[volt-nuts] Calibration of voltage standards

frank.stellmach at freenet.de frank.stellmach at freenet.de
Wed Feb 11 04:36:55 EST 2015


Hello Ken,
 
generally, the difference (bridge) method is always the most sensitive and precise one.
 
Thermal voltages and amplifier noise pose a lower limit to the achievable precisison.
 
About 10nV may be achieved at best, even in metrological environment.. 
 
That's the reason, why the 1e-18 stability / uncertainty between two JJ arrays against each other, (superconducting environment), can be transferred to the analogue world to 1e-9 uncertainty only (10nV / 10V).
 
 
 
To compare two 10V standards, an uncertainty of only 1µV is required, as this will yield 0.1ppm uncertainty of the 10V transfer - totally sufficient in practical terms. Btw.: the unit 'Volt' currently is 0.2ppm uncertain only in the SI system. (Will probably change in 2018!) 
 
Good low thermo voltage cables are required, maybe also a low thermal switch.
 
 
Therefore, both of your instruments can of course successfully and easily compare and null these two 10V references, and would also be able to give a precise difference reading, so to assign the DUT a calibrated standrad value, differing from a clean 10.000000V.
 
The 34420 is better suited, as it has better noise and higher resolution (100pV).
This instrument can even be used to make a transfer from a JJ array primary standard to a 732B secondary standard.
 
Rem.: There's a good application note from FLUKE, where they promote their 8508A "reference" multimeter as good replacement for differential meters like their famous 845A...
 
 
 
The 3458A on the other hand is capable of making ultra precise absolute value transfers (better than the 720A), i.e. you may measure the absolute value of your 732B reference standard, and then measure the DUT 732B also on the 3458A, to compare their values in absolute terms.
 
This transfer can also be made uncertain to < 0.1ppm, good thermal connections and maybe thermo voltage cancellation provided.
This again is equivalent to < 1µV level uncertainty, which is not so difficult to achieve.
 
Frank


---
Alle Postfächer an einem Ort. Jetzt wechseln und E-Mail-Adresse mitnehmen! Rundum glücklich mit freenetMail


More information about the volt-nuts mailing list