V
Hello all,
We are currently measuring 5 properties from a membrane roll, the pass/fail criteria the average of all points is between a tolerance. This tolerance set up by the client.
Currently we are measuring from 24 to 10 different points of the roll depending the property, from the west to the east of the wide.
Theoretically speaking, the properties across West to East should be the similar.
So potentially I can have a good chance to reduce the number of samples from 24 or 10 to 2.
For this I carried out a SPC analysis of the variance, determining the Between/ within sample variance. getting what is the variance associated to within the sample and the variance to sample to sample.
When I take the historical data, and I delete the measurements leaving just 2 the variance is barely the same.
Making a ratio btween the full testing and 2 points tested is really close to 1.
So this may show me that I am in the right direction, unfortunately this is not a proof or evidence enough to reduce the number of tests.
Ho can I address a testing or hypothesis to prove that the reduction of testing is OK?
Thanks for the feedback
We are currently measuring 5 properties from a membrane roll, the pass/fail criteria the average of all points is between a tolerance. This tolerance set up by the client.
Currently we are measuring from 24 to 10 different points of the roll depending the property, from the west to the east of the wide.
Theoretically speaking, the properties across West to East should be the similar.
So potentially I can have a good chance to reduce the number of samples from 24 or 10 to 2.
For this I carried out a SPC analysis of the variance, determining the Between/ within sample variance. getting what is the variance associated to within the sample and the variance to sample to sample.
When I take the historical data, and I delete the measurements leaving just 2 the variance is barely the same.
Making a ratio btween the full testing and 2 points tested is really close to 1.
So this may show me that I am in the right direction, unfortunately this is not a proof or evidence enough to reduce the number of tests.
Ho can I address a testing or hypothesis to prove that the reduction of testing is OK?
Thanks for the feedback