Feel like resolving temp to within 0.005C for less than $100?

[UberBlue]

Complete how to: Right here. (http://www.sciam.com/article.cfm?chanID=sa006&colID=22&articleID=0006869F-5263-1C6D-84A9809EC588EF21)

Looks like a fairly cheap and easy to best really expensive equipment. Calibrated with a know physical standard; the triple-point of water 0.01C.

I'm guessing accuracy would depend on your multimeter.

Two questions:

Is the linearity of platinum RTD's assured by the physical properties of platinum?

If linearity can't be assured, how do you calibrate it?

[murray13]

Only if you have a $1000+ DMM laying around. :cry:

[BillA]

you take the RTD to your cal lab and tell them the temp range over which you wish to use it, and the increment at which you wish the indicated vs. actual temps recorded
$80 to 120 or so
and it helps if the indicator is calibrated at the same time and/or with the same probe, as then there is another correlation resolved

a single point 'calibration' is pointless

[TerraMex]

Will a Keithley DMM 2000 do? :D.

[BillA]

Will a Keithley DMM 2000 do?
work through the specs (fair, but nothing special)
-> you will NEVER get 0.01°C accuracy, so what is the point of 0.005 ohm resolution ?
and
that is just an indicator
-> you still need to calibrate both against a known reference standard

[TerraMex]

I find them pretty good for a small lab.
We have a few. And a few flukes.
I usually use them with the GP-IB in Labview.
We have a few National Instruments and Keithley aquisition cards so the work is mostly automated.
Well , the keithley has a 0,0001 Ohm res with the 100 Ohm range , but the accuracy isnt that good. So , in real life conditions, the errors stack up. Still, , better than those 50 bucks DMM's with temperature reading :) .

-> you will NEVER get 0.01°C accuracy, so what is the point of 0.005 ohm resolution ?

Dick size? ;)

[UberBlue]

you take the RTD to your cal lab and tell them the temp range over which you wish to use it, and the increment at which you wish the indicated vs. actual temps recorded
$80 to 120 or so
and it helps if the indicator is calibrated at the same time and/or with the same probe, as then there is another correlation resolved

a single point 'calibration' is pointless

Bill,

Your word is one of the few I will take without question. With that being said, is this worth persuing any further? But, before you are quick to pass judgement let me share my thoughts.



PSRT's have a good approximation of linearity (link) (http://www.picotech.com/applications/pt100.html). Keep in mind they are talking about 100 Ohm PSRT's. A 1000 Ohm'er would be a magnitude of order better.

I have the nescessary multimeter. It was calibrated two months ago IAW ISO 9001 controls. It's accurate to .0001 Ohm (picked it up brand new in the box at a probate auction for $50, with the cal sheet).


With a single point calibration, it would be close enough for an apples to apples comparison in my own little world. I'm OK with .04C accuracy @50C (based on single point callibration with TP of water). What I really care about is the potential for outstanding resolution.

What I am planning to do:


Build two of these. I don't need the temperature range the one in the article is built for, therefore I can cut out the spot-welded nickel wires and the Pyrex tubing. That is the bulk of the cost.
Calibrate one (#1) against the triple-point of water as a starting point.
Put them both in a water bath and slowly heat them, while continously monitoring #1 PSRT
At a TBD interaval, measure both PSRT's. By doing this I can build a correction table for #2 PSRT and have two PSRT's callibrated against each other offering mediocre accuracy (to the rest of the world), but very accurate between themselfs and with kick ass resolution.


Can it get any better for $100?

Fire away.

[murray13]

Will a Keithley DMM 2000 do? :D.

Yes, that meter will get you about .02 to .03 deg C accuracy PLUS the accuracy of the reference. Good luck getting stable readings, it's not easy when looking at .001 ohm at 1000 ohms. (The resolution of that meter)

Oh, that's a $1049 meter new.

Edit: Just remember .0385 ohms is .01 deg C, with a 1000 ohm PRT.

[BillA]

many things can work effectively, bearing in mind the limitations of the equip
an Murray13 observes, the stability may be vexing