KEYSTONEIndustrial Services
Interactive trainer

Control valve & positioner

A control valve is the loop’s final element: a positioner strokes it to match a 4–20 mA command using instrument air. When a loop can’t hold setpoint, compare where the valve actually is to where it was commanded before you touch anything else. Six valves, four reads — position, signal, air, bench — including the two undershoots that look alike (stiction versus starved air) and the two wrong-position faults that look alike (a span mis-calibration versus reversed action).

By Francisco Padilla Interactive Technician level

Before you start — job brief

A control valve is the loop's final control element: a positioner strokes it to match a 4–20 mA command using instrument air. When a loop can't hold setpoint, the valve is as likely to blame as the transmitter or the controller. The discipline is to compare where the valve actually is to where it was commanded, then read the signal, the air supply, and the bench set to name why it isn't tracking.

  • The positioner drives the valve to match the 4–20 mA command through an I/P and instrument air
  • Compare actual travel to the commanded % first — a valve that won't track is the fault
  • Stiction jumps through a deadband; starved air can't stroke it; a mis-calibrated positioner offsets or spans the travel
  • Reversed action moves it the wrong way; a ruptured diaphragm won't hold — confirm with signal, air, and the bench set

The valve is the final element — and it fails like one

A control loop can measure perfectly and compute perfectly and still hold the wrong flow, because the last device in the chain — the control valve — did not go where it was told. The valve is the final control element, and on almost every modern valve a positioner stands between the 4–20 mA command and the actuator: it reads the command, drives an I/P converter that turns current into pressure, and uses instrument air to stroke the valve until a feedback linkage says the plug is where the command asked. When a loop can't hold setpoint, the transmitter gets the blame first — but the valve is just as likely, and it announces its faults in ways you can read directly if you look at travel before anything else.

Compare actual travel to the command — first, every time

The single most useful reading on a control valve is the one people skip: where is the plug, versus where did the controller command it? A valve that sits exactly where commanded is tracking, and the loop's problem is elsewhere. A valve that sits somewhere else is the fault — and how it misses tells you which fault. Read the actual travel against the commanded percent first, and only then reach for the signal, the air and the bench set to confirm the cause. Everything below is a pattern in those four reads.

Stiction and starved air both undershoot — but not the same way

Two faults make a valve fall short of its command, and they are the classic mix-up. Stiction — over-tight packing or a galled stem — is friction the positioner has to overcome before the plug will budge. The valve sits short, then breaks free and jumps past the setpoint, leaving a deadband around every move; the signal is correct and the air supply is full. Low instrument air is the opposite character: the positioner commands correctly, but with the supply header starved the actuator can't build enough pressure to stroke the valve, so it moves smoothly and stalls against a ceiling it cannot pass. Same symptom — an undershoot — but stiction jumps with the air fine at 60 psi, while starved air glides to a stop with the supply down at 30. The air-supply gauge and the shape of the motion separate them.

Field note — the two wrong-position faults that look identical

A valve is holding a position that isn't the one commanded, and the signal and air both check out. Two very different faults do this, and the fix for one does nothing for the other. If the miss is in the same direction as the command — 65% for a 50% call, too far open but the right way — the positioner's span or zero is mis-calibrated: recalibrate it. If the valve moves opposite the command — 30% for a 70% call, closing when you ask it to open — the action is reversed: fix the direction. The bench check names it outright: a span error versus a reversed action. Always ask not just "is the position wrong?" but "wrong in which direction?" — offset the same way, or flipped.

Zero signal problems, and the valve still won't hold

Some valve faults have nothing to do with the command at all. A ruptured actuator diaphragm leaks air as fast as the I/P can feed it, so the actuator never builds the pressure the positioner is calling for and the valve drifts to its fail position — fail-closed, fail-open, whatever the spring dictates. The signal is perfect, the supply is present, and the valve still won't stay put: the tell is that the actuator won't hold pressure. This is why the air read has two parts — the supply feeding the positioner, and the actuator pressure the diaphragm is supposed to keep. A good supply with an actuator that bleeds down points straight at the diaphragm or its casing.

The discipline

Four reads name every fault a valve throws. Position — actual travel versus the commanded percent — tells you whether the valve is the problem and, by the shape of its miss, which problem. Signal confirms the 4–20 mA command and the I/P output, so you know the positioner is being told the right thing. Air checks the supply that powers the stroke and the actuator pressure that holds it. Bench — the set range, the air-to-open or air-to-close action, and clean seating — is where a calibration or reversed-action fault convicts itself. Read all four against what they should be, and the valve tells you exactly why it isn't following the controller, instead of leaving you tuning a loop around a final element that was never going to track.