[volt-nuts] Basic question concerning voltage references

acbern at gmx.de acbern at gmx.de
Tue Feb 21 05:34:51 EST 2017

If you want to do a comparison measurement with consideration of the actual measurement uncertainty, then what is relevant is not the theoretical resolution of the meter but what the meter data sheet tells you with respect of that matter. It is ususlly called the transfer accuracy. Applies when you do two measurements with one meter, at similar voltages. E.g. the drift of a reference after it has been exposed to a temperature change. It takes into account also things like differential non-linearity of the A/D, short term noise/flicker/drift issues and so on.
For a 3458A at 10V, it is 0.05+0.05ppm (range and reading), so 0,1ppm (0.2ppm for two measurements in total). Now for 6.5 digit DMMs it is obviously less, and not always specified. As an alternative, if the value is not given, one can take the uncertainty referenced to the range (see Fluke 'Calibration: Philosphy in Practice'). The 34401 has the A/D linearity specified, it is 3ppm. That helps as well. No way arround analyzing such transfer measurement uncertainty for any given meter based on the data sheet values.


More information about the volt-nuts mailing list