By the end of this lesson, the technician will be able to:
The Loop That Wouldn't Sit Still: A steam pressure control loop has started "hunting" — the PLC trend shows pressure sawtoothing around setpoint, and the controller output endlessly ramping up and down a few percent. The first responder retuned the controller twice; the cycling survived both tunings.
You pull up the two trends that matter: controller output vs. actual valve travel (the positioner reports position feedback). The picture is textbook: controller output ramps smoothly from 42% to 47% — valve travel doesn't move at all — then at 47% the valve suddenly jumps to 48%, overshooting the needed correction. Pressure crosses setpoint the other way, the controller ramps down, the valve sticks again, then jumps. The valve is moving like a dresser drawer with a sock jammed in the slide: stick... slip. Stiction.
No controller tuning can fix friction — tuning only changes how fast the controller winds up before each slip. With the process isolated and the valve under LOTO, the mechanical inspection finds packing gland nuts over-tightened during last month's leak repair, binding the stem. Packing is adjusted to specification torque, the stroke check shows smooth proportional travel, and the loop holds setpoint with the original tuning. The lesson: when a loop oscillates, look at the valve's behavior before touching the controller's brain.
The control valve is where the control system finally touches the process — the muscle at the end of the measure–decide–act loop (Lesson 6.1). A "control valve" is really a four-part stack, and each part fails differently:
The pressure-containing body plus the internal trim (plug, seat, stem, cage) that actually throttles flow. Trim geometry sets the flow characteristic — how flow responds to travel:
Trim wears: wire-draw erosion, seat damage, and stem/packing friction are the mechanical aging modes a technician inspects for.
Converts energy to stem force. The industrial workhorse is the spring-and-diaphragm pneumatic actuator: instrument air on a diaphragm opposes a spring. The spring is the safety logic:
The fail action is chosen by process safety analysis — fuel valves fail closed, cooling valves typically fail open. A technician never has discretion to change, defeat, or "temporarily rig" a fail action; that configuration is a process-safety decision (and on safety-instrumented services, an SME/engineering domain). The actuator's bench set (the spring range, e.g., 6–30 PSI) defines the air pressures at which travel starts and completes.
The bridge between the electrical and pneumatic worlds: converts the controller's 4–20 mA into a proportional 3–15 PSI pneumatic signal (the legacy pneumatic standard, still everywhere). 4 mA → 3 PSI, 12 mA → 9 PSI, 20 mA → 15 PSI. Modern smart positioners integrate the I/P function internally; standalone I/Ps still serve simpler valves. I/Ps die from the air they breathe: wet, oily, or dirty instrument air fouls nozzles and orifices — which is why instrument air quality (dry, oil-free, filtered, per ISA-7.0.01) is a valve-reliability issue, not just a compressor-room detail.
The positioner is a feedback controller bolted to the valve: it compares the commanded position (input signal) against actual stem position (mechanical or magnetic feedback linkage) and drives the actuator with whatever air is needed to close the gap. It exists because actuators alone are open-loop — friction, process forces on the plug, and spring tolerance all push the stem off the commanded position. Smart (digital) positioners add diagnostics that matter to technicians: travel deviation alarms, total stem travel counters, air consumption signatures, and stored valve signatures from commissioning.
Control valve faults announce themselves in the relationship between three signals: controller output (CO), valve travel (feedback), and the process variable (PV). Three signatures cover most calls:
CO ramps with no travel response, then travel jumps past the target. PV sawtooths; the loop oscillates in a distinctive square-ish/triangular pattern that retuning cannot remove (it can only slow). Causes: over-tightened or aged packing, scale on the stem, seat galling. Confirmed by a manual stroke test: smooth signal ramp in, stepped jerky travel out.
When CO reverses direction, travel doesn't respond until the signal has moved through a gap (lost motion in linkages, worn pins, loose feedback connections). Signature: PV limit-cycles gently around setpoint and the loop always seems "a little behind" direction changes. Measured directly during a stroke check: command 50% from below vs. from above and compare actual travels.
A valve that strokes correctly but slowly, or that can't hold position under process load: inadequate supply pressure, a clogged airset filter/regulator, an undersized actuator for the line pressure drop, or leaking diaphragm/tubing (listen — a hissing actuator is telling you the diagnosis). Smart positioner air-consumption diagnostics flag diaphragm and tubing leaks before they become stroke failures.
The isolation logic chain for "the valve isn't doing what the controller says": verify the mA signal at the positioner/I/P input (Lesson 6.3 methods) → verify air supply at the airset gauge (typically 20–35 PSI supply for a 3–15 PSI device, per the valve spec) → verify the I/P/positioner output tracks the input → only then condemn the mechanics. Each step is a one-minute check that cuts the suspect list.
| Component | Function | Healthy Indication | Signature Failures |
|---|---|---|---|
| Body & trim | Throttle flow per characteristic | Smooth ΔP/flow response across travel | Seat leakage (passing when "closed"), trim erosion, plugging |
| Spring-diaphragm actuator | Convert air to stem force; enforce fail action | Stroke start/end matches bench set (e.g., 6–30 PSI) | Diaphragm leak (hiss, slow stroke), spring fatigue, undersized vs. process load |
| I/P transducer | 4–20 mA → 3–15 PSI | 12 mA in = 9 PSI out (±tolerance) | Fouled nozzle from wet/dirty air; output stuck or offset |
| Positioner | Close the loop on stem position | Travel matches command at 0/25/50/75/100% | Feedback linkage slip/wear (deadband), calibration drift, instrument air starvation |
| Airset (filter-regulator) | Clean, regulated supply air | Stable supply at spec (20–35 PSI typical) | Clogged filter (slow stroke), regulator droop under demand |
| Packing | Seal the stem | No leakage; smooth stem friction | Over-tightened = stiction; worn = emissions/leak |
Signal correspondence to memorize: 4 mA = 3 PSI = 0% command · 8 mA = 6 PSI = 25% · 12 mA = 9 PSI = 50% · 16 mA = 12 PSI = 75% · 20 mA = 15 PSI = 100%.
Boundaries first — this checklist's safety preconditions are absolute:
Procedure:
The stick-slip pattern is the most recognizable valve signature in loop diagnostics, and its defining property is exactly what the technician observed: tuning changes the oscillation's period, never its existence, because the root cause is static friction in the mechanics, not control aggression. Answer A doubles down on treating a mechanical disease with a software remedy — and a detuned loop responds sluggishly to real disturbances. Answer B blames a component whose failure looks different (offset, stuck, or dead pneumatic output — verified in one minute by checking I/P output pressure against input mA). Answer D is actively dangerous reasoning: overdriving supply pressure beyond the valve's specification doesn't cure friction — the positioner already applies full available force during the stick phase — and risks diaphragm and component damage. The professional sequence is signal → air → I/P/positioner → mechanics; everything upstream checks healthy here, and the diagnostics point downstream. Stiction is fixed with wrenches under LOTO, not with parameters.