Top ten KNX field problems and how to fix them

Patterns repeat. Once you have commissioned a few hundred devices, you start to recognize the same handful of issues. A short field guide.

1. Bus polarity reversed. Symptom: device powers up but does not respond. Fix: swap the two bus wires.

2. Power supply undersized. Symptom: random device dropouts under load. Fix: calculate total bus current draw and add 30 percent headroom; replace the PSU if needed.

3. Topology address conflict. Symptom: ETS shows an unreachable device. Fix: read individual addresses from each suspect device using the local programming button and reassign the duplicate.

4. Filter table set to forward all. Symptom: cross-line traffic floods the backbone. Fix: regenerate filter tables in ETS and reload the couplers.

5. Group address typo. Symptom: switch presses produce no response, the bus monitor shows the telegram on a slightly different address. Fix: correct in ETS, reload the misconfigured device.

6. Actuator response feedback missing. Symptom: visualization shows a stale state. Fix: verify the actuator's status group address is configured and that the actuator is set to send on change.

7. KNXnet/IP gateway unstable. Symptom: tunneling drops every few minutes. Fix: check IGMP snooping on the network switch and update the gateway firmware.

8. Long telegram queues. Symptom: a sluggish bus that lags by seconds. Fix: identify the chatty device with the bus monitor (sort by source address) and reduce its publishing rate.

9. Burnt-out coupler after a surge. Symptom: an entire line is unreachable, the coupler LED is dead. Fix: replace the coupler and add Class C surge protection if not already present.

10. Device stuck in programming mode. Symptom: a stubborn device that will not load. Fix: power-cycle the line, then attempt the load again with verbose ETS logging.