Weston Cells

I last bought one (for such a portable potentiometer) about 15 years ago. They are probably still floating around on the used market. They last a very long time if not broken or abused.

I can imagine some ways to make a modern replacement, certainly good enough for such a calibrator application, that wouldn't need a power switch, but it wouldn't be totally trivial. It would also need a reference for calibration.

Best regards, Spehro Pefhany

includes enough information about the Weston standard cell to let me build one - if I had access to a chemical laboratory.

It includes a reference to a paper which may include more specific information.

You need mercury, cadmium, cadmium sulphate. mercurous sulphate (all poisonous) and water (only life-threatening in large quantities). IIRR the standard cell was built up out of 15mm diameter soft soda-glass tubing to produce a more or less H-shaped cell - the tops of the H were open, and each of the bottoms was closed around a short length of platinum wire, which became your terminals.

One side of the H was loaded with enough mercury to cover the Pt wire, the other with amalgamated cadmium (cadmium metal dissolved in mercury.

You then dumped enough mercurous sulphate on top of the pure mercury electrode to cover the surface while staying below the crossbar of the H, and a similar amount of cadmium sulphate on top of the cadmium amalgam electrode.

Once you had done that you poured in enough saturated cadmium sulphate solution to fill the cell above the crossbar. Once the solution has percolated through the cadmium sulphate crystals (above the amalgam electrode) and the mercurous sulphate crystals (above the pure mercury electrode) you should have a working cell.

IIRR the open tops of the H cell were closed in some way to prevent evaporation and contamination, but I can't remember how.

I think I saw one on sale a while back at:

L.F. Hanney

44 (0) 1225 424811

You can build one but I don't think you can buy them any more because the chemicals are just too hazardous. What you actually need is a reference voltage of 1.0183 V, right? Did you know that NIST and CERN now use Zener circuits (as a secondary standard calibrated to a Josephson junction) because the latter is more stable than a Weston cell? You can DIY but zener diodes have tolerance issues.

If I wanted a chemical primary voltage standard I suppose I'd start stacking coins and paper disks soaked in saltwater.

OTOH do you absolutely need that last .0183 of a volt? If not:

Mark L. Fergerson

If *I* were wanting a voltage reference on the cheap I think I'd use one of the many bandgap chips that are available, add a heater so that ambient didn't enter into the picture, then calibrate it against a lab standard.

...Jim Thompson

...

If I were wanting to do it, I'd save myself a lot of time and buy a good one from a company just down the road from you, Jim:

Cheers, Tom

The standard cell has a fairly high tempco compared to most decent modern references (something like 50ppm/K), so I'm not sure an oven is called for. Natsemi has some with a built-in heater and temp control, but they are fairly pricey ($60 or something like that). Even the Thaler ones are gross overkill.

But here's a source (hey, $50 off regular price!) and an incentive to design your own if you go with one that requires a power source:

Maybe an 1.024V LM4140, with an op-amp buffer and trim network, but it probably will draw too much Iq to be left on all the time on batteries.

Best regards, Spehro Pefhany

Good link! They're probably using SiCr resistors and laser trimming. I didn't know they could be had that cheap! Probably a Burr-Brown spin-off... lots of spaciality companies in Tucson are.

...Jim Thompson

When I last looked - quite some time ago - the best buried zener references were a lot quieter and more stable than bandgap-based references.

Neither is going to be much use in a vintage thermocouple potentiometer, which won't have any kind of power supply, unless you sneak in a few lithium cells to power the voltage reference.

Nah, just add a coil and a cow magnet, and shake it whenever you need to use it. ;-P

Cheers! Rich

Does not work, get error message:

What about the use of those 5V regulators that Intersil makes (within

Don't you get buttermilk voltage then?

snipped-for-privacy@ieee.org wrote: ...

Really?? I would expect it to have some other power (typically a battery), because the thermocouple voltages will be well below the standard cell voltage, and you don't want to draw current from the standard cell. I would expect another battery driving the potentiometer through a rheostat. There would be a switch to set the galvanometer to compare the top of the potentiometer with the standard cell, and you'd adjust the rheostat to null that. Then switch the galvanometer over to compare the thermocouple voltage with the potentiometer output, adjust the potentiometer to null the galvo, and read the voltage. Double-check the ref to be sure. Read the included thermometer to correct the ref voltage... Ah, the "good old days." The aux power might even be arranged in a way that you really could use it to power the reference, too, though its voltage may be too low unless you have a reference that can run on a very low (e.g. 1.5V) input.

Cheers, Tom

High accuracy potentiometers did use separate power and voltage reference to avoid current draw from the voltage reference, but the typical thermocouple potentiometer used a mercury battery for both purposes (not an actual Weston cell, but similar). The battery was only connected when a momentary switch was held closed. Working Weston cells go for around $10 on eBay now and then, but they were not designed to handle the current draw of a potentiometer and probably wouldn't fit anyhow. Subsituting a modern battery and a voltage reference would be the best way to make this antique work; I would try it with an adjustable power supply first to see what voltage range the potentiometers calibration adjustment can accommodate, it might adjust to work with a standard voltage reference value like 1.0 V.

... Wow. Sounds like an inferior piece of junk. The only ones I ran across (in physical chemistry and physics labs) used standard cells only for calibration and avoided drawing any significant energy from them. The schematics I've found on the web are all for that type, and in fact show a resistor in series with the galvo, so you can keep the current low till you get close to balance, then a switch to short the resistor to get max sensitivity.

Why bother with a potentiometer with accuracy no better than you can get from a calibrated direct-reading D'Arsonval meter movement? Once you've gone to the effort to make a nice linear potentiometer, why not use it in a circuit that takes full advantage of it? Maybe marketing was as active "back then" as it is now. (Oh, yes. The theme of "The Music Man.")

Cheers, Tom

Yes, that's the kind I've run into. I did retrofit one with a regulator and some D alkaline cells to get rid of the unobtanium mercury D cells which powered the potentiometer. Current was only drawn (or put into) the standard cell during the calibration procedure- push a button down and null the galvo.

Best regards, Spehro Pefhany

Hi,

What about an old DVM?

I have dismantled a few over the years and come accross these strange merucry filled devices - look for an old bench top nixie tube base unit with

Philip

They did actually have much better accuracy than any calibrated direct-reading D'Arsonval meter movement when used with thermocouples. While they draw current from the mercury battery, they do not draw any current from the thermocouple when nulled, that was the whole point. The mercury batteries were remarkably stable over a very long life, and their small tempco was compensated with a wire-wound thermistor in the voltage divider chain.

I worked in a lab which had a bunch of these potentiometers some decades ago, along with a hundred or so thermocouples all made from the same batch of wire, and the potentiometer/thermocouple combinations pretty much all came in within a few deg F prior to adjustment on their annual ice/tin calibration at 449.54 F. These potentiometers could not be callibrated with a standard cell since they only went up to 30 mV, which was typical with thermocouple potentiometers, and could only be read to the nearest 10 uV anyhow, entirely adequate for the application.

My antique Guildline potentiometer is another beast entirely, with a