Author Topic: #1000: My (hi)story of the Weston cell, of the Volt, and of being a volt-nuts  (Read 17797 times)

0 Members and 1 Guest are viewing this topic.

Online Dr. Frank

  • Super Contributor
  • ***
  • Posts: 1697
  • Country: de
My collection of Weston cells, all salvaged from the dumpster, started in 1982.
That's marking my personal emergence as a volt-nuts, as for the first time, I collected Volt artefact fix-points, down to the ppm range uncertainty.

These were first measured in 1982, both to 5 1/2 and 6 1/2 digits, about 1.01882(4)V.
This was done by a Fluke 8502A, which was calibrated daily or weekly by means of 335D, 720A and an ovenized Standard Cell, which were frequently calibrated against a 2mV Josephson Junction, from a lab being located just across the street.
So these values were uncertain to about 10ppm, or better.




These unsaturated types, wrapped in tin foil, are from scrapped FLUKE 803B differential VTVMs.
They were used to calibrate the internal 500V reference.



The first one is from Eppley, now reading 1.01795V, that's a rate of about -25ppm/year.



The one on the left side in their box, is a Muirhead 845-C from 1961. Hello - that's my own vintage!
It still reads 1.01798V, about -1295ppm from nominal 1.0193V over 55 years.






A historical look at the Volt (*):
Weston cells defined the Volt until 1948, when the Absolute Volt was defined, and from then on served as mise-en-pratique (representation) for the Volt only.
The SI-Volt was defined in 1960, just before all these cells were made.
This SI-1960-Volt definition is still valid, up to date, and therefore still uncertain in the SI to about 0.3ppm, currently one of the worst defined units.



As there was no precise way at that time to do it better, all main National Standards Institutes had their own Volt, still based on the Weston cells, and the very difficult comparisons between the countries showed differences up to 20ppm.
In fact, there was the American Volt, the German Volt, the British Volt, and so on.

Between 1968 and 1972, the mise-en-pratique changed to using the single Josephson Junction, and Weston cells were used only as working standards.
The Volt at this time was defined by a Josephson constant of KJ(72) = 483594 GHz, so the absolute level of the worldwide Volt was reduced by about -10ppm.
As this JJ Volt was still difficult to transfer to 1V or 10V level, obviously there were still differences between the national labs, still on the order of a few ppm.
The German PTB claimed to be uncertain to 4*10-8, though.

The next time the cells were measured, was on 14th November 1989, another remarkable date in several aspects.

At first, the values were now determined to 8 1/2 digits, so that was obviously the first day, when I got a brand new 3458A at university.

Secondly, on 1st of January 1990, the SI Volt definition was changed again, exactly the day, when I started my PhD physics thesis.
The mise-en-pratique now was defined by the 10V Josephson Junction array, giving about 1000 times better uncertainty worldwide, from 10-6 to 10-9 level.
Coincidentially, the Josephson constant, was increased by about 8ppm, KJ(90) = 483597.9 GHz, bringing the Volt back to nearly the former level.



For my 2 month old instrument this meant, that it lost its calibration to SI Volt over night; the American Volt had to be increased by exactly 9.264ppm, so that my instrument was now reading about 9ppm high. https://www.nist.gov/sites/default/files/documents/calibrations/tn1263.pdf

This didn't matter for my work, as I made relative volt measurements, at much lower uncertainty requirements, only.

Then I salvaged 3 Muirhead 845-D Reference Cells from about the 1970ties, out of some dead gear.
These are also still fine, the one in the middle only decreased about -360ppm over maybe 45 years.




The saturated Standard Cell from 1965 also was measured first in November 1989, 1.01857420V @ 22.5°C, which gives U20°C = 1.018672V (SI-90), and 26 years later about U20°C=1.018639V, so that would be a drift of about -1.24ppm/ year.

           

I'm awaiting impatiently July of 2017. This will be the Due Date for the final measurements of the Watt-balances in U.S., Canada and UK, and the German Avogadro experiment.
Succesfully determining Plancks constant h within 2*10-8 uncertainty for one, and 5*10-8 between different experiments, will be the pre-condition for the next new definition of the SI-2018.
The definition of the kilogram, the mol and the Kelvin will be changed, and also will finally re-define the SI-Volt.
It will then have zero uncertainty by definition of the quantum Josephson Effect, with h and e having zero uncertainty. This will also be identical to its mise-en-pratique, at about 10-10 uncertainty level to real world, and within 10-19 @ cryogenic world.

