PID using OPAMPS or transistors

PI is very simple. you feed PV (process variable, or measured value) or SP (setpoint) into the - input through a resistor, and the SP or PV into the + input. The feedback from output to - input is a capacitor in series with a resistor. This configuration produces both integral and proportional effects on error (integral effect always takes place on error, since its job is to bring average error to zero, eventually). If you want to be able to move the setpoint, and have only integral response take you to the new setpoint, you can't put both effects in the same feedback path.

Adding derivative also involves some choices. Do you want the derivative to act on error (setpoint and process changes) or just on process change?

Then you start adding things like anti-windup prevention on the integral, etc.

1 opamp implementation: ||Cd ___ ||Ci .--------||-o--|___|--||-. | || | R2 || | in | ___ | |\\ | o--o-|___|-----o--|-\\ | out R1 | >------o--o .-|+/ | |/ === GND (created by AACircuit v1.28 beta 10/06/04

If you know the corner frequencies fd and fi and the required gain Ap. Ap= - R2/R1

fi= 1/(2pi*Ci*R2)

fd=1/(2pi*Cd*R1)

This is valid for Cd/Ci

Hi, Thanx for the solutions Il build upon it.

Rgds Vipin