Suppose you're called into the control room to correct an alleged loop tuning problem. Probably the worst thing you could do is to immediately begin changing a tuning parameter. There are several things you should do first.

1. Find out as much as you can about the loop. If this is not a new loop, then has this tuning problem just started? If so, what has changed? Is the process operating at a different condition? Is the set point different? Does this happen some of the time but not others?

2. Put the loop on manual. If the oscillations persist, there is certainly not a tuning problem with this loop; rather, some other loop is oscillating and that is being reflected by the oscillations of this loop.

3. Consider other equipment in the loop. Give particular consideration to the valve. Is it sticking? Is it operating very close to either end of its calibrated travel?

4. Understand the process phenomena. Is there something unusual about the process or its behavior?

5. If there is a controller tuning log available (it is a "must" for every control room!), then examine it. What has been the tuning history of the loop?

6. Finally, if you are convinced that this truly is a tuning problem, then make note of the existing tuning parameters. Try to improve the loop performance by the technique "improving as-found tuning parameters" described earlier, rather than by making either open-loop or closed-loop process tests.

Once you have found new parameters,...