I think, at that moment, we'll have to re-write Volt-Nuts, in capital letters.

Frank

(*) taken from:
"Volt Metrology: the Josephson Effect and SIS Junctions Arrays", B. Jeanneret
https://home.zhaw.ch/~spma/Scripts/Messtechnik/MT/Unterlagen_Lit&Co/VoltMetrology.pdf
« Last Edit: October 03, 2016, 01:11:13 pm by Dr. Frank »
 

Offline EEVblog

  • Administrator
  • *****
  • Posts: 29700
  • Country: au
    • EEVblog
Great 1000th post!  :-+
 

Offline CatalinaWOW

  • Super Contributor
  • ***
  • Posts: 3242
  • Country: us
Good read, even for us non Volt Nuts. 

I found it interesting that your three salvaged Muirhead cells have such a small serial number range.  Obviously the instruments they were salvaged from were purchased in a single lot.
 

Offline HighVoltage

  • Super Contributor
  • ***
  • Posts: 4252
  • Country: de
Your deep understanding of this subject is impressive.
What a great collection of Weston cells.

That is probably the reason, the Fluke 732 standards have still a 1.018 V output today?
There are 3 kinds of people in this world, those who can count and those who can not.
 

Offline Vgkid

  • Super Contributor
  • ***
  • Posts: 2507
  • Country: us
Nice 100th post. It has been a while since I have looked into the historical aspects of the volt.
If you own any North Hills Electronics gear, message me. L&N Fan
 

Offline TiN

  • Super Contributor
  • ***
  • Posts: 4045
  • Country: tw
  • xDevs.com/live - 24/7 lab feed
    • xDevs.com
Vote for "Volt-Nut" status for Frank  >:D

Great read and pleasure to see photos.  :-+
YouTube | Chat room | Live-cam | Have documentation to share? Upload here! No size limit, firmware dumps, photos.
 
The following users thanked this post: SeanB, wim.de.jong.59

Offline VintageNut

  • Frequent Contributor
  • **
  • Posts: 521
  • Country: 00
You have very cool toys and you know how to use them!

Your collection of voltage cells makes me jealous as well as your 3458A.

It is a pleasure to read your well-written and well reasoned posts.

working instruments :Keithley 260,261,2750,7708, 2000 (calibrated), 2015, 236, 237, 238, 147, 220,  Rigol DG1032  PAR Model 128 Lock-In amplifier, Fluke 332A, Gen Res 4107 KVD, 4107D KVD, Fluke 731B X2 (calibrated), Fluke 5450A (calibrated)
 

Offline Jeff_Birt

  • Regular Contributor
  • *
  • Posts: 193
  • Country: us
Dr. Frank,

Recently when searching for some manganin wire to rewire the 1K stage of an old Muirhead decade box I found a paper by R. Dzuiba and J.F. Mayo-Wells on the history of the manganin resistance standards used by NIST from 1901 until 1990. This paper mentioned that the Josephson junction array is now used (to calibrate manganin transfer standards). This got me to thinking about where all of these standards come from and how one can be specified without making reference to another, i.e. specifying one unknown by way of other unknowns. Personally I think this would make a fascinating mini-series on Discovery, etc. Thanks for the link to the NIST tn1263, I will attempt to digest it.
 

Offline quarks

  • Frequent Contributor
  • **
  • Posts: 821
  • Country: de
Very interesting. Thanks for sharing.
« Last Edit: October 05, 2016, 05:16:36 am by quarks »
 

Offline HighVoltage

  • Super Contributor
  • ***
  • Posts: 4252
  • Country: de
Dr Frank,

After reading your great post, I could not resist and got my first Weston Cell.
It was made by Cambridge Instrument Co. Ltd. in England

When you measure your Weston Cells are you using the multimeter in the 10 MOhm input impedance setting?
The load on to the chemical cell most likely will have an influence on the output voltage.

But much older multimeters had far less than 10 MOhm resistance, so I wonder what current can a Weston Cell supply,
without having a significant voltage drop.

I am measuring about 1.018, 310V but with 10 MOHm at the input of the 34461A I see a constant drop.



 
There are 3 kinds of people in this world, those who can count and those who can not.
 

