Playbook: VFD and Pump Faults

Adapted from HVAC FDD Reference v1.0, Remediation Playbooks (pp. 167–168).

Pump family index: faults/pmp.

Decision tree

Follow control authority before component replacement:

1. Confirm remote/auto/bypass state.
2. Compare speed command and feedback.
3. Determine whether the loop is pinned at minimum or maximum and read the process-error direction.
4. Check current, torque, safety, demand, and application limits plus mechanical restrictions.
5. Review tuning, process-sensor quality, and setpoint-reset behavior.
6. Inspect current/kW trends only after control-state findings are resolved.

Do not manually energize, transfer, or bypass a drive outside the site’s approved operating and electrical-safety procedure.

Step 1 — Confirm control authority (VFD-0005)

1. Verify the final drive enable is actually commanded, not merely an upstream system request.
2. Read the HOA/keypad selector and configured command source. “Auto” must mean the commissioned remote BAS source, not local PID auto or terminal control.
3. Confirm bypass from authoritative drive mode or contactor/power-path proof. Bypass available, ready, or commanded is not active bypass.
4. Check whether maintenance, emergency, fire/smoke, commissioning, or a functional test explains the state. If approved, suppress the diagnostic rather than defeating the safety sequence.
5. Restore authority only through approved procedures, then allow the loop and setpoint to settle before evaluating downstream faults.

Step 2 — Compare speed command and feedback (VFD-0001)

1. Verify command and feedback are both normalized to percent of rated speed; do not compare percent with Hz.
2. Check command-source configuration and fieldbus communication.
3. Check the drive display for current, torque, thermal derating, and fault codes such as overcurrent, overvoltage, or ground fault.
4. Check the cooling fan and heatsink — overheating can cause quiet derating before the drive trips.
5. Compare drive-reported speed with a tachometer only after confirming what each point measures.

Step 3 — Inspect loop saturation (VFD-0002, VFD-0003)

1. Confirm process value and setpoint are from the same loop, share units, and use a commissioned site-specific error threshold.
2. At minimum speed, use the error direction to distinguish overdelivery/a minimum set too high from underdelivery, obstruction, or bad sensing.
3. At maximum speed, verify the configured maximum and inspect current, torque, safety, demand, and application limits before calling the equipment undersized.
4. Inspect filters, coils, strainers, dampers, valves, belts, couplings, impellers, and the distribution path for restrictions or degradation.
5. Confirm the setpoint was settled long enough for the application’s time constant; a real reset/load step is not a capacity fault.

Step 4 — Inspect hunting (VFD-0004)

1. Acquire speed, process value, and setpoint together at 60 seconds or faster; resample faster acquisition to the rule’s legal fixed evaluator interval (14.3–150 seconds at the defaults, 60 seconds recommended). Change-of-value logs can hide crossings.
2. If only speed hunts, inspect command, feedback, drive limits, and mechanical backlash. If only the process hunts, inspect its sensor and external load.
3. If both hunt, compare phase before retuning: process movement leading speed suggests a real disturbance or sensor problem; speed leading process suggests aggressive tuning or actuator/drive behavior.
4. Exclude startup, staging, setpoint reset, smoke/purge, and manual tuning tests.
5. Change one tuning parameter at a time and preserve protective/current/torque limits. Observe for at least two evaluation windows after each change.

Step 5 — Electrical and drive service

1. Check VFD input and output power and calculate drive efficiency only after mode, scaling, and control-state findings are resolved.
2. Check harmonic distortion where symptoms or site requirements justify it.
3. Follow the manufacturer’s diagnostics for the exact drive and motor.
4. If the VFD is failing, replacement is typically $500–$2,000 depending on motor horsepower; confirm the diagnosis before replacement.

Step 6 — Separate pump delivery signatures (PMP-0001, PMP-0002, PMP-0003, PMP-0005)

1. Validate command and independent run proof first. An active PMP-0003 makes a stopped/running inference unreliable; resolve that before replacing a hydraulic component.
2. Confirm flow and DP belong to this individual pump branch. A common-header flow point cannot distinguish a failed lag branch or a passing check valve.
3. Running with no flow is PMP-0001. Running with high pump DP and low flow is the more specific deadhead signature PMP-0002. A stopped branch with flow is PMP-0005; do not treat these mutually different premises as one alarm.
4. For PMP-0005, verify meter sign and zero, then compare flow with other-pump status and header pressure. Inspect the discharge check valve, isolation valves, bypass paths, and approved gravity/free-cooling arrangements.
5. Remote fix: check the differential pressure setpoint — it may be set too high, forcing the active pump to work against closed valves.
6. Remote fix: implement a differential pressure reset sequence if one doesn’t exist — the same trim-and-respond logic as air-handler duct static pressure reset (see the missing-reset playbook). EEM-10: 0.5–2% site energy savings.
7. On-site: check for closed isolation valves and a blocked strainer.
8. On-site: verify the pump impeller and coupling — damage can produce no flow despite the motor running.
9. For variable-primary CHW systems, verify the minimum-flow bypass valve is functioning. Without it, the lead pump may deadhead when all AHU valves close.

Step 7 — Review starts and staging (PMP-0004)

1. Use per-pump proof and a fixed evaluator tick that can resolve the shortest cycle. At the defaults, acquire at 60 seconds and set count_scale=evaluation_window/tick; COV loss can hide starts.
2. Compare each start with plant enable, lead/lag transfer, DP/temperature demand, minimum on/off timers, and any approved exercise sequence.
3. Check whether VFD speed/process hunting (VFD-0004) is repeatedly crossing a run threshold. Fix the unstable loop before changing motor protection.
4. Inspect overload, safety, and drive fault histories for trip/auto-reset cycling. A chattering proof device can create the same count.
5. Apply the pump/motor manufacturer’s starts-per-hour limit; the library’s default is only a commissioning placeholder.

Step 8 — Compare actual and expected power (PMP-0006)

1. Confirm actual and expected kW cover the same pump motor/drive circuit and the expected model is ready, fresh, in-domain, and fitted on known-good data.
2. Compare power with speed, individual-branch flow, differential pressure, staging, and fluid condition. A model indexed on a different pump configuration is not a degradation finding.
3. Resolve same-drive VFD tracking or mode/bypass findings before trusting a baseline that uses those signals. Scope any suppression to this pump/drive.
4. Check power-sensor scaling and phase coverage before mechanical work.
5. After control and sensor checks, inspect strainers, impeller, coupling, alignment, bearings, seals, motor, and drive. Refit the baseline only after the equipment is known clean; never train the fault into normal.

Step 9 — Confirm resolution

1. Verify the drive remains in the commissioned remote-auto state with bypass inactive whenever normal operation is expected.
2. Verify output tracks command within the commissioned tolerance.
3. Verify the process returns within its allowance band without sustained minimum/maximum saturation or material hunting.
4. For pumps, verify expected branch flow/DP, zero stopped-branch flow, and a compliant per-pump start count over at least one full evaluation window.
5. Where PMP-0006 applies, verify actual power returns inside the commissioned residual band without refitting on the faulty interval.
6. Confirm suppressions release on the same equipment instance and observe for at least two hunting/baseline windows before closing an instability or degradation finding.