What is the difference between Process Variation and Tolerance?

[jeopie]

What is the difference between process variation and tolerance?

and what recommendation is there if a supplier submits a report and the process variation is in the 10-30% range but the tolerance is less than 10%?

 

[Ron Rompen]

Process variation is the observed difference from part-to-part which is caused by the limitations of the tooling, equipment, fixturing, etc.

Tolerance is the maximum allowed variation from the target dimension, which you are allowed per the customer print/specifications/requirements.

 

[Bev D]

jeopie - are you talking about a gage R&R?

 

[John Predmore]

What Ron R said is true.

Any time you measure parts, the data you collect include a bit of imprecision due to measurement variation. If your measurement system does not have adequate repeatability, you can’t be sure of what you are seeing in your process.

Using a designed experiment and assuming the parts themselves do not change between measurements, it is possible to statistically decompose the total observed variation into component portions. It is common in a Gage Repeatability & Reproducibility study (GR&R) to summarize what percent of the total observed variation is attributed to the parts, the operators and the (measurement) equipment. By consensus, the appraisal percentage (the repeatability of the “gage” and the reproducibility of the operators) should be less than 30% of the total. Your supplier is in that range.

When your measuring device is used for inspection, sometimes measurement variation is quantified relative to the total tolerance rather than process variation from the study.
When you wrote the supplier’s report shows “tolerance less than 10%”, I think you meant to say the portion which is measurement variability is <10% of the total tolerance. That would be a great result.

 

[bobdoering]

and what recommendation is there if a supplier submits a report and the process variation is in the 10-30% range but the tolerance is less than 10%?

The correct answer is: It depends.

If the gage is used to determine if the part is "good" or "bad", then you look at the %Tolerance value. In this case, no action is required.

If the gage is used to determine if the process needs adjustment, then you look at the %Process Variation value. In this case, there may be an opportunity for improvement.

If the %Process Variation is a better value than %Tolerance, then a close look at the part values used to represent the process is needed. Generally a sound study will have the %process variation greater than the %Tolerance.