Industrial Ethernet Troubleshooting Guide

When a device drops off the network, work from the physical layer upward: link lights, cable, addressing, switch and device. This guide gives an order that finds most problems fast.

Be careful on a production network. Scanning tools and heavy testing can disrupt older devices and running processes. Check with whoever owns the network, make small changes one at a time and know how to undo each one.

Match the symptom

Symptom	Likely causes
One device never connects	Cable or connector, wrong IP address or subnet mask, duplicate IP address, device not powered or faulted
Device connects, then drops now and then	Damaged or loose cable, a poor connector, electrical noise, speed or duplex mismatch, a switch port issue
Many devices lose communication at once	A switch or uplink problem, a network loop, a broadcast storm, power loss to a switch
Everything is slow	Too much traffic, a loop or storm, a device flooding the network, an overloaded controller
Works on the bench, not on site	Different IP range, a different VLAN, a firewall or a managed switch setting

Work through it in this order

Link lights. Look at the port lights on the device and the switch. No link light means a cable, connector, port or power problem. Fix that first.
Cable and connectors. Swap in a known good patch cable. Reseat connectors and look for bent pins, damaged boots or a cable pinched in a door. Copper Ethernet runs are limited to 100 m, and industrial cables and connectors should suit the environment.
IP address and subnet. Two devices can talk directly only when their addresses are in the same subnet. Check each device's address and mask, and use the subnet calculator to confirm they match.
Duplicate addresses. Two devices with the same IP address cause intermittent and confusing faults. Unplug the suspect device and ping its address. If it still answers, something else is using it.
Ping and ARP. From a laptop on the same network, ping the device's address. Then arp -a shows the MAC address that answered, which helps identify a duplicate or a wrong device.
Switch ports. On a managed switch, check the port's status, speed, duplex and error counters. A port with many errors points to a cable or a mismatch. Mismatched speed or duplex settings (one end forced and the other on auto) cause slow and unreliable links.
Loops and storms. A cable plugged into two ports of the same network can flood it. If many devices fail together and the switch lights blink furiously, look for a loop.
Noise. Look at where the cable runs. Cables beside motor or drive power wiring pick up noise. Use shielded cable where specified, ground the shield as the manufacturer directs and keep a separation from power cables.
Device settings. Check firmware versions, the update rate or packet interval and the number of connections the device supports. Update rates set too fast for a slow device or a busy controller cause timeouts.

Isolate

If you can't find it, shrink the problem. Connect a laptop directly to the device with a short cable and set a fixed address in the same subnet. If it works there, the problem is in the cable run, the switch or the network settings between. If it doesn't, the device itself or its configuration is at fault.

Write it down

Keep a record of IP addresses, subnet masks and where each device plugs into which switch port. Most network problems on plant floors are made worse by not knowing what should be where.

This is general guidance for training and reference. Network settings and diagnostics differ between devices; follow the manufacturers' documentation and your site's IT and OT policies.