The coal-feeder outlet valve (a DZW multi-turn electric actuator) at a circulating fluidized-bed unit drove the operators crazy: whether switched to remote or local, the valve cycled open-and-close by itself, endlessly, and the coal feeder could not start because of the abnormal valve state. The resulting forum thread drew nearly 90 replies and, taken together, amounts to a troubleshooting handbook for "a valve that cycles by itself."
The fastest localization method, widely agreed upon: disconnect the command wires.
- Disconnect the open/close command lines (remote and local) and watch: still cycling — the problem is inside the actuator; stops — chase logic and wiring;
- Even more rigorous: connect only the power leads and operate remote/local; then reconnect the open-command line and observe; then the close-command line — reconnecting one wire at a time always finds the faulty loop;
- Testing remote/local with only power connected also helps: if it misbehaves even locally, remote signals can be ruled out wholesale, narrowing to the actuator itself or the local circuit.
1. Failed internal control (main) board — the most frequent answer: "seen it several times; it was the main board; replaced and cured";
2. Travel (limit) switch failed or set wrong — "limits not seated or scrambled"; one plant's feeder kept tripping due to unstable limit-switch state signals until the switch was replaced;
3. Capacitor failure — abnormal start/run capacitors in on/off electric actuators also cause repeated cycling;
4. Failed remote/local selector switch — bad contacts mix remote commands with interlock signals.
One easily overlooked physical cause: the actuator has separated from the valve — "the actuator keeps turning, the valve never reaches position, so torque trips, resets, and it turns again" — a loop especially known on certain integral designs. A poor motor brake is another candidate.
If the valve stops cycling once the command wires are removed, the problem lives on the command side:
- Shorted open/close command wires: melted or chafed cable, vibration-induced loose joints — both commands arrive at once and the valve "cycles blindly";
- 220 V AC coupled into the control circuit: measured at one plant; behavior becomes completely erratic;
- Induced voltage: command cables routed parallel to power cables pick up enough induced voltage to drive the input channel — keep command/feedback cables separated from power cables;
- DCS-side faults: a failed DO card channel or terminal board, welded relay contacts, or a forced signal left in the logic all deliver phantom commands. Extreme case: one plant's DPU card failure set every valve, pump, and measurement flapping simultaneously.
Even mobile-phone/walkie-talkie interference made a control valve cycle at one plant until the interference source was moved away.
One I&C engineer published his plant's full post-mortem, a classic worth retelling:
> Coal-mill-head over-temperature trips the feeder and closes the outlet valve (interlock A), while another logic opens the outlet valve on over-temperature (interlock B, by design). But this particular outlet valve was a reverse-acting valve — its "open" is close and its "close" is open — wired as if it were a normal valve. Result: on every over-temperature alarm, interlock A commanded close and interlock B commanded open, the two commands tug-of-wared, and the valve cycled open-close-open-close until the belt burned, the vendor paid for it, and the PLC logic was finally corrected.
The lesson: a valve "acting by itself" need not involve any hardware fault — two simultaneously satisfied interlocks, or a fail-safe valve wired as a normal one, will make the logic fight itself.
1. Disconnection test to split actuator vs. signal/logic;
2. Actuator side: control board → travel switch → capacitor → selector switch → decoupled coupling / brake;
3. Signal side: shorted/damaged command wires → coupled/induced voltage (separate cabling) → card channels / relays / forced signals;
4. Logic side: audit every interlock's action direction, paying special attention to reverse-acting (fail-safe) valves and their wiring/configuration consistency;
5. Interference: walkie-talkie/phone tests, shielding and grounding checks.
"A valve cycling by itself" is a classic many-causes-one-symptom fault, but the procedure is clear: disconnect first to split internal vs. external; sweep the actuator (control board, travel switch, capacitor, selector switch); then audit the command circuit for shorts, coupled voltage, and card faults; finally re-examine the logic — especially the "reverse valve + two opposing interlocks" design trap. Among the 89 replies no two experts gave the same answer — yet every answer sits somewhere on the staircase above.
