[volt-nuts] Poll

Dr. Frank Stellmach drfrank.stellmach at freenet.de
Thu Dec 31 10:34:40 UTC 2009


As volt standards, I own a Fluke 5442A, and a home made double LTZ1000 
standard, which has still to be improved.
I got an HP34401A, and a 3458A since 1/2 year.

All instruments never got a calibration, when in my ownership.

Comparing 5442A vs. LTZ1000 vs. HP3458A during the last 1/2 year, I 
could not find an appreciable drift at the 1-ppm-level.

Currently, I search for the root cause of voltage spikes in the output 
of the LTZ1000 references (suppose shielding/grounding issues).

After solving, I would like to have the LTZ1000 reference been 
characterized/calibrated in a 2nd order lab, thats called DKD here 
(German Calibration Service). PTB would be too costly.

DKD labs mainly fall into 3 categories for Volt (see www.dkd.eu):
- 4-6ppm, obviously Fluke 5720A based
- 1ppm, obviously Fluke 734 / 7010 based
- sub 1ppm: Josephson based, Fluke Corp. only (agilents accreditation 
here in Germany has been deferred!)

Therefore, only few labs in Germany with 1ppm or less level may be 
capable to characterize my LTZ1000 standard, or to calibrate my 3458A. 
Let's see how expensive it will be..

Compared to U.S., a problem in Germany is to get the 'good' cal 
standards at all, and to get them for a reasonable price.

For linearity testing/calibration (e.g. of a Fluke 332A), I designed a 
precision one-decade divider, based on 11x Z201 resistors with 
Wheatstone Bridge, diff. linearity is less than 1ppm. This device 
replaces a Fluke 720A.

For the 5 Cardinal Points I designed a precision 10:1 / 100:1 divider.

It's built with 100x TK 0.3 metal foil resistors, and uses the Fluke 
752A self-calibration scheme.
Accuracy / stability of divider ratios is around 1-2ppm by design, and 
also measured with HP3458A and Fluke 5542A on 100V / 10V input.

1000V to 10V transfer accuracy is better than 4ppm, but Fluke's 5442A 
internal transfer and HP3458A 10M divider are currently the biggest 
error source.
This has to be further investigated by measuring the initial short-term 
output drift.

Resistance is maintained by a set of 8x 40 years old HP ww. resistors, 
TK1, from 10 Ohm to 1 MOhm.
They have been compared against calibrated 3458A's during 15 years, and 
they showed drifts below 5 ppm.

I also would like to build a VHP202Z/VHA512Z based resistor standard 
instead, but before, I have to investigate on capable cal labs..

I built a precision DC-current (1µA-100mA, < 10...30ppm) during my time 
at university for precise temperature measurements.

ACV is maintained by a Fluke 5200A which is partly calibrated by the 
3458A using Swerleins Algorithm.

My intension is to maintain Voltage, Resistance, Current (and also Time) 
as best and cost efficient as possible - just for obsession.

I also like to design and test all kind of home made standards, to 
reproduce the ancient metrology techniques.


More information about the volt-nuts mailing list