Peterbilt 567 Aftertreatment Issues

2015 Peterbilt 567 with ISX15. Truck came back one day saying its DPF was full. I was able to regen it and run it one more day while I was looking for options on a DPF and DOC. Then it came back with a derate code of 2976 DEF dosing unit temp - data erratic. We checked harness and volts. Everything ohmed out and took a load test successfully, and volts were within spec. I put on a new DEF dosing unit and codes 3559 DEF dosing unit supply low and 3558 DEF dosing unit supply high popped on. I checked voltage and everything was within spec. I put the original DEF dosing unit back on and then it would go between 3558 and 3559 being active separately and together randomly. We ended up taking out all the harnesses and inspecting them and didn't find anything. Now when we initially check codes with engine off, the following codes are active: 1921, 3559, 2976, 3135, 1879, 3133. When we start it and it runs for a minute, only 1921, 3135 and 2976 come active. I hate to say the ECM internally failed without a better diagnosis first. What else can I check?

Grounds?

still wouldn’t condemn the ECM yet. The main thing that gets my attention is the 2976 code for DEF dosing unit temperature erratic. Since it stayed there even after trying another dosing unit, I’d be looking hard at the circuit, connector pins, and harness before calling the ECM bad.





The 3558/3559 codes also make me think it could be a wiring or load issue. Voltage can look good with a meter, but the circuit can still fail under load. I’d load-test the wires, check pin drag/tension at the ECM and dosing unit connector, and watch the dosing unit temperature PID while doing a wiggle test, especially near the ECM and hot areas around the engine.





Also check ECM powers and grounds with the engine running, not just key on. Bad grounds or voltage drop can cause weird random aftertreatment codes.





If the harness passes under load, connectors are clean/tight, powers and grounds are good, calibration is current, and a known good dosing unit still gives the same 2976/3558/3559 behavior, then I’d start suspecting an internal ECM issue. But I’d want to prove all that first before replacing it.