TUR (test uncertainty ratio) based on Standard or Expanded Measurement Uncertainty?

[ScottBP]

Hey gang,

Simple question with an important impact I'm having a hard time finding an answer to: Is the Test Uncertainty Ratio (T.U.R.) based on the standard uncertainty, or the expanded uncertainty? For example, if I calculate a 1 volt source to have a ±2 ppm standard uncertainty and my meter has a ±9 ppm tolerance, the T.U.R. is a nice 4.5. However, if the uncertainty is expanded to k=2 (4 ppm), then the T.U.R. drops to a measly 2.25. Which is right? If I need to calculate a guardbanding factor because of a low T.U.R., how do I do that?

[Hershal]

I believe ANSI/NCSL Z540.3-2006 Clause 5.3 uses expanded uncertainty.....

However, that approach is two-tailed.....so twice the expanded uncertainty compared to twice the tolerance.....so long as tolerance is twice the uncertainty you have 4:1.....

I will post the specific language a bit later, I don't feel like digging it out right now.....

Now, if you have an extra couple of hundred bucks you can buy the Z540.3 Handbook which is finally out (only 2 years and 8 months after the standard).....it tells you how to do guardbanding to do the 2% false accept....

Hope this helps.....

[ScottBP]

In Reply to Parent Post by Hershal

Now, if you have an extra couple of hundred bucks you can buy the Z540.3 Handbook which is finally out (only 2 years and 8 months after the standard).....it tells you how to do guardbanding to do the 2% false accept....

Is it the same one available for download at the NCSLI store for $90, or did someone else write a companion handbook? (A search for Z540.3 at the ANSI store turns up zero.)