CAN-Lambda Problem <Resolved>

[krohelm]

TL;DR:  Add a 22uF capacitor to your CAN-Lambda, as close as humanly possible to the device.

 

Nasty problem with my CAN-Lambda.  On startup, watching the runtime values "ECU Status" tab, Lambda 1 box:

• Before startup, it's OK/diagnostic/waiting for rpm.
• After startup it cycles.
  Calibration
  Heating
  Error
• After heating
  code 33  Open circuit APE-IPE
  code 16 Heated too long
• While heating, the temperature very rapidly rises to 500-600c (2-4 seconds depending on exhaust gas temperature) before the code 16.

What I've tried:

• Carefully inspected wiring harness on o2 sensors.  No issues.
• Carefully inspected wiring harness on CAN-Lambda.  No visible issues, but I couldn't disassemble the pins to verify the crimps on APE and IPE wires (lack special tool or understanding of connector).
• Replaced o2 sensor with a brand new one from Amazon.  Exactly the same behavior.
• Allowed the engine to reach operating temperature with mixed low and high idle.
• Verified 0 CAN 1 Receive Error Counter.
• Verified 0 CAN 1 Transmit Error Counter.
• Verified 0 ECU Fault Codes
• Verified operating battery voltage 14.2v

Log file is attached.  Observe parameters Lambda 1, Lambda 1 Error, Lambda 1 Status, Lambda 1 Temp to see the cycle.  It's the same with both LSU 4.9 sensors I've tried.

Pclr is attached for 4age 20v engine.  It's a work in progress.

 

Log 2017-02-16 9;36;52 pm.llg

4age 20v.pclr

Edited November 24, 2018 by krohelm
Solution added, re-marked as correctly resolved.

[krohelm]

By the way I'm in the United States and I need this thing working within a week or so (have dyno time scheduled shortly after, makes shipping back & forth with NZ challenging).  I'm willing to try lots of things.  I still can't figure out how to disassemble the wiring harness on the CAN-Lambda non-destructively, if someone knows how I'd love some tips...

I've got a spare DTM-4P connector I'm going to make into a test harness for the CAN-Lambda.  Any bright ideas as to what I should look for?  Voltage under load at the connector maybe, but what else?  Maybe a set of tests I can do with a DMM on the o2 sensor or the CAN-Lambda?

Edited February 18, 2017 by krohelm

[Ducie54]

Is it getting good voltage when power up?

[krohelm]

Made a test harness to check voltage under load.  It does not draw the feed line below 13.85 volts while the heating element is on at the point of connection to the CAN-Lambda.

Video of voltage cycling with load as the CAN-Lambda cycles the o2 sensor through calibration, heating and errors 33 & 16. https://goo.gl/photos/DijeiVstLmUe1QgB8

[Ducie54]

did you happen to chop off the original plug off the lambda sensor?

[krohelm]

No.  I installed the o2 sensor the CAN-Lambda kit came with as it came.  Then 2 days later installed another new o2 sensor from Amazon.  No modifications, just screwed them into the m18 boss that came with the CAN-Lambda.  

I had the boss welded into the original o2 sensor location; it's elevated 25-35 degrees above horizontal, a couple feet down from the header and before the flex pipe.

[Ducie54]

What firmware are you running? There maybe a fix for this if your using a old firmware

[krohelm]

5.6.5, I keep up to date like an addict.

[Ducie54]

Sorry but im out of ideas

[krohelm]

I really appreciate you taking the time to offer your thoughts :-).

[Dave Kriedeman]

Hi Krohelm,
is there any difference if you turn the wideband to ON / RUN WHEN STALLED so that the LSU 4.9 is heated before running the engine.
Just curious.
Regards
Dave.

[krohelm]

Hi Krohelm,
is there any difference if you turn the wideband to ON / RUN WHEN STALLED so that the LSU 4.9 is heated before running the engine.
Just curious.
Regards
Dave.

Thanks for the reply Dave!  I have not seen this option in PCLink, how can I set it this way?  I wanted to try this but didn't want to force the mode by leaving the CAN bus unplugged for >= 2 seconds after power up.

