Grounding and Shielding on a Building Automation Bus
You are installing a BACnet MS/TP control bus (RS-485) for a rooftop unit and VAV boxes in a commercial building. The bus runs about 600 ft total, daisy-chained through 12 devices, using shielded twisted pair. The electrician bonded the shield to earth ground at both ends and also at two intermediate VAV controllers. The bus is terminated with 120-ohm resistors at both physical ends, but a third 120-ohm resistor was left installed at a mid-bus controller. Devices intermittently drop offline, especially when the rooftop unit's VFD-driven fan starts. The controls contractor says 'just add more repeaters.'
1. Identify the wiring/termination errors on this RS-485 bus and explain how each causes intermittent communication failures.
Read the worked answer
Errors:
- Shield grounded at multiple points (both ends plus two midpoints). Multiple ground bonds create ground loops; differing ground potentials drive current on the shield, coupling noise onto the signal pair. This is a prime cause of intermittent dropouts.
- Third 120-ohm termination at a mid-bus device. RS-485 should be terminated only at the two physical ends. A mid-bus termination adds excess load and reflections, degrading the signal - classic cause of marginal, intermittent communication.
- VFD noise coupling. The rooftop VFD starting correlates with failures; with a poorly grounded/looped shield and wrong termination, the bus has little noise immunity, so VFD switching noise tips marginal nodes offline.
'Add more repeaters' treats symptoms; the physical-layer faults must be corrected first.
2. Describe the correct shield grounding and termination practice for this bus, and how you would verify it.
Read the worked answer
Correct practice:
- Termination: 120-ohm resistors at only the two physical ends of the bus; remove the mid-bus resistor. Confirm the two ends are the actual cable extremities.
- Shield: bond the shield to ground at one end only (typically the controller/head-end end); leave the other end and intermediate devices' shields unbonded (splice the shield through, do not ground at each device). This avoids ground loops while still draining interference.
- Topology: maintain a true daisy chain, no stars or long stubs; keep device drops short.
- Noise: route the bus away from VFD/motor feeders where practical; ensure the VFD and its cable are properly grounded per the drive manufacturer.
Verification:
- Continuity check that the shield is grounded at one point only.
- Measure resistance across the pair at an end with power off to confirm termination value is as expected (not doubled/tripled).
- Monitor node online status through a VFD start/stop cycle to confirm dropouts stop.
- Check bus voltage levels and error counters on the controller.
3. The controls contractor insists repeaters are the fix. How do you handle that professionally, and what would justify a repeater?
Read the worked answer
Respond with evidence, not opinion:
- Correct the physical layer first - single-point shield ground, remove the extra termination, verify daisy-chain topology. Document before/after node status through VFD cycles.
- Show the data: if dropouts stop after fixing grounding/termination, the repeater was unnecessary. Present the monitoring logs.
- Explain when a repeater IS justified: RS-485 segment length/device-count limits being genuinely exceeded (e.g., beyond the standard segment length or node count for the baud rate), or a required topology change - not to compensate for ground loops or mis-termination.
Professional handling: put the finding in writing, fix the identified faults, and only add a repeater if segment limits are truly exceeded after the physical layer is correct. Adding hardware to mask a wiring fault wastes money and leaves a latent reliability problem.