PREDICTING 10 VDC REFERENCE STANDARD OUTPUT VOLTAGE

摘要

Zener type dc voltage standards (DCVS) today serve as the primary dc voltage reference in many standards and calibration laboratories of the world. Originally designed to function as working or transfer standards in support of high-end calibrators and voltmeters, their excellent predictability, good stability and low noise characteristics are now recognized as suitable for maintaining the dc volt in primary laboratories. In recent times, they have been shown capable of transferring the volt from the standards laboratory to the customer's site with accredited uncertainties of 0.1 ppm or better. This remarkable capability is consistent with the reports of interlab comparisons of Josephson array system operators claiming a 2 sigma scatter of less than 0.03 ppm between laboratories. Most metrologists agree, I think, that the zener DCVS is very well suited for establishing and transferring the volt. There is less agreement, however, on their capability to maintain the volt, especially at the fractional ppm level. Maintaining the volt requires either good stability or, at least, good predictability. Metrologist prefer the former. The bad news is that zener type dc standards tend to drift. They drift significantly more than saturated standard cells. The good news is that their drift is reasonably predictable. Perhaps the biggest metrology problem in modeling and predicting future values, is the estimation of their uncertainties. This problem was addressed in a paper delivered at NCSL in 1996. This current paper summarizes the zener model developed in the NCSL paper and describes a study conducted to test the validity of that model.