Close loop DC control is the best solution for TEC operation. It provides most effective use of TECs.

In practice two control methods are in use:

For most applications PI Control looks is enough. But RMT recommends to use most complex PID control loops.

Usually there is not so much difference in hardware of such two types (PI and PID), only one additional coefficient - a derivative circuit to the PI controller which improves the transient response of such control.

Effectiveness of PID/PI controls depends on correct adjustments of parameters of the Loop:

Calculation of the parameters requires knowing or to examining some specific properties of TEC. One of the methods - Ziegler-Nichols tuning approach.

By these parameters it is easy to calculate parameters of both PI or PID controls:

where  K = time constant / (process gain * deadtime) .

In many cases tuning of the PID parameters is done by trials and errors. It is quite long way and requires experience. There are high-end TEC Controllers which provide auto-tuning functions. Even with PID loop parameters obtained this way,  final fine tuning is very useful.

Knowledge of object of the control (TEC) is very important for such working. Rough estimations on the basis of the knowledge make PID tuning  easier and more effective: