Text preview for : deaver_msc01.pdf part of Fluke deaver msc01 Fluke appnotes deaver_msc01.pdf
Back to : deaver_msc01.pdf | Home
PREDICTABILITY OF SOLID STATE ZENER REFERENCES
David Deaver
Fluke Corporation
PO Box 9090
Everett, WA 98206
425-446-6434
[email protected]
Abstract - With the advent of ISO/IEC 17025 and the growth in laboratory accreditation, more complete, detailed
uncertainty analyses are required. This often points out the need for improved accuracy starting from the top of the
traceability chain of a calibration laboratory. This paper describes a program that has been in place for several years
to provide better uncertainties for solid state 10V zener references. This program is an on-site calibration service that
can model a reference's drift performance resulting in a very predictable projected value. An evaluation of the
accuracy of the predictions is presented as well as a description of the tools used to make the predictions.
INTRODUCTION
Prior to the development of the solid-state zener reference standard (DCVS), standard cells were used by most
laboratories to maintain the volt. Pampered, they are quiet and have very predictable drift characteristics. However, they
are poor travelers making it very difficult to send them for calibration or to use them to deliver accurate voltages to a
production line or remote site. Zener references were conceived as a means to easily transfer dc voltage anywhere that it
was needed. For manufacturers of precision calibrators, the need was to deliver 10V accurate to about 1-1.5 parts in 106.
Soon, many laboratories wanted to be able to maintain their own zener references so calibration uncertainties were
improved to about a part in 107. With multiple standards, some history and the use of regression models, 0.3 parts in 106
can be maintained with annual calibrations [1]. With further characterization and care, zener references can be used for
intercomparisons between Josephson Arrays with transfer uncertainties of about a part in 108 [2].
Though zeners travel much better than standard cells, the demand for increasingly better uncertainties also increased the
reluctance of some owners to allow them to do so. Many of them use an on-site calibration service that uses a well-
characterized zener reference with demonstrated ability to travel well. A proficiency test is performed, then the
calibration using a procedure from the calibration service provider. This process was accredited in 1995 by NVLAP and
in 1998 by DKD, the German accreditation body. After calibration, users employ a number of methods to re-assign the
value throughout the calibration interval. Some use the calibrated value throughout the interval and increase the
uncertainty by the drift specification. This works well if the more accurate workload can be scheduled early in the
calibration period. If the uncertainty demands aren't as great, a simpler approach is to increase the uncertainty for the
maximum drift immediately and use the same uncertainty as well as assigned value throughout the interval. A more
sophisticated approach is to use a drift model for the reference and change the assigned value and uncertainty during the
calibration interval. To execute this strategy, one must have some previous calibration history for the unit, knowledge of
the behavior of similar devices, a mathematical model to apply, the ability to perform the drift and uncertainty
calculations and create a table of assigned values and uncertainties to be used throughout the calibration interval. As the
manufacturer, we were in the best position to look at large populations of the references to develop the models and
evaluate the results. In 1996, we started supplying projections based on linear and non-linear drift models upon the third
calibration event. In 1998, a study was performed [3] to evaluate the reliability of those projections. This paper repeats
the study for a much larger population, separates the results for 732A and 732B references and provides drift data for the
populations.
LINEAR AND NON-LINEAR DRIFT MODELS
A linear regression model for the zener standard is calculated using historical calibration data. The output voltage of the
standard is estimated by the following equation:
2
2 1
V = ax + b