Maximum Permissible Error for Monitoring and Measurement Devices

M

mickng

Hi,

I read the article before which related to Permissible error for monitoring and measurement devices. However, I can't find this article anymore...

I was told there are few ways to determine the permissible for the devices;

1) Follow the testing standard recommendation ( if there is available ) or
2) Follow supplier recommendation ( if the salesman is technical enough)
3) Rule of thumb - own calculation based on 1/3 of the product's tolerance.

ISO10012:2003 saying that the permissible might determined by metrological function, then does the above method acceptable?

I was learn that previous standard stated 1/10, then 1/4 and latest recommendation is 1/3, am I right?

Anyone can offer a help to solve my doubt..

TQ

Mick:truce:
 

BradM

Leader
Admin
Well, hello there!

I'm really not a subject matter expert, but I'll give it a stab.

When you state permissible error, to what are you comparing? If you are talking about how much error an instrument can have, that will depend on what calibration tolerances you establish for that particular instrument.

Or, do you have in mind some comparison of that instrument tolerance to some specification within the process?

As far as what type of accuracy your instrument needs, a general rule of thumb is 4-1. NOTE: That's general, and there are many factors that need to be considered. But in general, the accuracy of your measuring instrument should be four times greater than the process requirements.

Not sure if that helps.
 

harry

Trusted Information Resource
Welcome Mick,

If I am not mistaken, these are rules of thumb.
1:10 normally refers to resolution of instrument
1:4 is what Brad mentioned earlier
1:3 refers to tolerance and uncertainty. Tolerance should be greater than 3 times uncertainty.

Can any fellow covers here confirm this or point us in the right direction.

Thanks in advance.
 
M

M Greenaway

Harry

Those rules of thumb are ok as a broadbrush guideline, however the only way to determine if a piece of measuring equipment (or indeed a measurement process) is good enough is to conduct MSA study. Certainly this must be done on critical or safety characteristics.
 

harry

Trusted Information Resource
Harry

Those rules of thumb are ok as a broadbrush guideline, however the only way to determine if a piece of measuring equipment (or indeed a measurement process) is good enough is to conduct MSA study. Certainly this must be done on critical or safety characteristics.

Thanks. I am in complete agreement with your advice especially after some experience with TS implementation.
 
M

mickng

Hi,

I managed to find MIL-STD-45662A, found out that the previous recomendation is 25% of the product tolerance. I am still looking for ISO10012-1,-2:1998 Version which I read before that indicate the 1:3 of the product tolerance.

I agreed of MSA but it does not indicate much on accuracy requirement which is the permissible error of the devices, it focus more in precision study. The question is if TS auditor ask on how to read the calibration certification, then auditee need to answer how he compared the measurement uncertainty with permissible error. MSA indicates that we need calibrate our device prior GR&R study...right.

Anyway I have both MIL-STD-45662A and ISO10012:2003, let me know if anyone interested..

TQ

:thanx:
 
P

potdar

MSA has to be conducted on calibrated instruments only. It encompasses too many variables within the results.

The shop floor, quality and calibration people use different language. I shall therefore use the schoolboy language and define what I say - at a risk of sounding too elementary. These are thumb rules that the standards will overrule.

Product / dimension tolerance limit

MMD least count - prefarably 1/10 of tolerance, 1/4 of tolerance max.

Calibration error - Difference between standard value being measured and measurement shown by the MMD.

Acceptable calibration error - defined by the instrument standard or 1/3 of least count max.

Comparison between actual calibration error and acceptable calibration error will decide whether the the MMD is 'in calibration'.

I have examples of my clients using some 'calibrated' vernier callipers in their foundry division. These were considered out of calibration by their machining division located in the same compound.
 
M

M Greenaway

Doesnt MSA also indicate bias, which is roughly speaking what 'calibration' is ??
 
Top Bottom