How to choose parameter levels for Full Factorial ( FF ) and Variable Search ( VS ) experiments

Though there are a lot of similarities between the two - the purpose of the 2 experiments is totally different. FF helps you to fine tune process parameters, to improve results; whereas VS is to determine which parameters are critical or important, or rather to weed out noncritical parameters.

?First one carries out VS to determine critical parameters, which will be in most cases not more than 4. And then carry out FF on these 4 or less parameters to determine the optimum one.

In both experiments, it is recommended that one set of parameters be always those which are in use, as they are supposed to be the known as the best,( that is why they are being used.) Unless you have specific reason not to choose them. The second set of parameters are chosen differently.

For VS choose second level which is worse or marginal( which should be otherwise in use in normal production.)Normally it is known to persons who are managing the process.  Do not choose a level which will give 100% rejection or worst possible results.

Whereas in the case of FF, the second parameter should be the one which is likely to be better than the original level. This is less likely to be obvious, that is why it is not being used. Many times it could be because of interaction effects. The purpose of the experiments is to confirm the not so obvious combination of parameters which would give better than current results.