When the proportional gain is very low (which could be used with a PID driving another setpoint), it is possible to not reach setpoint in steady state.
This is due to a combination of events.
This can only happen if the actuator is not reaching setpoint, by margin a. This margin is multiplied by the anti-windup gain of 3. If p equals 3a, we are stuck.
The integrator is increased with p every cycle.
The antiwindup is (pid result - achieved output) * 3.
When these two cancel each other out, the integrator is stuck in a steady state value, without the error being zero.
elcojacobs
commented
7 years ago
When the proportional gain is very low (which could be used with a PID driving another setpoint), it is possible to not reach setpoint in steady state.
This is due to a combination of events. This can only happen if the actuator is not reaching setpoint, by margin
a. This margin is multiplied by the anti-windup gain of 3. Ifpequals3a, we are stuck.The integrator is increased with p every cycle. The antiwindup is
(pid result - achieved output) * 3. When these two cancel each other out, the integrator is stuck in a steady state value, without the error being zero.