LADX
Start free
← Resources
Practice·8 min read·

PID tuning without the maths

A method that gets a loop stable in twenty minutes, and the three symptoms that tell you which term to change.

Short answer

Start with proportional only, increase gain until the loop oscillates steadily, then halve it. Add integral until the offset disappears, slowing it if the loop starts hunting. Leave derivative at zero for most loops, because it amplifies measurement noise and is rarely needed on temperature or level.

The closed loopSP-PIDValveProcessmeasurement closes the loop

Tuning is usually taught with transfer functions and applied with trial and error. The trial-and-error version, done in a disciplined order, works.

The order

1. Proportional only. Set integral and derivative to zero. Increase gain until the loop oscillates with a steady amplitude that neither grows nor decays.

2. Halve it. That gives you margin.

3. Add integral. Increase until the steady-state offset disappears. If the loop starts hunting slowly, back it off.

4. Leave derivative alone unless you have a specific reason.

Reading the symptom

  • Fast oscillation. Proportional gain too high.
  • Slow hunting. Integral too aggressive.
  • Never reaches setpoint. Not enough integral.
  • Jumpy output with a noisy measurement. Derivative, or a measurement that needs filtering.
A loop that is slightly slow and completely stable beats a fast one that oscillates. Operators turn oscillating loops to manual, and a loop in manual is not controlling anything.

Before you tune anything

Check the valve. A sticking valve, a badly sized one, or one with significant hysteresis cannot be tuned around, and hours get spent trying.

Common questions

Should I use derivative in a PID loop?
Usually not. Derivative responds to rate of change, which means it amplifies measurement noise, and most industrial loops are stable and fast enough with PI alone. It earns its place on slow, well-damped processes such as large thermal masses where anticipating the change genuinely helps.
Why does my PID loop oscillate?
Most often proportional gain is too high, so each correction overshoots and the next correction overshoots back. Halve the gain and see whether the oscillation decays. If it oscillates slowly rather than quickly, integral action is too fast and is winding up before the process has responded.

Keep reading