— From the Perspective of THINKTANK Control Valve Engineers
Why we never accept “stroke test passed” as final acceptance
In many chemical plants, valve commissioning ends once a control valve passes a basic stroke test.
Typical checks include:
- 0% position
- 25% position
- 50% position
- 75% position
- 100% position
If the valve reaches all five points correctly, it is often considered “qualified.”
From our experience at THINKTANK, this is one of the most common and dangerous assumptions in control valve commissioning.
A valve can pass a five-point stroke check and still behave poorly in real process control.

Why positioner error and deadband matter in real operation
The valve positioner is the core device that translates controller output into valve movement.
Its performance is defined not only by stroke range, but by:
- positioning error,
- deadband,
- and response consistency.
If error or deadband exceeds acceptable limits, the valve may:
- hesitate to move,
- overshoot,
- or fail to respond to small control signals.
In a real process, these behaviors amplify flow fluctuations and directly stress the control loop.
A typical accident scenario we help customers analyze
One case we investigated occurred shortly after startup.
What commissioning showed:
- Five-point stroke calibration passed
- Valve position feedback matched controller output
- No mechanical issues were observed
What happened during operation:
- Process flow fluctuated abnormally
- Controller output changed continuously
- The valve response lagged and jumped
- The control loop became unstable
- The unit eventually tripped due to interlock activation
From the operator’s perspective, this looked like a “process instability.”
From our engineering analysis, it was a positioner performance issue.
Root cause: incomplete positioner testing
The root cause was not valve sizing or actuator selection.
It was simple and very common:
Positioner error and deadband were never tested.
Only stroke positions were checked.
Error, deadband, and response behavior were not verified according to calibration requirements.
In this case, the valve used a smart positioner.
Incorrect internal parameter settings caused a mismatch between:
- actual valve position, and
- controller output signal.
This mismatch was small enough to pass stroke checks, but large enough to destabilize the control loop during normal operation.

Why this problem only appears after the valve is put into service
Positioner deadband and error rarely cause immediate, obvious failures.
Instead, they:
- distort small control corrections,
- create oscillation,
- and gradually push the process toward instability.
This is why these problems often appear:
- after startup,
- during load changes,
- or under dynamic operating conditions,
rather than during static commissioning tests.

How THINKTANK engineers evaluate positioner performance
When we suspect a positioner-related issue, we go beyond basic stroke checks.
Our evaluation focuses on:
- positioning accuracy over the full range,
- deadband under small signal changes,
- repeatability during increasing and decreasing signals.
We compare actual valve movement directly against controller output, not just position feedback.
This allows us to identify whether the positioner is truly suitable for the required control performance.
Engineering corrections we typically recommend
Based on field experience, our recommendations are clear.
- Positioner commissioning must follow full calibration requirements, including error and deadband testing
- Passing a five-point stroke check should never be considered sufficient acceptance
- For control loops with high stability requirements, any positioner that cannot meet performance criteria should be replaced, not tolerated
From our perspective, adjusting tuning parameters to “mask” poor positioner performance only postpones failure.
Our engineering conclusion
A control valve that causes frequent oscillation or unexplained trips is often blamed on the process.
In reality, it may be revealing a positioning problem that was never properly tested.
A valve that reaches its stroke positions correctly
is not necessarily a valve that can control smoothly.
For high-demand control applications, positioner error and deadband testing are not optional — they are essential.