Group work sampling calculation

[DennyO]

WE have a large material equipment fleet in a distribution facility. There are three general categories of equipment in use. I am trying to determine the utilization by work sampling, but to pin this down on an individual machine basis would take around 1,000 observation tours of this large facility, and this is a steep time investment.

If I am happy to get the utilization by equipment class only, not individual machine, within a couple of percentage points how could I calulate the sampling needed for that tolerance of accuracy? +/- 2% on a smaller element of say 5%, and +/- 3% on a larger one of say 20%?

Any references available?

The smallest equipment class has 7 machines, the largest has 32, if that helps.

 

[BML]

Could you look at it from the point of view of determining the probability that any one of the machines in the group is active? Or would it be possible to have some sort of electronic data capture from all the machines collectively?