In many industrial cases, faults are measured as the number of problems per thousand operations. Examples include the number of wrong items per thousand items dispensed from a vending machine, or the number of paper jams out of a thousand sheets of paper fed through a photocopier. Is this a good measure?
No! The characteristic y = number of paper jams per thousand sheets is a poor quality characteristic. It is also wasteful of paper and machine wear.
|
|
How do we do it?
|
Our strategy is as follows:
A larger operating window that takes account of the force that causes the paper jam would provide a better quality characteristic. The idea of the operating window is explained with reference to a simplified paper feeding mechanism. The force applied on the paper tray depends on the force f maintained by the spring. Suppose:
f = x is the force (misfeed force) at which one sheet of paper starts to feed (paper feeding threshold) and
f = z is the force (multifeed force) at which more than one sheet of paper starts to feed (paper multifeed threshold).
If the objective is set to:
Minimize x as a smaller-the-better characteristic and
Maximize z as a larger-the-better characteristic, then
PM = -10 log (MSD of x ) - 10 log (MSD of z )
and the problem is reduced to one of maximization. This technique is also called the Operating Window technique.
|