[time-nuts] temperature sensor

[Attila Kinali]

> To satisfy my curiosity and get actual data I'd like to place 6 or more
> tiny analog high-resolution temperature sensors all around the OCXO of a
> Trimble Thunderbolt. That's high-resolution both in temperature and in time.
> In other words, no fake accuracy "averaging" allowed. The goal is to observe
> thermal gradients in real-time and see how good, or how bad, the correlation
> is among crystal temperature, case temperature, and DS1620 temperature
> sensor (which is mounted a considerable distance from the OCXO). The same
> technique, and maybe even the same conclusions, might apply to Rb.

Ok... a couple hours of reading later.... ;-)

My excursion into temperature measurement has brought some results:

1) PT sensors can be secondary standards for temperature calibration.
   But standard industrial sensors do not have the stability or linearity
   of the standard grade sensors. But at least they do not break when you
   glare at them. Those in ceramic housing are supperior to those in glass
   or metal housing. Thin film are inferior to wire wound. (in terms of
   stability and accuracy, thermal coupling is a different matter)
   (the price for a commercial standard grade PT sensor seems to be in
   the order of 3kusd)

2) The uncertainty of the calibration of the standards grade PTR seems
   to be in the order of 100uK to 10uK.

2) Making a triple point of water cell for calibration with an accuracy
   better than 10mK seems to be quite simple and doable at home, most
   likely something around 1mK is achievable. Also judging the quality
   of the cell is quite simple: make multiple of them, the one with the
   highest temperature is the most accurate one.

3) A well done ice bath gets you into the ballpark of 10mK accuracy.
   Most of the error is due to impurities and gas in the water.
   The air pressure effect is much smaller (and thus inconsequential)
   unless living on a high mountain. Also an ice bath is easier to
   do than using an triple point cell.

4) There are people on ebay who sell very pure Gallium and Indium that
   could be used for (not so accurate) melting/freezing cells for
   ~29.7°C and 156°C.
   (if anyone knows what the non-nut would use those for, please tell me)

5) The book "Traceable Temperatures" by Nicholas and White is a very good
   reading on temperature measurement and calibration. It explains the
   procedures with what can go wrong and what accuracies are achievable.
   It also contains a list of references for further reading. I did not
   have a look at those yet, but from the titles they look very reasonable.


				Attila Kinali

-- 
I pity people who can't find laughter or at least some bit of amusement in
the little doings of the day. I believe I could find something ridiculous
even in the saddest moment, if necessary. It has nothing to do with being
superficial. It's a matter of joy in life.
			-- Sophie Scholl