Control loop will not hold setpoint
A loop sitting steadily away from setpoint is a different problem from a loop that cycles, and it has a much shorter list of causes. In almost every case the controller has either been prevented from correcting the error or has no mechanism to correct it in the first place.
What causes it, most likely first
No integral action
most common tuning can fix this
Proportional action needs error to produce output, so a proportional-only loop settles wherever the proportional term balances the load — not at setpoint. The offset grows as the load moves away from where the bias was set. On averaging level control this is deliberate. On flow or pressure it almost never is.
- How to confirm it
- Check the controller structure and the reset value. Either a P-only structure or a reset value of zero (in time-per-repeat units) gives you this.
- What to do
- Confirm whether proportional-only was deliberate. If not, integral action needs to be enabled.
Output saturated
most common tuning cannot fix this
The controller is asking for more than the final element can deliver, so the error persists no matter what the tuning is.
- How to confirm it
- The output sits at 0% or 100%.
- What to do
- Deal with the constraint. See the saturated output page.
The loop has lost the selection in an override scheme
common tuning cannot fix this
In a selector scheme only the winning controller drives the output. A loop that has been overridden is not in control and will sit away from its setpoint by design.
- How to confirm it
- Check which controller currently has the output. A loop whose output does not match its error is a strong hint.
- What to do
- Nothing at the loop. Understand the override scheme and why it is active.
Reset so slow the offset never closes
common tuning can fix this
Integral action is present but slow enough that the offset persists for a very long time.
- How to confirm it
- The gap is closing, just extremely slowly. Watch over several minutes rather than seconds.
- What to do
- Speed the reset up in steps.
A calibration difference between measurement and reference
less common tuning cannot fix this
The loop is holding its own measurement at setpoint perfectly, and the measurement disagrees with whatever you are comparing it against.
- How to confirm it
- Compare the transmitter against a local gauge or a laboratory sample.
- What to do
- Calibration, not tuning.
Questions that come up
Can I just bias the setpoint to compensate?
You can, and on a proportional-only loop that is exactly what the manual reset or bias term is for. But it only holds at one load. As soon as the load changes, the offset changes with it and the bias is wrong again.
Related
Last reviewed 2026-08-01.