Poor Man gets his candies (6X weston cells)

[MK]

or in other words about the same as a k-type thermocouple...

[Kleinstein]

There seems to be some common drift up on all cells (or the meter drifting down). With the limited way to measure the voltage there is a limited use in sending out a cell for calibration / check. It might need more time to see which cells are actually good. One might as well take the absolute values of those cells as a kind of primary standard (as it once was) - good enough for a 5 digit meter and it is not sure that an external calibration will help much as transport can add quite some errors.

[Kleinstein]

The temperature during transport is only one thing. There are also mechanical forces and possible tilting that can be a problem.

For a portable constant temperature box, I would not use a Pelletier element - they have quite a thermal leakage path and thus don't behave that well when used with low power. Just a heater and good insulation is likely the better way.  One may not need to set the temperature very high as the cell itself has essentially not self heating.

Form the shown data it is hard to tell a difference. It could be the 3rd temperature reading was off a little and this way caused slightly odd temperature readings. There are too little point and to much noise / uncertainty to really tell a difference.

[Conrad Hoffman]

FWIW, the Julie oven I had consisted of a heavy walled aluminum box with heaters or heater wire wrapped on it, and a mercury thermoregulator, inside another aluminum box set up the same way. The outer box was significantly above ambient, and the inner box just slightly above the outer box. A sensor and bridge were included so you could detect temperature fluctuations. Everything was insulated from everything else, and it was enclosed in a wooden case. The problem with saturated cells is that you need a really good thermal control system that can be trusted not to drift over a long time period.

[MisterDiodes]

zhtoor:
I think you need to listen to everyone and maybe think about another way besides wet cells.

I would suggest don't even worry about spending limited resources on getting a wet cell calibrated.  You don't really have any equipment that's going to see any benefit - and it's very very unlikely you're going to successfully get a wet cell to the lab and back without destroying the calibrated measure anyway.  In other words:  The cost to benefit ratio is very, very low for getting a calibrated absolute measure on a wet cell.

The other problem is with wet cells - the more you try to take a measure, the more you're changing the discharge curve.

For your situation getting something like a working 3456a (or similar), get that calibrated to within a few ppm (usually very inexpensive), and you're miles ahead for having a known stable way to measure your wet cells to a reasonably accurate absolute value.  That's probably the most bang for your buck to get some ppm measuring ability.  Those old meters are big and slow - but they can offer really good yearly drift specs (often better than 3458a).  Used, functional 731b's work very very well also and they don't break the bank - those can work well down to very low ppm per year drift also.  Which is still going to outperform and outlast your wet cells if you are using them at all.

We just measured a 731b recently that really wasn't drifting more than a ppm per year, and they don't get too picky if you keep them powered on or not.  A bit noisier than a 732a or battery, but they can be rock-solid stable as a value Vref.

You can still use the wet cells as a fairly quiet voltage source as required - as long as you're willing to discharge.  They will -always- discharge, not matter what you think you're doing to draw no energy...

But I wouldn't use the wet cells to transfer an absolute voltage measure to your location.  There's a ton of better ways to do it that are more realistic for your modest uncertainty of measure capability.

[MisterDiodes]

I think that's still probably a better chance of getting an absolute measure transferred to your lab.  Unless your calibrated  measurement facility was across the street where you could walk your wet cell (in a thermal box) over to get tested - there's just not a realistic good way to handle wet cells getting form point A to point B without disturbing the measure.

Remember:  You're not a calibration lab and you don't have extremely tight measurement uncertainty.  But that doesn't mean you can't run Vref experiments. 

The mark of a GOOD metrologist to is to recognize exactly where your limits of measurement uncertainty are, and keep your measuring results and analysis within your actual real measure capability.

In other words you learn when to stop putting digits to the right of the decimal point, and realize what data is meaningful and what data is just noise.

[MK]

I think that's still probably a better chance of getting an absolute measure transferred to your lab.  Unless your calibrated  measurement facility was across the street where you could walk your wet cell (in a thermal box) over to get tested - there's just not a realistic good way to handle wet cells getting form point A to point B without disturbing the measure.

how about driving 300 Km with the cell in a "thermal box"?

-zia

It would need to be heated, and how smooth are those 300k of roads, the more it is bounced about the longer it takes to settle down each time, it could be a month to get stable again.

[HighVoltage]

When I take one of my weston cells and move them a little bit, they immediately change the value.
I think they are only made to sit still and don't move at all.

Moving them 300 km in a car is a bad idea.

[Edwin G. Pettis]

The 3N164 is quite close to the 2N3609 and should work with out problems.  These MOSFETs come with a shorting ring around the pins, do not remove the ring until the FET is soldered into the circuit and remember, the gate is extremely sensitive to static discharge, the slightest zap will puncture it.