for 2008 -2012 (38770-MCS-R12). The last 3 digits are different and I thought this only represented a different manufacturer or some other update. Does anyone know if these are interchangeable? The later one is about $200 less MSRP and I thought I remember reading that the mapping was improved on the later one.

Ashley said:

Does anyone know if these are interchangeable?

Click to expand...

They are not.

The way to tell in general is to look up the part in the parts fiche. If the old part says it was superseded by the new part, you can use the new part. The other thing you can do is look up the new part by number, find out where it's used and see of the bike is in that list. (Many places that sell parts online have this feature.)

You can learn about what the part codes mean here: CLICKY.

--Mark

Blrfl said:

They are not.

The way to tell in general is to look up the part in the parts fiche. If the old part says it was superseded by the new part, you can use the new part. The other thing you can do is look up the new part by number, find out where it's used and see of the bike is in that list. (Many places that sell parts online have this feature.)

You can learn about what the part codes mean here: CLICKY.

--Mark

Click to expand...

Didn't somebody also mention that the 2003 harness won't plug in the later ECU?

eBay/gmilne said:

eBay Post

ECU from a running US 2004 ST1300A.

I purchased a new ECU in an attempt to clear a code 26. Unfortunately, the new ECU didn't help, but I left it in the bike and am selling the original. The bike ran perfectly except for the FI code, which I learned that many just live with it with no issues.

Click to expand...

Wondering if eBay gmilne is on this board. Had not heard of the "many just live with it" observation and was wondering how wide a poll he might have conducted!



.

MidLife said:

Don B,

Couldn't find the rest of your story in this thread, on chasing after your code 25/26 issue. Did you post it somewhere else? Did you end up having to replace the ECM?

Click to expand...

Below is a copy of a PM that I recently sent to you:

I haven't had the tupperware off of my ST since the problem first cropped up (shame on me), so I haven't yet been able to do any troubleshooting to determine the cause of the failure. I still don't hesitate to ride the bike anywhere I want, I just try to keep it under 70mph and ride on. As you well know, keeping the bike at or below 70 isn't easy, but I don't get concerned about it if I lite up the FI since I'll notice a slight drop in RPMs when the code is set, so I know what it is and that no harm is likely to occur from continuing my ride with the FI lite lit, just an annoyance. At any rate, I really do intend to take the time to remove the plastic before too much longer and do a little routine maintenance while I'm there. I plan to do the regular multimeter checks for insulation and continuity, and go from there. hopefully it will be a wiring problem that should be relatively easy to fix. Trying to check the wiring integrity all the way back to the ECU connector appears to be easier said than done because of the poor accessibility to the ECU, so I plan to check it back to a connector between the knock sensors and the ECU connector. It's been awhile since I looked at it in my Honda SM, but I think the connector I'm referring to is the 9-pin connector under the seat. If I find the wiring is all good, I guess I'll have to bite the bullet and replace the ECU, but if it comes to that I'll try taking my chances with one from e-bay or a salvage yard first. I really would hate to have to cough up the $ for a new one from Honda, would have to think twice before doing that.

Sorry I couldn't really answer your question since I haven't yet tamed my code 26 problem, but hopefully I may have brought up some ideas you hadn't considered. If you tackle yours, let me know how it turns out. I thought about trying a couple of the M/C salvage yards around if it comes to that, but it's been several years since I've been to one that I don't even know if they still exist or not. I used to go to one on Long Drive a short distance from the south loop 610 many years ago, but I think that place is long gone by now.

By the way, I remembered that I located at some time in the past an easier way to gain access to the ECU than the SM describes. Just follow the first 9 steps in the following instructions.

Since your ST doesn't have ABS, it may be that you can access the ECU from under the seat and not have to remove the tail section. I've never seen under the seat of a non-ABS ST, so I have no idea.

http://www.powercommander.com/downloads/120/install/411/eng120-411.pdf

Can anyone post a picture of the knock sensor wires up by the ECU? Or maybe the same wires go through a plug under the seat or tank somewhere? A picture and what color the two wires are would really help those of us just starting to trace the problem. Thank you.

Sent from my SAMSUNG-SM-G935A using Tapatalk

Don B said:

By the way, I remembered that I located at some time in the past an easier way to gain access to the ECU than the SM describes. Just follow the first 9 steps in the following instructions.

Since your ST doesn't have ABS, it may be that you can access the ECU from under the seat and not have to remove the tail section. I've never seen under the seat of a non-ABS ST, so I have no idea.

http://www.powercommander.com/downloads/120/install/411/eng120-411.pdf

Click to expand...

Thanks for the info.

Just had a look at the wire by the sensors on mine, it all looks very good, no melting, chafing or lack of slack. Nothing like some of the pictures I've seen.

As you say, my 03 is non ABS and I can see the ECU under the rack.

Has anyone taken a replace ECM and opened it up to see if there's any obvious electrical damage which might be corrected?

I may be mistaken, but I think I read at one time that the ECU is fully sealed in potting compound. Is that possible?

Mellow said:

Has anyone taken a replace ECM and opened it up to see if there's any obvious electrical damage which might be corrected?

Click to expand...

Looks like some have already looked into this. Check this thread and post.