Anyone have experience troubleshooting intermittent faults on a PLC-controlled robot?
We've got a robot that will run fine for hours and then randomly stop with a communication fault. Reset it, and everything runs normally again.
We've checked the cables and connections, but nothing obvious stands out. The frustrating part is that we can't seem to reproduce the issue when we're actually trying to troubleshoot it.
Intermittent comm faults that clear on reset are usually the physical layer or network load, and the trick is to capture the state at the moment it drops instead of troubleshooting afterward. A checklist that usually narrows it down:
1. Record the exact fault code and timestamp on both the robot controller and the PLC. A connection timeout and a lost link point to different causes. 2. Read the port diagnostics on the managed switch. CRC or alignment errors, or a climbing link up/down count on the robot's port, point to cable or connector, especially in the dress pack where the cable flexes with the arm. Check that the cable through the arm is flex-rated. 3. Correlate fault times with other events: drive or servo starts, welding, a large contactor, shift change. If they line up, check shield termination and routing next to motor leads. 4. If it is EtherNet/IP, check the RPI and timeout multiplier against network load. A broadcast storm or a laptop plugged into the wrong port can push a connection past its timeout. 5. Rule out a duplicate IP address, which gives exactly this random, clears-on-reset pattern.
If you can, mirror the port and leave a packet capture running across the next fault.
1. Record the exact fault code and timestamp on both the robot controller and the PLC. A connection timeout and a lost link point to different causes.
2. Read the port diagnostics on the managed switch. CRC or alignment errors, or a climbing link up/down count on the robot's port, point to cable or connector, especially in the dress pack where the cable flexes with the arm. Check that the cable through the arm is flex-rated.
3. Correlate fault times with other events: drive or servo starts, welding, a large contactor, shift change. If they line up, check shield termination and routing next to motor leads.
4. If it is EtherNet/IP, check the RPI and timeout multiplier against network load. A broadcast storm or a laptop plugged into the wrong port can push a connection past its timeout.
5. Rule out a duplicate IP address, which gives exactly this random, clears-on-reset pattern.
If you can, mirror the port and leave a packet capture running across the next fault.