thanks for the observations, i am sure u will be happy with the results and the costs of using thermistors .

one specific area we looked into was the use of transmitters, and how to maintain end to end accuracy in the field.

a big advantage of thermistors is one can use rather high impedance units, eg 10kohms, 30kohms, etc and thus reduce the errors caused by length of wiring. for example, at 25degC the 30k thermistor has slope rate of 800ohms/1degC, while a 100ohm prt would be below 0.4ohms/degC.

one high accuracy scanner/voltmeter like the hp3457 would be cheaper than an array of high accuracy transmitters, which themselves would require power supplies, recalibration, drift with changes in field ambient temp, etc. if needed the scanner could also read one calibration channel with low tempco resistor to keep one's peace of mind. the error due to wire could also be nulled in some sense, since we can short the wires at the thermistor end and measure the ohms. i would not think it is possible to reliably maintain 0.03degC accuracy using transmitters, and for our application we are shooting for zero calibration for more than 5 years.

for a reference above 50degC i am not aware of "primary" standards like the gallium cell--there was some mention in the literature of some
hydrocarbon compound but looks like no one is making them commercially. i would use an oil bath, with the prt to verify the accuracy of the control setting.

rgds
leelock

 

Typical temperature sensors used by climatologists have uncertainties down to a few millikelvins for temperatures less than 373K.

Climatologists currently measure to traceable NIST ITS-90 standards.

There must be some confusion between garden-variety industrial sensors (e.g., thermocouple and RTD) and those used by agencies monitoring the deteriorating global climate.

FB

 

I read an interesting article the other day.

Seems that surface temperature measurements that seem to show a very slight increase in surface temperature over the last century or so, do not correlate to temperature measurements taken of atmospheric temperatures, both from balloons and from satellites, which show no temperature increase.

If nothing else, this shows that we need to be sure of what it is we are measuring. And maybe more importantly, proper analysis of the data
collected.

Humorous anecdote follows:

A number of years ago, an acquaintance relayed the following tale to me. He was involved in a relatively simple project that had a steam fired
heat exchanger on it. He was tasked to add an additional temperature sensor to the heat exchanger outlet so that in the event of a primary
sensor failure, they could detect it and shut down the process.

The two identical sensors (RTDs as I recall) were located in the fluid stream, in identical thermowells, with the tips about 1" apart in the
fluid stream. One would be tempted to think that the temperature measured would be very close, wouldn't one?

Well, the temperaure measured at the two sampling points showed about a 5 degree differential. After recalibration, sensor switching, transmitter
switching, etc., they finally decided that there was indeed an inexplicable temperature difference between the two points being measured.

BTW - the spec called for controlling the temperature +/- 1 degree. I do not recall if he ever said what they did about this situation. I seriously doubt they shut the plant down.

Bob Peterson

US Filter, Rockford Number 815-877-3046 X576

 

With regards to problems with poor satellite correlation with surface temperature readings (specifically from 1979 to 1995), this problem arose because the satellite data was being misread.

Further analysis revealed that the data did correlate once corrections for orbital decay, different satellite types and other factors had been included with the data.

The result was a confirmation of surface data.

FB

 

We had a similar experience in a chemical plant we were building.

Flow through numerous ports of a reactor needed to be monitored to determine if there was a build-up that would lead to a blockage. If a blockage
occurred the result would be catastrophic.

Due to cost, it was decided to monitor the flow indirectly by monitoring temperature. A change of + 0.5 degrees would be sufficient to indicate
imminent danger and the process would be shut down.

However the inaccuracy of the RTDs, to say nothing of the inaccuracy of the balance of the monitoring system plus the variance within the process itself, resulted in every temperature reading being different.

By making the assumption that the flow in a new system was optimal, we forced all readings to be the same. We relied on the higher accuracy of
repeat reading, a feature of RTDs, to detect the 0.5 degree shift.

Of course, one never knows if every monitoring system would be affected exactly the same with a 0.5 degree change in temperature and I heard, a few years later, of at least one catastrophic incident in that plant.

Bob Pawley
250-493-6146