Valve Stiction vs Poor PID Tuning: Evidence to Check

An oscillating process value does not establish that the PID controller is poorly tuned. Static friction in a valve assembly, measurement problems, interacting loops, disturbances and controller settings can create different or overlapping patterns. A useful investigation compares the controller request with independent evidence about what the final control element and process actually did.

Stiction concerns resistance to the start of motion. Controller tuning concerns how control action responds to error and dynamics. They can coexist, so a forced either/or diagnosis is often too simple.

Record A shows a position indication remaining still while the command changes, followed by a larger movement. Record B shows the position following each displayed command. This invented record is a reasoning exercise; four observations cannot validate a mechanical diagnosis or a tuning model.

Record A gives a reason to investigate the command-to-motion path, including static friction, actuator or positioner behavior, and whether the position feedback is current and trustworthy. It does not prove stiction. A stale or coarsely sampled position signal could also create an apparent hold-and-jump pattern.

Record B does not show that pattern in these samples. If the process value still oscillates, the investigator would need other evidence about disturbances, loop interaction and dynamics. Position following the request does not prove that tuning is correct, nor that the valve’s flow characteristic is suitable.

Manufacturer documentation and authorized diagnostic records can help distinguish a mechanism from a trend impression. Do not infer real valve travel from controller output alone.

• Align timestamps for setpoint, measured process value, controller output and available valve-position information.
• Record operating mode, limits, sampling interval and relevant process changes.
• Check whether the trend display or data collection can hide short movements or interpolate between samples.
• Separate observations from hypotheses: write what moved, what did not move and which evidence is missing.
• Keep hardware, measurement and controller explanations open until discriminating evidence is available.

Retuning cannot physically remove friction, and replacing a valve does not automatically settle every control-design issue. The appropriate next action depends on the diagnosis and the site’s authorized procedures, not on this illustrative table.

The control-loop diagnostics guide explains performance evidence and competing causes. Review MON300 monitoring and industrial statistics training for its published topics. The gain-versus-proportional-band example addresses a separate parameter-convention problem that should not be confused with mechanical behavior.

Technical definitions are supported by the primary references below. The labelled numerical records, timelines and checklists are original teaching examples for this article. They do not establish site-specific design approval or a new course entitlement.

• Emerson: Control Valve Handbook
• Valmet: valve diagnostics manual

Browse the process-control learning library for the wider subject guides and course routes.