Felis Corax
Harmless
Posts: 35
Registered: 26-7-2019
Location: These Uninted States of Murica
Member Is Offline
Mood: Pre-project euphoria
|
|
Making a triple point of water cell
While getting ready to do some thermometer calibration (the horror!) I ran into an old article on creating a triple point of water cell (Tackling the Triple Point). I thought I'd see if anyone here has made such a cell before or otherwise has usable experience/insight before I
begin.
In the absence of a "straight-walled flask" I'm using a 41x225mm borosilicate test tube. It's reasonably heavy walled (guessing 1-2mm) and has a good
chance of handling a little external pressure. If not... well, that's what face shields are for! 
I'm planning to fabricate the thermometer well out of some 10mm ID borosilicate tubing as I haven't been able to source test tubes that are suitably
long and narrow. My glassblowing skills are minimal, and all I have is a map-pro torch, but that should be sufficient to close off one end of a tube,
flair the other end, and flame-anneal the whole thing.
As I've never done flame annealing before I'm going to make a polariscope to verify my results. A couple of sheets of polarizing film and some 3d
printed parts should suffice. I may even be super lazy and set the whole thing up in a ring-stand.
Any thoughts or suggestions would be welcome. Particularly if you know any way I can further over-complicate the ostensibly simple process of
calibrating a thermometer in an ice bath. 
Nothing is pure, nothing is perfect, nothing is clean.
|
|
|
Sulaiman
International Hazard
Posts: 3558
Registered: 8-2-2015
Location: 3rd rock from the sun
Member Is Offline
|
|
Unless the resolution / readability of your thermometer is less than 0.1oC then I think that distilled water-ice in distilled water will be
sufficiently accurate.
it's the 100oC end of the scale that is difficult due to
. atmospheric pressure
. superheating of the water (I have found 1oC to 3oC error due to superheating)
CAUTION : Hobby Chemist, not Professional or even Amateur
|
|
|
Felis Corax
Harmless
Posts: 35
Registered: 26-7-2019
Location: These Uninted States of Murica
Member Is Offline
Mood: Pre-project euphoria
|
|
How is THAT supposed to help me over-complicate anything? Sheesh! 
But in all seriousness... you're not wrong. I could probably get away with just an ice bath, but I thought making a triple point cell would be neat.
And not a total waste of time as I'm using it to calibrate a pt1000 RTD for use as a reference to calibrate a bunch of ds18b20 thermometers which
suffer from linearity problems but are quite repeatable.
My understanding is that the resolution for a platinum rtd is primarily a matter of calibration and ohmmeter repeatability. I'm using an old fluke
multimeter at the moment, which ought to be pretty good until I can build something better. With the correct constant current source and voltage
reference the ADC on an arduino should be accurate to within 0.1% of the measurement range. But I'm getting a little off in the weeds here... every
project compounds into another project onto infinity...
As for the high point... I'm not even going to try with water's boiling point. It's my understanding that a platinum RTD should have near-perfect (for
my purposes) linearity within it's design range and, as such, I should be able to use a gallium melt cell as a second reference point and extrapolate
from there. But that's a project for later...
Nothing is pure, nothing is perfect, nothing is clean.
|
|
|
andy1988
Hazard to Others
Posts: 135
Registered: 11-2-2018
Location: NW Americus ([i]in re[/i] Amerigo Vespucci)
Member Is Offline
Mood: No Mood
|
|
Apparently the pencil eraser is a thing. Years ago I'd thought some PhD physicists I worked with were just jimmy rigging on-the-fly whatever it was
they were doing.
Table 2 adjust for atmospheric pressure and hydrostatic head of the liquid in the cell: https://www.bipm.org/utils/en/pdf/ITS-90.pdf
The physicists I worked with also used gold and a different metal as calibration points for their platinum rtd and... different probe (gold was too
hot for the non-platinum one IIRC?). It's been so long I don't remember. Be very gentle with the platinum rtd, bending it will throw off the
calibration and you'll want to re-calibrate IIRC.
Maybe look into calibrating the multimeter as well as you said, maybe renting one within calibration if it's super important/official. Calibration
certification only lasts a few years(ish?) per whatever standard, and there exist companies which rent out calibrated instruments. The ones I used
needed to be on for so long (an hour?) to "settle" prior to measurements per the manual's instructions.
This website I have bookmarked as a reference, helped me make sense of things: http://www.mosaic-industries.com/embedded-systems/microcontr...
It is neat to see that Scientific American article thank you... it helps me better understand the bigger scope of what I did years ago.
[Edited on 7-5-2020 by andy1988]
|
|
|
![[*]](images/xpblue/default_icon.gif){kind=icon}
