FI code 26

Don B

Don B

Joined
Oct 15, 2007
Messages
1,295
Age
82
Location
Alief, TX, USA
Bike
ST1300
OK, time for me to cry "Uncle". I have had the code 26 for a long time now, and have lived with it, since I seldom have occasion to ride over 70mph anyway. However, I would like to engage in something a little more spirited than 4000rpm from time to time. I had to remove the LH center cowling last week to fix a headlight issue, so now have a good opportunity to go after the code 26 problem. What I need to know is what connector and where it is in between the knock sensors and the ECM. I am assuming that the sensors must be connected to a sub harness that plugs into the one that goes to the ECM. If that isn't the case, I may just live with what I have, since taking the rear end of my bike apart to access the ECM will be a last resort for me. I suspect that if there is an open or short circuit in the knock sensor circuit, it will likely be in the sub harness that includes the sensors.
If not, I may dig a little deeper once I get my pony (Mustang) back on the street again.
 
I checked my Honda SM and took a look at the LH cylinder head to ID the knock sensor, and think that I have located the connector from the subharness (https://www.partzilla.com/catalog/honda/motorcycle/2005/st1300a-a-st1300a/wire-harness-st1300a} to the main harness. Since you have replaced yours, I hope that you can verify that it is the grey 10P connector just behind the RH cylinder head. I agree that it wouldn't be a quick fix to replace that harness, since it also has connectors to the injectors and I assume the O2 sensors, but it would be much easier to check the insulation and continuity through that harness first than all the way to the ECU, which I would probably never do such it appears to be a real PIA. Since it is not unusual for the SUB-HARNESS, INJECTOR 32103-MCS-770* to fail, I'm surprised that the official Honda SM doesn't include that check in it's troubleshooting steps. Removing the RH middle cowling will be a piece of cake for me at this point, much easier than accessing the ECU. I can't remember the last time that I've seen the red FI light after starting the bike, and if I ever exceed 4k rpm long enough to lite it up all I have to do is slow down below an indicated 75mph and turn off the red light and all is good again.

* https://www.partzilla.com/product/honda/32103-MCS-770?ref=1ae0274c94b87c48b75c96d7c50b50e70445c094
 
Last edited:
bayoubear said:
mine did the same {2003 just like yours} turned out to be the knock sensor wiring harness.
Click to expand...

I am assuming that the ENGINE SUB-HARNESS 10P (GRAY) CONNECTOR shown below scanned from the SM is the one I want to disconnect to check the knock sensor circuit for an unwanted open or short to ground:
SM 1-45C 001.jpg
 
yes thats the connector . the usual culprit is the connector that plugs directly to the knock sensor . they just fall apart cause they can't take the heat . IMO this should have been a recall by Honda. The newer ones are made of different stuff.
 
I thought 26 was right side. Mine was coding 26 and it ended up being the ecu. Find one you can swap and see if it goes away before ripping wiring apart.


Sent from my iPhone using Tapatalk
 
I have been searching for the past 10-15 minutes to no avail. but have read earlier (a long time ago) that codes 25 and 26 are reversed in the SM. Doesn't really matter in my case, if I remove the RH middle cowling to gain access to the 10P connector to the FI sub-harness I can check both knock sensor circuits for proper insulation and continuity. MarkFeit may be able to direct us to the reference that I just made.
 
Andrew Shadow said:
I don't find that the links are sufficiently evident with this new software. They are not as obvious as with the old software.
Click to expand...
I adjusted the software to add an underline when a non-http link is used like you did with the word 'here' so hopefully that will help some.
 
Andrew Shadow said:
Don:
I gave you the link to that answer in post #8 above back on Monday. I only entered the word "Here" however. I guess in retrospect I should have entered more text to make it more visible. I don't find that the links are sufficiently evident with this new software. They are not as obvious as with the old software.
Click to expand...

I opened the above referenced link, but failed to see the Service Manual Corrections from stwiki.
 
There were some posts a few years back related to knock sensor problems (codes 25 & 26} that included measured voltages at the disconnected sensors at different RPMs, and some images of oscilloscope waveforms IIRC. I haven't been able to find any of these so assume they have been "purged" from the forum. Has anyone saved these by any chance who would be willing to share? I have a DMM that I use from time to time, but it isn't a sophisticated high-tech meter like many of the Flukes, etc. However, my next-door neighbor is a master mechanic with 35+ years experience under his belt and a man cave full of all kinds of tools including diagnostic equipment, and he has offered to help me trace down the problem. I'm guessing that a logical place to start would be to simply swap the LH and RH sensors to see if the code changes, unless getting those buggers out of the heads is a real bear of a task.
 
Yes, that’s is the plug. I have code 26, and with the right knock sensor unplugged, a short to ground.

I unplugged that connector and the short to ground is not between the plug and the knock sensor.
So, I removed the tail section and unplugged the ecm. The short to ground is somewhere between the ecm connector and the above plug.
Does anyone know where the screened cable for the knock sensor is connected to ground as I feel that is where the fault is.
I have a length of 28swg screened cable I am going to use to bypass the wiring loom but would rather find where the centre core is shorted out.
 
I had the FI code 25/26 on my bike when I bought it. I verified the wiring for both knock sensors was good to the ECM and that the knock sensors themselves were good. I wound up locating a used ECM, replaced it and no more problems. That was 1 1/2 years and 10K miles ago. Follow the procedures in the service manual to verify the wiring and the knock sensors with a DVM before you start tearing the wiring loom apart.
 
IvorS said:
Yes, that’s is the plug. I have code 26, and with the right knock sensor unplugged, a short to ground.

I unplugged that connector and the short to ground is not between the plug and the knock sensor.
So, I removed the tail section and unplugged the ecm. The short to ground is somewhere between the ecm connector and the above plug.
Does anyone know where the screened cable for the knock sensor is connected to ground as I feel that is where the fault is.
I have a length of 28swg screened cable I am going to use to bypass the wiring loom but would rather find where the centre core is shorted out.
Click to expand...

You can go through all of that effort if you'd like, but doing some research on the forum will show that in the vast majority of cases, replacing the ECU ends up making the issue go away. FYI.
 
Wayhay!
Running a new shielded cable to bypass the “shorted to ground” loom cable has fixed it!

I ALWAYS got the FI code 26 if I ventured over 4,200 rpm for several seconds. I have read all of the forums posts about the ECM and how it always leads to replacement ECM. With this in mind, for a few months I have been watching ebay for an ECM with HISS and key, and the first one to pop up I purchased - a couple if weeks back.

So, this weekend was time to pull the fairings off and trace the knock sensor wiring back to the ECM which also involves removing the top-box plate and tail section.

I tested the wiring, left side good, continuity of central core with no short to ground. Then I tested the right side and sure enough, short to ground with short to ground at the ecm connector too.

So, next I ran a new shielded cable from the ecm pin (black cable) to the right knock sensor. I connected the shield at the ecm end to the ecm/abs ecm metal casing. Checked continuity of both shield and core and no short between.

Tidied up, plugged ecm back in, started bike, let it warm up, then held throttle at around 6,000 rpm for about a minute or so. No code shown .

So, the ecm was telling me “right knock sensor fault” and it was! It was shorted to ground!

I’ll keep the hiss/ecm/key I bought a couple of weeks before putting it back on ebay just in case.

I’m a happy bunny
 
You must log in or register to reply here.
Back
Top Bottom