Question calculating process capability on box sortation system

[calpoly]

Greetings fellow IE's,

I work for a distribution center where we have a sorter system that scans boxes from overhead and diverts the boxes down the shipping lanes and to the trailers. I am trying to calculate the Cp/Pp,Cpk/Ppk, but am having trouble figuring out if my Cp is accurate. I am calculating Cp on "no-reads" (failed scan reads) that are recirculated and passed through the scanning tunnel again until there is a read. Our volume fluctuates between 50k-120k thorughout the year and of course with the higher volume we have more no-reads. How can I determine an accurate Cp? If volume was a constant it would things easier, but is not. I thought of isolating no-reads for volumes at 50k, 60k, 70k, etc, but I do not believe it will truly give me an accurate Cp. The range is still quite large. Any feedback would be greatly appreciated. Thanks.

 

[rhodie]

I think you need an Xbar - R chart, a chart that tracks variable data along a continous scale.

 

[vilchws]

If you used a p-chart, which would chart the fraction or percent of non-conforming events – no reads, then you should be able to first see if you have a stable continuous distribution. This type of charting would help you deal with the changes in volume.

 

[calpoly]

Thank you for the responses so far. Please keep them coming.