# Stability in MSA

#### Bev D

##### Heretical Statistician
Super Moderator
I am not confused. I know objective of control charts in SPC and MSA both.
You have hurriedly replied without referring to MSA manula. Pl refer Page 86 of MSA manual and read last sentence on that page.
So I finally got a copy of page 86. And I have done as you commanded. This refers to a study of the stability - or consistency - of the measurement system over time using one part. (For those who also don’t have access to the manual). In this case there are actually no “trial limits”. The limits are used for a constrained time study and is not an ongoing monitoring; so no phase 1 and phase 2 limits.

Of course if there is a stability problem - a drift or shift over time - then one would repeat the study after removing the cause of the drift or shift to demonstrate that they were successful. New data would be calculated and limits calculated from that data.

Care must be taken to ensure tha tthe shift or drift over time is due to the measurement system or the part itself…and stability studies are not required unless it is believed that the measurement system itself is susceptible to changes over time or environmental conditions.

#### Miner

##### Forum Moderator
In this case there are actually no “trial limits”. The limits are used for a constrained time study and is not an ongoing monitoring; so no phase 1 and phase 2 limits.
The stability study would be based on phase 1 limits with no phase 2 follow-up. While we could argue semantics about there not being a phase 1 without a phase 2, the concept is identical to a phase 1, which is why I used it.

#### Rameshwar25

##### Quite Involved in Discussions
Thank you miner, My question is well answered now.
I perceive that we need to calculate trial control limits with a smaller group and to use those limits to draw cntrol chart for stability study.
Although this study is rerely used, I was trying to understand the method.

#### Bev D

##### Heretical Statistician
Super Moderator
Just for clarity for those who don’t have the manual: the process outlined is to measure one part 5 times. This constitutes one subgroup. The example used 20 such subgroups (part was measured 5 times (this gives you a within subgroup variation = measurement repeatability) each day for 4 weeks resulting in 20 subgroups). If you use more than one part you will need control limits (and the resulting chart) for each part.