Relative or Combined Type A & Type B Uncertainty Calculation?

[amanbhai]

Is it neccessary in the sandard to calculate the relative unicertainty calculation? What if we calculate the uncertainties of type A & Type B but do not combined them or expand them?

[Benjamin28]

Your AB will most likely give you a minor for not calculating Expanded Measurement Uncertainty for your instrument/calibration. If you've already calculated the A B contributors it is relatively simple math to calculate expanded uncertainty even if you need to use sensitivity coefficients the guides available for UNC will give you a good clear method for calculating this, therefore there really shouldn't be any reason you can't calculate expanded uncertainty.

http://metrologyforum.tm.agilent.com/download3.shtml

Something like this may also be useful for calculating uncertainties.

[Benjamin28]

I probably should have also mentioned that ISO 17025 isn't very specific about MU, rather it mentions the requirements to calculate uncertainty and then refers the reader to ISO 5725 for further direction. If you're looking for specifics the GUM is your best bet. If you're a test lab 17025 also makes an exception for UNC calculations, recognizing that exact calculations may be prohibitive in some cases they do accept reasonable estimations. At the minimum they will want to see that you've identified the appropriate sources of uncertainty and made an effort to estimate the combined uncertainty.

[DrBobby]

Be careful of referencing the GUM manual unless you know that you are using it. As an auditor when the GUM manual is referenced, I will ask someone in the lab who is taking the data for uncertainty what the equation A-3 on page 13. It is a second order partial differential equation. It is better if you just rely on a known software calculator like UnCalc4. One must be able to explain why they are using a source and the GUM is difficult if you have a limited background in mathmatics. However, it is an excellent source that will answer many questions.

[Jim Wynne]

Originally Posted by DrBobby

Be careful of referencing the GUM manual unless you know that you are using it. As an auditor when the GUM manual is referenced, I will ask someone in the lab who is taking the data for uncertainty what the equation A-3 on page 13. It is a second order partial differential equation. It is better if you just rely on a known software calculator like UnCalc4. One must be able to explain why they are using a source and the GUM is difficult if you have a limited background in mathmatics. However, it is an excellent source that will answer many questions.

Welcome to the Cove, DrBobby

That's good advice there. Never cite that which you haven't read and/or don't understand.