Online dr.diesel

  • Super Contributor
  • ***
  • Posts: 2124
  • Country: us
  • Cramming the magic smoke back in...
Super Cool, thanks for sharing!    :-+

Online Dr. Frank

  • Super Contributor
  • ***
  • Posts: 1697
  • Country: de
Dr Frank,

After reading your great post, I could not resist and got my first Weston Cell.
It was made by Cambridge Instrument Co. Ltd. in England

When you measure your Weston Cells are you using the multimeter in the 10 MOhm input impedance setting?
The load on to the chemical cell most likely will have an influence on the output voltage.

But much older multimeters had far less than 10 MOhm resistance, so I wonder what current can a Weston Cell supply,
without having a significant voltage drop.

I am measuring about 1.018, 310V but with 10 MOHm at the input of the 34461A I see a constant drop.

Nice cell, I assume, it's a saturated standard cell.

These should not be loaded at all. Therefore, these measurements were always made in differential mode, Null (e.g. Fluke 845A) against an equal comparison voltage, (e.g. Fluke 750A or 720A), giving virtually indefinite high resistance, and that was possible even many decades ago.

Nowadays, the low bias of modern bench DMMs is sufficiently low; 10MOhm impedance must not be used.

They may survive short loading with nA, even a few µA for short time, but this means, it takes time for the cell to recover, but maybe the change is irreversible.

Frank
« Last Edit: October 12, 2016, 08:08:29 pm by Dr. Frank »
 

Offline tautech

  • Super Contributor
  • ***
  • Posts: 16019
  • Country: nz
  • Taupaki Technologies Ltd. NZ Siglent Distributor
    • Taupaki Technologies Ltd.
So just for clarity are DMM's with 10G input impedance OK to measure these Weston cells ?
Avid Rabid Hobbyist
 

Online Dr. Frank

  • Super Contributor
  • ***
  • Posts: 1697
  • Country: de
So just for clarity are DMM's with 10G input impedance OK to measure these Weston cells ?

I don't like at all this characterization ".. 10 GOhm.. impedance.."

These instruments are better characterized by their bias current.

A low bias current of several ten pA on the 1V or 10V range therefore is allowed for standard cells .

PS: These "10 GOhm" are totally misleading. These were somehow theoretically calculated by dividing the voltage range by the bias current, or if you would do it mathematically correct,  by calculating the differential dU/dI .
This bias current in first order is constant over the input voltage, so unreasonably high resistances might occur, or even negative ones, if the bias current decreases with increasing voltage.

Frank
 
The following users thanked this post: SeanB, cellularmitosis, tautech

Offline HighVoltage

  • Super Contributor
  • ***
  • Posts: 4252
  • Country: de
Well, luckily I had the 10 MOHm multimeter only hooked up shortly, so hopefully the cell was not damaged.

At a high-Z setting for the 34461A, the voltage of the cell is very stable and has no drop.
So, may be it is OK to have a high resistance multimeter or an Electrometer hooked up in parallel to the Weston Cell?

Looking at your pictures, it looks like you have the 3458A hooked up directly to your cells.
May be not?


 
There are 3 kinds of people in this world, those who can count and those who can not.
 

Online Dr. Frank

  • Super Contributor
  • ***
  • Posts: 1697
  • Country: de
Well, luckily I had the 10 MOHm multimeter only hooked up shortly, so hopefully the cell was not damaged.

At a high-Z setting for the 34461A, the voltage of the cell is very stable and has no drop.
So, may be it is OK to have a high resistance multimeter or an Electrometer hooked up in parallel to the Weston Cell?

Looking at your pictures, it looks like you have the 3458A hooked up directly to your cells.
May be not?

10MOhm create 100nA current, that should not damage the cell if you loaded for only a few seconds.
The voltage drop is  caused by the about 200..600 Ohm internal resistance of these cells.

Of course I used HiZ for the 3458A, when I measured the cells. Its < 20pA bias do not harm the cells at all.
 
The following users thanked this post: eplpwr

Online Macbeth

  • Super Contributor
  • ***
  • Posts: 2521
  • Country: gb
The best #1000 post ever!

