PID devices are typically shipped by their manufacturers with default settings programmed for the proportional, integral and derivative. Operators installing and using any type of PID instrument must first calibrate the device, making sure it is properly programmed and adjusted to suit the specific needs of the industrial process in question. This also includes ensuring that the environmental parameters that it is required to operate between are appropriate for the potential variables occurring within that process. Until this process has been completed, the PID controller cannot be left to automatically handle its assigned workload.
This is equally true for PID temperature controllers, and there are a number of different ways in which this tuning can be achieved. In practice, all the different methods aim for the same result, which is to adjust, or tune, each of the proportional, integral and derivative terms individually in order for the module as a whole to deliver the desired performance.
Simple trial and error is often viewed as the most practical method of tuning PID controllers for many systems and scenarios. It is based on installing the device in a working system and effectively zeroing out all the settings. The operator then begins with proportional value, adjusting the gain upward until it reaches a point where it is oscillating around the setpoint.
Once this state is reached - which would be unsatisfactory in an accuracy-critical system and exposes one of the key limitations of proportional-only temperature control devices - the values for integral and derivative can be adjusted respectively. The former, if done correctly, should start to limit the oscillation rate down to almost zero, whereas the latter will increase response speed as optimum settings for the system and device type installed are neared.
An alternative to trial and error is the Ziegler-Nichols method. This approach involves observation of both continuous cycling and damped oscillation under a closed-loop system. However, this method has certain limitations and many operators still prefer to use trial and error to achieve their desired results.