Is there a rule of thumb to follow when choosing a calibrator? 2

[davidd314152wind4]

Is there a specific accuracy or other specification one should strive for?

For instance, if calibrating a DMM how much more accurate must the calibration meter be?

Windo

 

[IRstuff]

The traditional MIL-STD-52B benchmark was 10:1 test accuracy ratio (TAR), but more recently, the trend has been a 4:1 TAR. This makes a reasonable compromise, since independent calibrator errors would therefore would only contribute about 3% to device's measurement error.

The closer you get to require device accuracies comparable to NIST calibration limits, the lower the TAR that you can achieve.

TTFN

FAQ731-376

 

[logbook]

TUR (Test Uncertainty Ratio) in modern terminology.
If you use a TUR of 10:1 you tend to ignore the uncertainty in the calibrator completely. At 4:1 you may then want to consider "guard banding". there are some excellent articles on the web on this subject. Google "TUR" and "guard band".