Linearity study on a non-linear gage

[beaser3]

Hello,

I am not a calibration/MSA person and I have what may be a few basic questions regarding MSA. How do you determine if a gage is linear or non-liners? Can linearity studies be done on non-linear gages (CMM's, vision systems, etc...)?

Thanks in advance for any input.

 

[Miner]

I have a blog on Linearity that should answer your first question.

Regarding the second question, why do you call these gages nonlinear? It it because they are capable of 3D measurements?

 

[JoShmo]

Hello,

I am not a calibration/MSA person and I have what may be a few basic questions regarding MSA. How do you determine if a gage is linear or non-liners? Can linearity studies be done on non-linear gages (CMM's, vision systems, etc...)?

Thanks in advance for any input.

A cmm isn't linear? Run, Forrest, run!

 

[beaser3]

Thanks for the link. That does help me understand linearity better.

Regarding gages that are linear or not....I'm not even sure if any gages are non linear.

This is being driven by an audit finding that we received for not calibrating over the full range of a gage. Our calibration house calibrated toward lower end, middle and near upper end but not complete range. I got pulled in last minute to review our response and found that our procedures do not require us to calibrate over full range, only 3 spots along the linear travel. Our MSA procedure states that we have to perform linearity studies on new variable gages along the full range of the gage. That would cover the full range requirement. So, I'm not sure the audit finding is legit. I pointed this out to the auditor and her response was as follows:

You can do it the way you are talking about if you differentiate measuring devices that are ‘linear’ and those that are not, because then you have to specify the range over which to calibrate for the ‘non-linear’ devices. That is why most all organizations choose to specify the range or points to calibrate at over the range. It keeps it simpler for them rather than dealing with linear and on-linear. Your choice.

Now I'm just trying to figure out what she is referring to by 'non-linear". I was told that non linear would be gages like cmm, vision systems, etc that can measure in a full area vs along a linear straight line scale.....


Thoughts???

 

[Miner]

A linearity study would establish whether a gage's response is linear or not. If it is linear, you can use the approach that you currently use. A CMM would be trickier because you would probably have to study all three axes as well as diagonal.

 

[beaser3]

ok, so linearity does not apply to the gage itself, it applies to the measurements of the gage.

Thank you so much!

 

[JoShmo]

ok, so linearity does not apply to the gage itself, it applies to the measurements of the gage.

Thank you so much!

Gas gauges in moderm cars usually aren't linear. The gas tank shape is all kindsa weird. The gauge might try to be linear, but what it's saying isn't.

"Now I'm just trying to figure out what she is referring to by 'non-linear". I was told that non linear would be gages like cmm, vision systems, etc that can measure in a full area vs along a linear straight line scale....." Who is "she"? I'm not sure I'd trust any comment from a person who told me a cmm isn't linear...

 

[beaser3]

She would be the auditor but she did not state that the CMM was not linear. That was my assumption (and obviously a bad one) based on the little I know about this. I now understand that the linearity does not apply to the gage but rather the result on any given variable gage. Below is the comment that our auditor made. The reference to "non-linear" devices was throwing me off.


"You can do it the way you are talking about if you differentiate measuring devices that are ‘linear’ and those that are not, because then you have to specify the range over which to calibrate for the 'non-linear’ devices. That is why most all organizations choose to specify the range or points to calibrate at over the range. It keeps it simpler for them rather than dealing with linear and on-linear. Your choice."


Thanks