I have to admit I restrict myself to 6.5 digits in the volt nut stuff. I am tempted further, but then one of my dogs pass by and wags their tail and there is a very noticeable effect immediately visible on my Keithley 2000/2015  :-DD I quickly realise my lab (not the labrador, the laboratory) is useless for anything in the microvolt level due to unavoidable environmental factors!  :-DD

However I still find all the volt-nuttery very interesting indeed. Please carry on! :-+
 
The following users thanked this post: Vtile

Offline lowimpedance

  • Super Contributor
  • ***
  • Posts: 1060
  • Country: au
  • Watts in an ohm?
Nice to see there is still interest in these obsolete, but not forgotten, metrological artifacts.
Attached is a 1958 reprint excerpt from the IRE Transactions On Instrumentation. by George D. Vincent.
Hope you find this an interesting read.

Had to split it into 2 PDFs to suit the forum file sizes !.

ps: I promise this is the last upload on weston cells from me  :)...............            maybe.
The odd multimeter or 2 or 3 or 4...or........can't remember !.
 
The following users thanked this post: Dr. Frank

Offline VK5RC

  • Supporter
  • ****
  • Posts: 2246
  • Country: au
Thanks for a concise and understandable summary of your professional experience and development of voltage standards.
It must have been, and still is I hope, an interesting journey!  :-+
Rob.
Whoah! Watch where that landed we might need it later.
 

Online Dr. Frank

  • Super Contributor
  • ***
  • Posts: 1697
  • Country: de
Many thanks to lowimpedance for that interesting article about the building techniques.

I have another one from NBS / NIST, which is complementary.
It's too big to upload, so here's the archive address, for a tidy copy: https://archive.org/details/standardcellsthe84hame_0

It also depicts the changeful volt-history based on Weston cells.

Frank

 

Offline HighVoltage

  • Super Contributor
  • ***
  • Posts: 4252
  • Country: de
I have found another nice document about the drift of the Weston Cell (see attachment)
"THE EVALUATION OF WESTON CELL RELIABILITY ON THE BASIS OF THE ELECTROMOTIVE FORCE DRIFT"
This one is from 2013

Also interesting ...
I set my 34461A to HighZ input impedance and in this setting the instrument has about 131 GOhm of input resistance (measured).
The voltage drop of the Weston Cell can clearly be observed within a 30 min time frame and drops about 13 uV



 
There are 3 kinds of people in this world, those who can count and those who can not.
 

Online Macbeth

  • Super Contributor
  • ***
  • Posts: 2521
  • Country: gb
@HighVoltage - Blimey that seems a heck of a lot for such a huge input impedance. Does the cell recover at all afterwards?
 

Online Dr. Frank

  • Super Contributor
  • ***
  • Posts: 1697
  • Country: de

Also interesting ...
I set my 34461A to HighZ input impedance and in this setting the instrument has about 131 GOhm of input resistance (measured).
The voltage drop of the Weston Cell can clearly be observed within a 30 min time frame and drops about 13 uV

Hello HighVoltage,

I assume, that the 34461A does not induce any current spikes from the AZ chopping, otherwise you probably wouldn't have measured 131GOhm.
What about the room temperature over these 30min? Did you monitor the cells temperature?
A standard cell has about -40µV/°C, so a temperature rise of 1/3°C simply explains your observation.

Frank
« Last Edit: October 18, 2016, 09:09:52 pm by Dr. Frank »
 

Offline HighVoltage

  • Super Contributor
  • ***
  • Posts: 4252
  • Country: de
@HighVoltage - Blimey that seems a heck of a lot for such a huge input impedance. Does the cell recover at all afterwards?
Yes, the cell actually recovers very nicely after about 1 hour time, the 13 uV are gained back.
Since this is my first Weston Cell, I have no idea if this is a normal behavior even for such small currents.

I have also hooked up my Keithley 614 Electrometer ( >50 T Ohm input impedance) and spot checked the cell  every few hours and did not notice any drop, based on the 614 being constantly hooked up in parallel to the cell.
There are 3 kinds of people in this world, those who can count and those who can not.
 

Offline enut11

  • Frequent Contributor
  • **
  • Posts: 284
  • Country: au
  • Love building/modifying/restoring test equipment
Dr Frank, thanks for the wonderful journey back in time wrt to voltage standards. It reminded me of what happened some 40 years ago. I used to spend Sat mornings wandering through our local scrap merchants electronics section. One day I found a Western Cell wired inside a big box. I went back to my car to get my cutters but on returning, it was gone. I was heart broken. I would have destroyed it anyway as at the time my best multimeter was 20kohm/volt!
Enut11
an electronics nut from wayback...
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf