[volt-nuts] Fluke 720A Kelvin Varley Divider Questions

[Dr. Frank Stellmach]

>Hi to everyone,

>I am new to this group and have a keen interest in precision and accurate voltage references and metrology.  I am using a pair of Fluke 732A Reference Voltage Sources for my voltage standards and a Fluke 5440B Direct Voltage Calibrator in my home lab. I am thinking about acquiring a Kelvin Varley voltage divider for a number of lab uses including scaling of my 10 volt standards so that I can calibrate my Fluke 5440B and, in turn, various DMMs.  The most precision DMM I own is an HP 3457A 6 ½ digit multimeter which can be extended to 7 ½ digits by adding the contents of the 7th digit register to the displayed reading.

>My question is how stable are the Fluke 720A dividers over time, i.e., if I have one that is in calibration (+/- 0.1 ppm accuracy and linearity) will it tend to stay that way over the course of years assuming that I don't apply excessive voltage, dissipate excessive power, or otherwise abuse the divider?  Or will it need to be calibrated on an annual basis to keep it within reasonable specifications?  I don't require +/- 0.1 ppm accuracy but I would like to maintain somewhere between +/- 0.5 ppm accuracy and  +/- 1ppm accuracy over time.

>All advice and comments are welcome.  For me, this is a new area of learning and adventure.

>Thanks.

>Charlie



Hallo, Charlie

welcome, I think you share exactly the same keen interest as the others on this mailing list.


I think, that for the calibration of a 5440B, you don't need and can't use a 720A.

The 5440B, according to its cal manual, requires the 5 Cardinal Points 100mV, 1V, 10V, 100V and 1000V, precisley related to your 732As 10V.
The 720A is not precise enough especially for the high voltage transfer 1000V =>  10V, maybe also not precise enough, or "on the edge" for 100:1 and 10:1 transfers, as required for 100V, 1V and 100mV.
Check 5440B vs. 720A specifications in the different voltage ranges on that aspect.

The 752A reference divider is a much better instrument for calibrating the 5440B, required also in the cal procedure of the 5440,
but the 720A can well be used for checking its linearity, although this can not be calibrated.

I own a 5442A (mainly less precise internal reference than 5440B ), and I studied the functional description very thoroughly.
I think, the self cal function is fully sufficient for calibration of the 1kV, 100V ranges, if the 10V range is externally calibrated against a 732A before.

The manual is not very exact at this point; why a separate external calibration (by means of a 752A) should be necessary, ie what benefit this gives over the internal autocal, is not totally clear.
The internal switching of the range resistors, additionally thermally stabilized, does exactly the same job as the 752A.

I think, that Fluke did not  have the heart at that time to promote a full auto cal instrument, as they did for the later 5700/20A.
The 3458A then was the only instrument with full autocal function, which was accepted by the "metrological community".

Read also the Fluke articles about artefact calibration, explaining this acceptance problem at that time.

http://assets.fluke.com/appnotes/Calibration/dbmsc98.pdf
http://support.fluke.com/calibration-sales/Download/Asset/2114953_6200_ENG_A_W.pdf

On the 5440, 1V and 100mV cannot be auto-calibrated, as an additional fixed  10:1 / 100:1 divider is implemented.

Therefore you need a reference divider for that only, or a 3458A.


Anyhow, if you can get a 720A, together with a 845, it would be a nice gadget for your metrology interest.


The 419A is also a nice, stable Null Voltmeter.
Both instruments use similar optical chopper amplification (influence of spikes?).
I can fingerpoint only that for both instruments, the leakage current is not specified, although this would be the most important parameter for null /bridge measurements.
So I cannot judge if the 419A is better or worse than the 845A.

That's perhaps also a question of the acceptance of the 845 in the metrological community, and HP did not want to get in the way of Fluke in the business of analogue metrology.


regards - Frank