Symptom
The valve position and the process variable oscillate together with a period of seconds to minutes; the actuator is constantly moving; the positioner air is constantly bleeding; the process never settles at setpoint; the trend shows a regular cycle.
The short answer
Valve Hunting
A hunting valve is an oscillation, and the first question is who is oscillating: the control loop, the positioner, or the valve itself. Put the controller in manual with a fixed output; if the valve settles, the loop tuning or the process is the cause, and the gain or the integral is reduced or the loop is re-examined for interaction with another; if the valve continues to cycle on a fixed command, the positioner gain is too high, the feedback or the command is noisy, or the valve has stiction that makes it stick and then jump past the target, producing a limit cycle that no tuning cures. An oversized valve controlling near its seat has a high and nonlinear gain that makes any tuning marginal; a valve fed by a supply pressure that fluctuates hunts with the supply. The fix follows the finding: tune the loop, tune or replace the positioner, fix the noise, repack or repair the valve, or replace an oversized valve with one that works in its middle range.
Key points
- Controller in manual: if the valve settles, it is the loop; if it keeps cycling, it is the positioner, the signal, or the valve.
- A limit cycle from stiction has a square-ish position trace that jumps; tuning does not cure it.
- A positioner gain set for a small actuator on a large one overshoots on every move.
- An oversized valve controls at the bottom of its travel where gain is highest; nothing tunes well there.
- Two loops that share a variable can fight; detune one or decouple them.
Possible causes and what to check
| Possible cause | What to check |
|---|---|
| Loop tuning too aggressive | The oscillation stops with the controller in manual. Reduce the gain or lengthen the integral time; retune from a step test. |
| Positioner gain too high or wrong for the actuator | The valve cycles on a fixed command; the positioner overshoots on every step. Reduce the positioner gain per its manual, or run its auto-tune with the correct actuator parameters. |
| Stiction in the valve | On a fixed command the valve does not move for small changes and then jumps; a step test shows a dead time then an overshoot. Repack, lubricate, or repair the valve; a positioner with stiction compensation helps a little. |
| Noise on the command or the feedback | The command or the feedback signal shows noise on a scope or a fast trend; the valve chases it. Shield and route the signal cable; filter the command lightly; fix the feedback pot. |
| Valve oversized, working near the seat | The valve operates below 20 percent open; small movements cause large flow changes. Replace with a smaller valve or a reduced trim, or add a smaller valve in parallel for low flows. |
| Two loops interacting | A pressure loop and a flow loop on the same line each correcting the other. Detune the faster one, or cascade them properly. |
| Supply air pressure fluctuating | The supply gauge cycles; the valve follows. Fix the regulator, the compressor, or the dryer. |
| Process dead time and a fast controller | A pH or a residual loop with a long sample delay tuned as if it were fast. Retune for the dead time; a slower integral. |
| Deadband in the controller set too small for the actuator | The output changes by less than the actuator can resolve, and the actuator hunts around the target. Set the controller output deadband to the actuator resolution. |
The manual test
- 1
Trend it
The setpoint, the process variable, the controller output, and the valve position feedback on one trend at a fast rate. Note the period and the shape.
- 2
Manual
Put the controller in manual with the output frozen at the current value. Watch the valve position and the process.
- 3
Read the result
Valve settles and the process settles: the loop was oscillating; tune it. Valve settles and the process still cycles: another loop or a disturbance is driving the process. Valve keeps cycling on a fixed output: the positioner, the signal, or the valve.
- 4
Step it
In manual, step the output by 2 percent, then 5, then 10, and watch the position. No response to small steps then a jump is stiction; overshoot and ringing on every step is positioner gain; a clean response is a healthy valve, and the loop is the problem.
- 5
Fix and retest
The fix from the finding, then automatic with the trend running through a setpoint change and a disturbance.
