[volt-nuts] JJ standards and the 3458A

Frank Stellmach frank.stellmach at freenet.de
Mon Jan 20 02:00:57 EST 2014


JJ standards are always primary ones, i.e. don't need to be calibrated.

Preconditiion is only, that the setup is deigned correctly, mainly that 
the JJ voltage on cryongenic level is transferred to room temperature 
level with high accuracy.

The JJ voltage is uncertain to 10e-17 on cryogenic level, i.e. on 
comparing/transferring from one JJ to another by means of a SQUID.
Transfer to RT is limited by thermal voltages to around 10e-9 uncertainty.

That's also the practical limit in the real, macroscopic world.

A 9.5 digit macroscopic or analogusly made DMM is unimaginable due to 
following reasons:
1. Stability of Zener references are limited to around 0.1ppm in time, 
temperature and noise
2. Same goes for the analogue range amplifiers, based on resistors.
3. The 3458A still has got the best linear A/D, around 0.02ppm of input.
4. Range transfer to the 5 Cardinal Points 1kV, 100V, 10V, 1V, 100mV is 
also limited to around 0.2 - 0.5ppm (see Fluke 752A)

A JJ array could not be designed into an analogue DMM, due to the fact, 
that the setup needs cryogenic devices as dewars, and lHe4 or lN2. 
Microwave generator, frequency and constant current source can be made 
compact, as there already exist "portable" JJ array setups for 
comparison between primary labs. (Those comparisons always were precise 
and limited to 10e-9 uncertainty).

Anyhow, as the Volt still is uncertain to about 2*10e-7 in the SI 
system, a 9.5 digit DMM currently makes no sense at all.
Let's see, if the BIPM will decide on the new SI by end of this year or 
in 2 / 4 years only.

Then, I would propose to build a 9 digit DMM, but based exclusively on 
cryogenic electronics, i.e. a JJ array as an ultra linear D/A and as a 
primary Volt standard in one.

The only unsolved problem then would again be the range transfer, which 
is not yet possible by cryogenic standards, afaik.

Similar limitations apply also to an quatum Ohm standrad / DMM, as the 
Hall effect standard requires several Teslas of magnetic field, and 
cryogenics, which cannot be built into an analoguos DMM, and also the 
Ohm transfer to analogue resistors is again limited to 10e-9 by thermal 


More information about the volt-nuts mailing list