# What should be the Order of Activties in a MSA Study?

Discussion in 'Gage R&R and MSA - Measurement Systems Analysis' started by Juggler_IN, Oct 20, 2015.

Tags:
1. ### Juggler_INNew Member

Joined:
Oct 20, 2015
Messages:
2
0
Trophy Points:
1
For a variable Gage R&R study we study the following:

Gage Precision:
- Repeatability
- Reproducibility

Gage Accuracy:
Bias
Linearity
Stability
Calibration

Gage Resolution:
Distinct Categories

Now, as per my understanding, as far as the sequence is concerned, the process flow for Gage R&R studies is:

Resolution
Bias
Linearity
Calibration
Repeatability
Reproducibility
Stability

Let me know if the order is any different.

2. ### Bev DModeratorStaff Member

Joined:
Jul 30, 2015
Messages:
620
679
Trophy Points:
92
Location:
Maine
By resolution do you mean the number of distinct categories? if so, you cannot calculate the NDC unless you have the repeatability value.
If you mean simply the resolution of the gauge vs the tolerance tehn you can and should look a this first. You don't need parts or any measurements to do this.

Calibration can be done next - but it only establishes 'accuracy' and can be effected by repeatability depending on hwo the calibration is performed. typically calibration is done to a standard and doesn't involve fixturign or any other real part that can effect repeatability...

Repeatability is the next measurement test. Without establishing repeatability, none of the other calculations are reliable. IF you will be using multiple operators to determine bias, stability, or linearity, you should etablish the reproducibility next.

3. ### Juggler_INNew Member

Joined:
Oct 20, 2015
Messages:
2
0
Trophy Points:
1
So, as I understand the flow would be as follows:

Step 1. Gage Resolution vs. Tolerance
Step 2. Gage Calibration against a Standard
Step 3. Gage Repeatability
Step 4. Reproducibility (with multiple operators)
Step 5. Bias
Step 6. Linearity
Step 7. Stability

Is this a fair order?

4. ### Bev DModeratorStaff Member

Joined:
Jul 30, 2015
Messages:
620
679
Trophy Points:
92
Location:
Maine
I do refer you to Miner's blogs in the resources section of the forum (see upper left tabs to move from the forums to the resources) as he is an expert in this subject.

Traditionally, Bias and Linearity are part of calibration. Bias will tell you how 'inaccurate' the device inherently is from a 'gold standard' method or standard object of known geometry or property.
Linearity will tell you if the bias is consistent across the response of the thing to be measured.
Typically these two are assessed together as part of the calibration study.

Stability is how the gauge - or method - changes over time. Some systems will be susceptible to transients such as temperature or humidity or vibration. Some will be susceptible to a systemic temporal forces such as age, wear during use, etc. So stability is typically a larger - and longer study. It can be done after the initial studies are complete that would validate the use of the system. Stability then would cap the length of time, or number of uses, or set limitations on environmental conditions.

So I might perform these in this order:

resolution against the tolerance
calibration against standards (including bias and linearity)
repeatability
reproducibility (alternately a single large study can be performed for both repeatability and reproducibility at the same time. this is quite common)
stability. (unless I have reason to be concerned about stability under typical use conditions...then I might combine repeatability and stability together before approving the system for use. I would do this after the calibration step and then perform reproducibility last)

Mark Paul and Miner like this.

Joined:
Sep 21, 2015
Messages:
14