Reading the trace
| Trace shape | Likely cause |
|---|---|
| Smooth sine wave in output and process, stops in manual | Loop gain too high or integral too fast |
| Position jumps in steps, square-ish process cycle | Stiction limit cycle |
| Position rings after every output change | Positioner gain too high |
| Small rapid jitter on position | Noise on the signal, or deadband too small |
| Oscillation only when another loop is active | Loop interaction |
| Oscillation only at low openings | Oversized valve, high gain near the seat |
| Oscillation with the supply pressure | Air supply |
Stiction
A valve with tight packing, a corroded stem, or a worn positioner linkage does not move until the actuator force overcomes the friction, and then it breaks free and overshoots. The controller sees no response, integrates, increases its output, the valve jumps past the target, the controller reverses, and the cycle repeats: a limit cycle whose amplitude is set by the stiction, not by the tuning. Detuning slows the cycle without removing it. The cure is mechanical: loosen or replace the packing, clean and lubricate the stem, repair the linkage, and then retune. A positioner with a stiction compensation feature can reduce the amplitude on a valve that cannot be fixed immediately.
Oversized valves
A valve sized for a future flow controls today near its seat, where a percent of travel is a large fraction of the flow and the characteristic is at its steepest. The loop gain is high at low openings and low at high openings, and tuning that is stable at one is unstable at the other. A reduced trim, a smaller valve, or a small valve in parallel for the low range gives the loop a valve that works in its middle, and the hunting stops without heroic tuning. The valve position trend shows where the valve lives; a valve that spends its life below 20 percent is oversized.
Frequently asked questions
- I reduced the gain and it still hunts, just slower.
- That is stiction: the amplitude stays and the period lengthens as the controller slows. Do the step test in manual; a valve that does not respond to small steps and then jumps has friction that tuning cannot cure. Fix the valve.
- The valve hunts only at night.
- At night the flow is low, the valve is near its seat, and the gain is high; or the air supply pressure changes as other users drop off. Look at the valve position at night on the trend, and at the supply pressure.
- Can I add deadband to stop the hunting?
- A deadband on the controller output that matches the actuator resolution stops the actuator chasing tiny changes, and it is legitimate. A large deadband to hide a limit cycle leaves the process cycling inside it. Use the smallest deadband that stops the jitter, and fix the valve.
- The positioner auto-tuned and the valve now rings.
- The auto-tune was run with wrong actuator parameters, or with the valve not at operating conditions. Set the actuator type and size, run it again with the process pressure on the valve, and then reduce the gain a step from what it chose; auto-tunes prefer speed over damping.
Related topics
- PID Control for Pumps and ValvesWhat each term actually does, how to tune a loop on a real pump station, and the handful of mistakes that cause most of the oscillation people blame on tuning.
- Deadband and HysteresisWhy a single setpoint chatters, how a deadband turns it into a start point and a stop point, how wide to make it, and where deadbands belong in alarms, level control, and controller outputs.
- How to Create a PID Loop in a PLCSet up a PID instruction from scratch: scale the PV and CV, pick the action, set the execution rate, configure limits and anti-windup, tune conservatively, and test the manual and auto transitions before it controls anything.
- Valve Slow to RespondA valve that takes far longer to move than it should: a pneumatic supply restricted or leaking; a positioner tuned slow; an electric actuator whose speed is fixed; a stem binding in cold or packing; or a rate-limited output. Measuring the stroke time.
- Position Feedback WrongThe HMI says one position and the valve is at another: a feedback potentiometer worn or miscalibrated, a linkage that slipped, scaling that does not match, a reversed signal, limits set wrong, or noise. Comparing indicator, feedback, and process.
- TrendingUsing historian trends to understand a process and diagnose a problem: pens that belong together, scaling and time ranges, raw versus averaged data and what compression hides, overlays and cursors, event markers, and the patterns that point at each cause.
Direct contact
Have a controls question?
Reach Eric Sullivan directly about anything on this site, a controls or automation topic, or one of his personal projects.