GM/Cat J1587 Comm loss

Ask for and share advice on using the PicoScope kit to fix Heavy Duty and Off Highway machinery here
Post Reply
SafetyLighting
Newbie
Posts: 4
Joined: Thu Nov 17, 2022 10:47 pm

GM/Cat J1587 Comm loss

Post by SafetyLighting »

Hello,
I am currently attempting to troubleshoot a sudden loss of communication between transmission control unit, engine control unit, and possibly the throttle position sensor. I have begun to decode the J1587 using PS7 and the 4425a. It appears to only happen at WOT. I am about to road test the unit while capturing. My first of probably what will be many questions, is what is the best way to monitor for the communication loss? I have set up PS7 according to the J1587 thread here, and it appears to be decoding. In case it matters, I'm using channels C & D.

Thank you,

SafetyLighting
Newbie
Posts: 4
Joined: Thu Nov 17, 2022 10:47 pm

Re: GM/Cat J1587 Comm loss

Post by SafetyLighting »

So after driving up and down a local highway for an hour, I've found the issue occurs after WOT/top speed(66 mph) after approx 5 to 10 minutes. First symptom is the speedometer loses signal and drops to zero and the vehicle begins to decelerate. My laptop was running PS7, but I also had a second laptop running diag software which allows me to see parameters with a graph. I had several items being graphed, fuel, pedal position, etc. So when the speedometer loses signal, all the items on the graph continue to read signal, but appear to be locked into the last data they received. So even though the engine shuts down, the fuel pressure is still at 2000 PSI, and the pedal is still showing as activated. So all the signals become flat-lined, but not at zero.

ben.martins
Pico Staff Member
Pico Staff Member
Posts: 539
Joined: Tue May 16, 2017 1:02 pm

Re: GM/Cat J1587 Comm loss

Post by ben.martins »

Hi SafetyLighting,

Sorry for the delay in getting back to you.

Sounds like an interesting fault! I take it that because the data just flatlines we don't have any fault codes?

To help catch the fault when it occurs we could take a signal such as crankshaft to give us the engine speed. Using am advanced trigger we can trigger the scope to only start capturing when the engine speed starts to drop. Having the CAN on at the same time means we can decode and see what happened. As you have an extra channel still left over you could look at the throttle signal as well.

Happy to help with the decoding as well if you wanted to post the file here.

Kind regards

Ben

SafetyLighting
Newbie
Posts: 4
Joined: Thu Nov 17, 2022 10:47 pm

Re: GM/Cat J1587 Comm loss

Post by SafetyLighting »

Hi Ben,

You are correct, no fault codes.

I have what appears to be valid data, but am now trying to decipher it. Being new to this, I'm also not positive I've captured it correctly. Please take a look when possible. Tried to upload the file, but it's too large.

https://www.dropbox.com/s/l16o5mnb95kzg ... sdata?dl=0

-John

ben.martins
Pico Staff Member
Pico Staff Member
Posts: 539
Joined: Tue May 16, 2017 1:02 pm

Re: GM/Cat J1587 Comm loss

Post by ben.martins »

Hi John,

Just want to let you know that I am looking into your file you sent. There is a lot to pick through but we can decode the data for certain which is great and yes you have set it up correctly. It's just a little more difficult to piece it back together compared with J1939 or J2534 as we don't have a dedicated decoder.

I mentioned the use of an advanced trigger which we can set the scope to start capture when say the engine speed starts to drop. Let me know if this is something you would like to try.

We will get there I'm sure but it's going to take a little time.

Kind regards

Ben

SafetyLighting
Newbie
Posts: 4
Joined: Thu Nov 17, 2022 10:47 pm

Re: GM/Cat J1587 Comm loss

Post by SafetyLighting »

Well, I found the problem. I suspected the ignition feed circuit breaker might be shutting off, but it was in fact the battery feed to the ECM was being cut off. The ECM on this engine, and most CAT engines I guess, has two battery feeds and one ignition feed. Both battery feeds are on the same circuit. This being an older truck, the feeds are through circuit breakers rather than fuses. I back-pinned into the battery feeds and the ignition and took the truck for a ride. The circuit breaker for the battery feeds was opening up and causing the ECM to go offline with no chance for DTC's to register. I swapped the 20 amp circuit breaker to a 20 amp fuse, road tested and the problem is resolved.

The 12 volt signals are really dirty and I don't know if that is my settings, or just that truck.

https://www.dropbox.com/s/942a6xewdapdd ... sdata?dl=0

Seeing the circuit breaker operate on the scope is interesting.

ben.martins
Pico Staff Member
Pico Staff Member
Posts: 539
Joined: Tue May 16, 2017 1:02 pm

Re: GM/Cat J1587 Comm loss

Post by ben.martins »

Hi John,

Good to hear you found the issue. Quite common for ECM's to have at least 2 battery feeds with sometimes 3 and an ignition feed as well. You've captured it very nicely on buffer 8.
GMC circuit breaker.png
Not the greatest idea putting both supplies on the same circuit breaker as you'll end up in this situation!

The 12V signals you're seeing where the signal is fluctuating signals that something is either pulling the voltage down at a constant frequency. As the circuit breaker is reacting to an over current it might have been nice to have seen the current as well to see if it was on the ECU side which was pulling the current over the limit.

I do find that a good ground is somewhat tricky to reference too on the bigger machinery. As the 4425A has floating inputs we can use those to our advantage and use electrical grounds used for the engine ECU or for a sensor. One example for this use case is when looking at the rail pressure sensor.

If you pushed your time base to 500ms/div you would have been able to watch signal coming into the scope and graphed from left to right of the buffer. As you have signals where you looking for a drop out a mask could also have been used to help 'automate' the software whilst on a road test. I've added the mask around channel A and you can see where the issue happens in the mask and on the waveform navigator.
Mask on channel A.png
Hope this helps and thanks for sharing.

Kind regards

Ben

Post Reply