Actually, this is a CAN-Lambda setup.  I found the setting in "Help Browser" but it looks like that setting is only applicable to internally controlled lambda sensors - which Monsoons don't support.  :-/  I'll try to work out a pre-heat test.  There are a few features I wish I had, I wish I could upgrade this ECU!

Edited February 19, 2017 by krohelm
read more in help

[krohelm]

I hope this doesn't affect my warranty coverage, I'm just trying to debug my CAN-Lambda!  I observed a low o2 sensor voltage error one time just now while experimenting with powering the CAN-Lambda while the engine is off and no CAN connectivity back to the ECU for first 2+ seconds of power.  Since I've verified that the voltage at the connector I've provided is 13.9v while heating and the engine is running (and the o2 sensor requires >= 10.8v according to the spec sheet) maybe there's a problem with the CAN-Lambda Deutsch plug.  So I disassembled the pwr & gnd pins on the CAN-Lambda plug.  The pwr wire is not assembled totally correctly, so I'm going to re-solder it.

Notice that the shielding of the red wire goes into the Deutsch solder pin...  it's hard to see in the picture, zoom in if you can.  This is not the recommended means of assembly:  These connectors are technically supposed to be crimped (not sure what connector places have against solder, but apparently even these solid pins are supposed to be 100% crimp connectors) but be that as it may, wire jackets don't belong inside the solder pocket.  Black wire looks great, please don't take away my warranty coverage Link people :-(

 

Reference http://www.alliedelec.com/m/d/39051ed66cf0bac8360059bb809f47cb.pdf for the reference to the technical document 0425-205-0000, found at https://laddinc.com/products/deutsch-products/deutsch-information-drawings/ which lists crimp tool HDP-400.  Looks stupid expensive, I'll just make the red wire look like the black one.

[krohelm]

Serviced the power wire connection and checked the car.  Still does the same thing.  Power to the connection is well within spec, the CAN-Lambda connection is now visibly electrically sound.  Log file like the original post one attached...

Log 2017-02-18 7;09;13 pm.llg

[Ducie54]

I would be asking Simon as to why there soldered and not crimped. Crimping tools are not that expensive, the cheap ones worked will on my harnesses.

I can't understand why the power wire has the shield attached to the pin. If it's grounded in the CAN lambda module that will give you issues. 

I'll be pulling my LINK knock block apart to see if it's soldered now. 

[krohelm]

I would be asking Simon as to why there soldered and not crimped. Crimping tools are not that expensive, the cheap ones worked will on my harnesses.

I can't understand why the power wire has the shield attached to the pin. If it's grounded in the CAN lambda module that will give you issues. 

I'll be pulling my LINK knock block apart to see if it's soldered now. 

Yep, I'll be chatting with him in the morning.  I don't think I'll bother about solder vs. crimp on the connection, I'm more interested in getting it to work for the moment.

The power wire was just trimmed too short and pushed in.  I rectified it really quickly with my fx-888d.  If I really get feeling ambitious maybe sometime I'll replace these pins with crimped pins.  The CAN-Lambda module is 100% non-serviceable near as I can figure or I would definitely be checking each solder pad with a magnifying glass looking for any hint of a defect that might explain the operation.

Both of the brand new o2 sensors I've tried do heat up, but then code 33 or 34 happens.  It always happens right at about 520c after heating looking at the logs.

[Dave Kriedeman]

Hi all,

unless there is some unknown reason to me,

A shield wire should be encasing/surrounded or wrapped around signal wires normally.

The end of the shield should normally be terminated to sensor ground and the opposite end left non-terminated.

Soldered connections are less desirable due to possibility of dry or poor joints.

The other main reason is the heating of the wires during soldering can create an issue with the longevity of the wire cores and them becoming brittle, which through movement or vibration can or may cause complete breakage or partial disconnection leaving only a few cores to carry the signal, current what ever it may be.

This will increase resistance obviously.

This reply is a well known fact however i feel it is worthy to mention the reasons behind , for those whom may not know or understand.

So crimping is always the preferred method unless otherwise instructed with accompanying explanation.

Regards

Dave.

