Identify Variation

On your previous thread..
https://www.pcdmisforum.com/forum/pc...594-cmm-health
I gave you a Gage Correlation Document. It will get you past an auditor asking "How do you know your cmm program is accurate/yielding repeatable results?"

What if the auditor asks "How do you know your process is accurate/yielding repeatable results?"

When it comes to "Process Capability Studies", the industry standard for aerospace & medical is performing a "CPK Study".
What is cpk?
Cpk is a stats term. In your example you'd measure the diameter 30 times & record all of that data. The data gets checked for two things...
-Are all data points within tolerance?
-If all data points were chatted on a graph, how "close"/distributed are the points to eachother?
If the cpk number is above 1.33 your process is mathematically proven to yield repeatable results

At a previous role, my main responsibility was to:
-Write CMM program to measure a new medical product that was in the prototype phase
-Measure a lot of 30 pieces & output the dimensional data from those parts to minitab
-Use minitab to output a CPK study and a few other charts/graphs that bosses love to oogle at
-Present this info to management team, customers, and FDA

If you can't get minitab you can go online & find an excel form that has cpk stuff baked into it.

With the data you shared above you're like ±0.00015 on everything...they definitely shouldn't be crying unless your tolerance is crazy tight?


Hope this rant helped lol

On your previous thread..
https://www.pcdmisforum.com/forum/pc...594-cmm-health
I gave you a Gage Correlation Document. It will get you past an auditor asking "How do you know your cmm program is accurate/yielding repeatable results?"

What if the auditor asks "How do you know your process is accurate/yielding repeatable results?"

When it comes to "Process Capability Studies", the industry standard for aerospace & medical is performing a "CPK Study".
What is cpk?
Cpk is a stats term. In your example you'd measure the diameter 30 times & record all of that data. The data gets checked for two things...
-Are all data points within tolerance?
-If all data points were chatted on a graph, how "close"/distributed are the points to eachother?
If the cpk number is above 1.33 your process is mathematically proven to yield repeatable results

At a previous role, my main responsibility was to:
-Write CMM program to measure a new medical product that was in the prototype phase
-Measure a lot of 30 pieces & output the dimensional data from those parts to minitab
-Use minitab to output a CPK study and a few other charts/graphs that bosses love to oogle at
-Present this info to management team, customers, and FDA

If you can't get minitab you can go online & find an excel form that has cpk stuff baked into it.

With the data you shared above you're like ±0.00015 on everything...they definitely shouldn't be crying unless your tolerance is crazy tight?


Hope this rant helped lol