Temperature Measurement Uncertainties - Verify a PRT at 0?C specified ?0.02?C

Temperature Measurement Uncertainties - Verify a PRT at 0°C specified ±0.02°C

Hello people,
i'm about to verify a PRT at 0°C specified ±0.02°C,but i'm not going to use an ice bath or TPW cell, instead i'll be using a stable liquid bath and another calibrated PRT (±0.006°C @ 0.0°C) as the reference standard. i'll also be using only one HP 3458A to measure the resistance of both PRT's (maybe one on the front the other on the rear terminal) then convert it to temp using ITS90 equations. Of course I need to consider all uncertainty contributors but my only concern is with regards to the error contribution of HP 3458A.

Do i need to include only 1 uncertainty calculation for the HP3458A contribution? or 2x since I use it twice ,one for the UUT and one for the standard? or I should not include it at all since its error contribution is cancelled out. Since i'm making this comparison:
TempUUT = TempStd
TempUUT + DMMerr = TempStd + DMMerr
ideally the error TempUUT-TempStd = 0
is this right? (TempUUT +DMMerr) - (TempStd + DMMerr) = 0
and DMMerr is cancelled out???

The 3458A uncertainties - from its accredited calibration as reported on the certificate - is a Type B and should be included in both calculations. The uncertainty number will be the same in both calculations. Note that if the range/function on the 3458A changes then the uncertainties may also change and that must be accounted for.

The PRT is a Type A of course, and may vary Std to UUT. Realize if you were to put this in a ratio format, you are looking at 1:1 because the Std and UUT are (effectively) identical.

Keep in mind the lead resistance and any thermal effects, ambient conditions, the fluctuation of the liquid and other factors. The liquid will fluctuate temp regardless of how well it is controlled.

Hope this helps.

Hershal

You should be able to use the 24 hour specs for four-wire resistance measurement. There is a strong correlation between the two measurements, i.e. a (small) offset in the meter becomes a second-order effect. With great care, you could get down to maybe 2-3 ppm for the resistance ratio. Then comes of course all the temperature related effects and the stability...

Hans.

Not using ice bath

celon said:

Hello people,
i'm about to verify a PRT at 0°C specified ±0.02°C,but i'm not going to use an ice bath or TPW cell, instead i'll be using a stable liquid bath and another calibrated PRT (±0.006°C @ 0.0°C) as the reference standard.

Click to expand...

Why don't you want to use an ice bath if you are about to verify a PRT at zero degrees Centigrade? Can you explain to us?

There is a fundamental physical property associated with an ice bath and you may be overlooking it. The so-called "stable" liquid bath will not be "stable" and will show temperature fluctuations, as noted in this thread. The temperature scales that we use were developed many centuries ago by simply DEFINING the temperature of an ice bath to be zero degrees Centigrade.

It would seem to me if you want to know if a PRTs accuracy to .02 deg C at 0 deg C you should be testing it at 0 deg C. A temp bath will not give you the kind of stability you will be needing like a Ice point will. Makes more sense to test it with a ice point or TPW for a good measurement.