[krohelm]

Thanks for the reply and explanation, Dave.  I chatted with Simon and it appears that I may have gotten a pre-production CAN-Lambda from my dealer based on the connections being soldered.  I'm now in the process of having it replaced.  This is my first experience with Link support physically, I hope it works out as well as their Internet support which I've been really impressed with!

[krohelm]

Got the replacement CAN-Lambda.  Installed it, put in the new o2 sensor, same problem.  I drove the car around for a little over 10 minutes, no change in the condition.

Sometimes error code 33, sometimes error code 34 with no rhyme or reason for which one it is as it goes through its reboot cycle.

33:  Open circuit APE-IPE

34:  Open circuit RE-IPE 

• This is the 3rd brand new LSU4.9 I've used with the exact same behavior.  2 from New Zealand and 1 from Amazon in USA.
• Still, no errors with the CAN bus.
  I used high quality twisted, shielded, grounded wire terminated with a 120 ohm resistor after the tap for the CAN-Lambda - I didn't expect a problem there.
  PCLink shows all CAN checks OK and error counts 0.
• No problems supplying power.
  13.9v while it's heating.
  I jumped directly to the battery to rule out the possibility of low voltage that my Fluke was somehow missing.
• O2 sensor threads in freely by hand, no binding or twisting.
  It's mounted in full compliance with the CAN-Lambda manual.
• Wire connectors are not pulled or stretched - all connections exhibit proper relief.

 

Log file from drive attached.  What else can I try?

Log 2017-02-23 7;48;46 pm First drive no lambda.llg

Edited February 24, 2017 by krohelm
attached log file from drive & added detail

[Ducie54]

How have you wired it in? I didn't think u needed the resistor.

Is the ground wire large enough? Is it grounded to the chassis or battery?

 

Edited February 24, 2017 by Ducie54
Added more

[krohelm]

+12v and ground come from the old o2 sensor wires.  Power comes via the efi relay and the ground appears to be the engine head but I didn't trace it through the loom.  I jumpered directly to the battery to see if power supply was a problem and it made no difference, status rotation was exactly the same.

There's no voltage error from the CAN-Lambda and voltage does test fully in spec while it's heating...

Resistor is for the CAN bus termination. 120 ohms is needed at both ends of a high speed CAN bus line. Link ECU's have the one end covered internally, and since the CAN-Lambda doesn't have a termination resistor you have to provide one on the main bus wire after you tap it for the CAN-Lambda.

0 CAN bus errors and data is constantly flowing as you can see the CAN-Lambda temperature and status continually rotating in the logs.

[krohelm]

Anyone else have one of these that's working?  What PCB REV is yours?

[rrob]

Using the old 02 as a power supply is not ideal as the earth is usually ecu controlled. 

[Adamw]

Using the old 02 as a power supply is not ideal as the earth is usually ecu controlled. 

Although I agree with this comment in general, We think in this case we have already eliminated power supply issues during initial troubleshooting by powering it with jumper leads direct from the battery.

 

Anyone else have one of these that's working?  What PCB REV is yours?

Hi Kenny,  Unfortunately I dont think you are going to get the help you need from the forum in this case.  Your problem is one that we havent seen before and without many strong clues that we havent already eliminated so we are having a hard time understanding what is causing it.  I understand being on opposite side of the world with the awkward time zone differences will make it very frustrating for you but as I explained on Friday night, the quickest way you are going to get this sorted is by making contact with Tech support as early as you can in NZ Daytime hours.  That way when you call, tech support will have the whole engineering team beside them to work through it.  Unfortunately I think there is still probably a little more toing and froing to do yet.        

[krohelm]

Hi Adam, no worries, I just have to throw out the net as wide as possible at this point.  This now has me pouring back through the FSM for my 20v to see exactly where the ground is terminated - which is definitely not a bad exercise!  I'll be online tomorrow morning NZ to continue the troubleshooting.

Thanks for the note, rrob, will follow up if that is somehow the issue.  :-)

edit:  Yeah, it's definitely a proper ground.  Diagram is attached.  For reference, E1 on the ecu is "Engine ground."  The O2 sensor grounds to the engine head, I know the post it's in good condition.

BTW I did not use OX and HT for CAN bus  I ran a shielded twisted pair just for it.

Edited February 26, 2017 by krohelm
followed up on rrob's tip