Thomas J. Witt

Using the Allan variance and power spectral density to characterize dc nanovoltmeters

Stochastic serial correlations in nanovoltmeter data are often ignored and the experimental standard deviation divided by the square root of the number of observations is incorrectly used to describe the scatter. This paper demonstrates the use of power spectra and the Allan variance to characterize the performance of digital and analog DC nanovoltmeters.

The BIPM Watt Balance

The Bureau International des Poids et Mesures (BIPM) has recently started work on a watt balance to link the kilogram with the Planck constant, in view of a future redefinition of the kilogram. This paper presents the essential ideas of the BIPM watt balance and reports the progress to date

International comparisons of Josephson array voltage standards

Comparisons of 1-V Josephson array voltage standards provide a good estimate of some of the limiting factors in the uncertainties of such devices. The BIPM has participated in nine such international comparisons. This program was preceded and followed by the comparison of the two BIPM Josephson array devices. A summary of the results, which shows agreement at the level of 1-7 parts in 10/sup 10/, is presented. Some of the problems encountered are discussed. >

Report on the 1990 International Comparison of 1 Ω and 10 kΩ Resistance Standards at the BIPM

In order to evaluate the agreement among representations of the ohm based on the quantum Hall effect and to establish a firm basis for international traceability among the laboratories concerned, the BIPM, at the request of the CCE, organized and carried out an international comparison of 1 Ω and 10 kΩ resistance standards from eighteen national laboratories and from the BIPM. The weighted means of the results obtained by the ten national laboratories which, like the BIPM, based their measurements on their own QHE resistance standard differ from the BIPM results by (-0,032±0,024) parts in 106 at the 1 Ω level and by (0,009±0,022) parts in 106 at the 10 kΩ level.

Pressure coefficients of some Zener-diode based electronic voltage standards

We present results of measurements of the pressure dependence of the 1.018 V and 10 V outputs of twenty electronic voltage standards. Seventeen Fluke 732B and three Fluke 732A instruments were studied, these being the types most often used with Josephson voltage standards and in international comparisons for which the smallest uncertainties are demanded. Fifteen of the 732B instruments have relative pressure coefficients near 1.9/spl times/10/sup -9//hPa for both outputs; the other two have negative coefficients, the absolute values of which are smaller by about a factor of ten. Small, but statistically significant, coefficients of both signs were found for the three Fluke 732A instruments. The pressure effects are highly reproducible and so they permit the use of corrections to eliminate errors which, in extreme cases, can attain relative values of nearly four parts in 10/sup 7/.

The stability of some Zener-diode-based voltage standards

We report on a study, carried out from the point of view of users of Josephson array voltage standards, of factors limiting the precision of Zener-diode-based electronic voltage reference standards of the type (Fluke 732A) most widely used for Josephson voltage measurements. Frequency-dependent voltage noise was studied with a very low-frequency spectrum analyzer, and was found to be the major factor limiting short-term stability to about one part in 10/sup 8/. For two of the three reference standards tested, humidity effects were the principal factors limiting the medium- and long-term stability of the 1.018 V outputs. These effects were studied by subjecting the standards to step-function changes in relative humidity. For example, increases in ambient relative humidity of 0.01 were found to provoke decreases of 0.05-0.15 /spl mu/V in the 1.018 V output of one Zener. The time lag between the humidity change and the voltage change is characterized by a time constant /spl ges/19 days. >

Using the autocorrelation function to characterize time series of voltage measurements

The standard deviation of the mean is the most common basis for specifying the statistical uncertainty of repeated measurements, yet it is often calculated incorrectly. The variance of the mean, , of a time series of correlated measurements of a weakly stationary process is correctly expressed in terms of the autocorrelation function (ACF) at lag k, ρ(k). This approach is used to evaluate for white voltage noise measured at regular time intervals τ0 through a low pass filter of bandwidth B by four methods: (1) by developing the expression ρ(k) = exp (−4Bτ0k) evaluated by estimating B from the sample spectrum; (2) by noting that ρ(k) = k, = exp(−4Bτ0) < 1 is the ACF of a first-order autoregressive process, AR(1), for which is readily evaluated in terms of ; (3) by estimating from the sample ACF, , using the cut-off lag for an AR(1) process; and (4) by applying the general method recently proposed by Zhang (2006 Metrologia 43 S276–81), to estimate from , assuming that the data may be described by a moving average process with a cut-off lag deduced from the themselves. The values of from the four methods are in good agreement. This provides firm support to Zhangs method; this is important because of this methods wide scope of application.

Comparison of quantum Hall effect resistance standards of the NIST and the BIPM

An on-site comparison of the quantum Hall effect (QHE) resistance standards of the National Institute of Standards and Technology (NIST) and of the Bureau International des Poids et Mesures (BIPM) was made in April 1999. Measurements of a 100 0 standard in terms of the conventional value of the von Klitzing constant, RK-90, agreed to 12 parts in 101.0with a relative combined standard uncertainty 1Lc= 20 X 10-10. Measurements of 10000 0/100 nand 100 0/1 0 ratios agreed to 59 parts in 1010 with 1Lc= 55 X 10-10 and to 38 parts in 1010 with 1Lc= 31 X 10-10, respectively.

CPEM 2006 round table discussion 'Proposed changes to the SI'

This report summarizes a round table session held last July at the CPEM 2006 to discuss recently proposed redefinitions of some base units of the International System of Units (SI) based on defined values of some fundamental constants. The aim of the session was to inform CPEM delegates of the various proposals and to promote a wide discussion of the issues arising from them. An interdisciplinary panel of six experts from national metrology institutes, the academic community and the industrial metrology community briefly presented their views and their concerns. The presentations were followed by a session in which the panel answered questions and heard comments from the audience.

Allan variances and spectral densities for DC voltage measurements with polarity reversals

Using a polarity reversal technique, precise dc voltage measurements were made with both digital and analog voltmeters. The results were analyzed by calculating Allan variances and spectral densities for polarity-reversed measurements and comparing them with similar analyses of unreversed or unipolar measurements of the same processes. The Allan variances and spectra for the two measurement methods were found to agree quite well, particularly for sampling times corresponding to the 1/f noise regime for the measurement of Zener-diode voltages. The results show that polarity reversals do not affect 1/f noise of the Zeners. In polarity-reversed measurements, a rectangular voltage waveform is presented to the detector. Since this waveform can be synthesized by ac waveforms, the results also demonstrate that ac detection does not remove 1/f noise in the voltage source being measured. We also estimated the spectral density and Allan variance of some thermal (Johnson) noise processes and found that the results agree with the Nyquist expression to within 4%. This provides direct validation of our measurement and analysis methods.