Is it necessary to do MSA every time after regular calibration?

If it is a docuemnted Custoerm requirement then it is a requirement.
however in the published standards and by just plain common sense (if you know what an MSA is for) then there is no overarching requirement. even doing it once a year is uneccessarry...

 

And if it isn't, charge them!

 

Or as Andy suggested charge them more for it as it is a useless requirement. It’s just busy work.

 

First of all, MSA consists of 1) bias, 2) linearity, 3) stability, and 4) R&R. Bias and linearity (1 & 2) are already addressed by a well planned calibration system. Stability (3) is usually an issue with extremely tight tolerances where temperature is a factor, and with older electronic equipment where warm-up is an issue. I recommend that stability studies are done when the risk warrants it, not on all equipment. Finally, R&R (4) is a result of the design of the measurement system. R&R measures the spread of measurement variation which is independent of the location (bias/linearity/stability). If the design of the measurement system has not changed (no change in gage, etc.), there is no need to repeat the study.

 

That's very difficult, given that if they understood what they were asking for, they wouldn't be asking for it. To Miner and Bev's excellent descriptions, it's worth pointing out to your customer that all the act of calibration actually does is tell you how far "off" the nominal value being measured, the equipment is going to be. That might not actually change over a very long time so, for certain the results of the study won't change. If you can discover some calibration records which show no changes over the year they want the study to be done, then provide that, explain why it's not going to demonstrate anything useful and explain to the purchasing people that they will have to bear the cost of an ineffective task - which no-one else does, by the way.

 

Agree with Andy regarding the effectiveness of this conversation with your Customer. It seems someone there doesn't understand the difference and purpose of MSA and calibration.

to add to Miner's excellent points:

Calibration looks for the 'system' difference between what the gauge by itself measures and what truth is. calibration is typically done with every gauge using a traceable standard not the parts as the truth must be known in calibration. This is a very narrow look at measurement variation. Since individual gauges can get dropped, worn or otherwise 'damaged' or 'altered', calibration is typically performed on some periodic frequency, or after the operator believes the gauge may be altered.

MSA looks for variation in repeated measurements by the same operator, different operators and potentially different gauges of the same type. This is a very broad look at measurement variation that includes all elements of the measurement system and uses the actual parts and features - or properties - that are being measured. This variation typically doesn't change unless a new untrained operator starts, you take some action to deliberately improve the variation or the feature/property is changed. This is why MSA is usually only performed once and not on any periodic basis.

Think of it this way:
Calibration is the average location of the individual gauge's measurement ability - in technical terms, it's accuracy.
MSA is the variation in measurements regardless of it's location/accuracy. It is the standard deviation of the measurement system - in technical terms it is it's precision.

I find it helpful in these situations to draw the pictures (graphs) of what these two things really mean...