Text preview for : aug96pp2.pdf part of Fluke aug96pp2 Fluke appnotes aug96pp2.pdf
Back to : aug96pp2.pdf | Home
MAINTAINING 10 VDC AT 0.3 PPM OR BETTER
IN YOUR LABORATORY
Ray Kletke
Fluke Corporation
Everett, Washington
Abstract
There is a need today to maintain dc voltage at an uncertainty of 0.3 ppm or better. We
believe that most standards laboratories can do this without using an expensive Josephson
Junction Array System. DC Reference Standards commonly in use today are capable of
this performance providing that they are properly calibrated and that the effects of
temperature, pressure and seasons are taken into account. This paper develops supporting
uncertainty equations and applies them to useful scenarios.
Introduction
High end multifunction calibrators and digital voltmeters today require calibration
uncertainties that only primary standards laboratories maintained in recent history.
Uncertainties of 1.0 to 1.5 ppm are required for 10 VDC at the time of test. These
standards, in turn, must be supported by reference standards having a NIST traceable
uncertainty of 0.3 to 0.5 ppm if a reasonable Test Uncertainty Ratio (TUR) is to be
maintained.
It is possible today to maintain 10 VDC at 0.3 ppm or better in most standards
laboratories without a Josephson Junction Array. However, care must be taken to
minimize the effects of certain stimuli that usually contributes negligible error and,
therefore, is overlooked. This paper examines some of those effects and recommends
how they can be controlled so as to achieve 0.3 ppm performance or better.
Classical Approach
The classical approach for estimating the uncertainty of the dc voltage standard is simply
to combine its stability, as specified by the manufacturer, with the uncertainty of the
calibration as follows:
= stab + U
2 2
U tot cal
Modern zener type dc voltage standards typically have a stability of 2 ppm per year.
Assuming this value and a calibration uncertainty of 0.1 ppm, the total uncertainty can be
calculated as a function of time. Figure 1 gives the total uncertainty for a single cell and
triple cell dc voltage standard (DCVS) as a function of its calibration cycle in months. It
shows that the 0.3 ppm goal can be approached for the singe cell DCVS only if it is
calibrated with a Josephson Junction Array System every two months or less. This may
1 .2 0
1 .0 0
0 .8 0
ppm
0 .6 0
0 .4 0
3 C e ll G r o u p
0 .2 0
S in gle C e ll
0 .0 0
2 2 .5 3 3 .5 4 4 .5 5 5 .5 6
C alib r atio n C ycle in M o n th s
Figure 1. Total Uncertainty for DCVS Array Voltage Standard
or may not be practical depending on the turn-around time of the calibration supplier. If
the number of independent cells is increased to three, the required calibration cycle time
to maintain 0.3 ppm is increased to 3 months. Although this may be workable, it is
inefficient and costly, requiring a total of 12 calibrations per year (4 on each cell).
Characterized Performance
The largest component of uncertainty in the above calculations is the stability of the
standard. You might postulate that the typical performance for a standard must be better
than its specifications since the manufacturer must achieve a high yield in his
manufacturing process. This in fact is true. Therefore, if the standard is characterized
using historical calibration data, its actual performance will usually be better than its
stability specifications.
A linear regression model is frequently used for this characterization. Historical
calibration data is used to calculate the parameters of the regression line and the
components of the estimated regression uncertainty. The output voltage of the standard
and its uncertainty is given by the following equation:
V std
= a + bX