I sent a PICList post on this topic to a friend of mine who I knew to be
expert in the use of thermistors for temperature measurement and got back
the following reply. I've removed a few of the more libellous parts :-).

This comment should be of value to those interested in temperature
measurement.




RM

______________________________________

From: Ken and Louise Mardle <louisem@adr.co.nz>
Subject: thermistor temperature measurment


>I would suggest that thermistors are now the technology of choice
>for many temperature measurement applications, and the ONLY viable
>technology in some cases.
>
>A huge variety of thermistor devices are now available ranging from
>extremely inaccurate devices costing cents and intended for thermal
>protection, low-cost devices designed for temperature compensation in
>appliances like TV's and computer monitors, to very high precision devices
>costing 100's of dollars intended for very accurate temperature
measurement.
>
>We routinely use precision NTC types from YSI, Themometrics, Alpha
>Thermistor, and several other manufacturers with accuracies ranging from
>+/-0.5 degrees Celsius to +/- 0.01 degrees Celsius.  Prices for these
>devices range from around US$0.25 to about US$90.00 depending on quantity,
>precision, temperature range, etc.
>
>Used within the manufacturers recommended temperature range, a good quality
>thermistor designed for temperature measurement is very stable over time,
>and has a repeatability much better than its accuracy grade.  This often
>allows the accuracy to be improved (over the manufacturer's out-of-the-box
>spec) by specific calibration.
>
>This class of thermistor (generally termed a curve-matched type) is also
>distinguished by a high degree of similarity between devices of the same
>type, giving both a good initial calibration and allowing
interchangeability
>between devices without recalibration.
>
>Low cost (epoxy encapsulated) thermistors are generally rated for
>temperatures in the range -25 to +110 degrees C although some will go as
low
>as -80 degrees C or as high as +150 degrees C.  Glass encapsulated types
>cost more but extend the temperature range in both directions.  Recent
>developments in thermistor technology allow temperatures in excess of 500
>degrees C, and types for cryogenic and extra-terrestrial applications are
>readily available.
>
>Basic thermistor devices are non-linear, and while this requires
>linearisation for purposes of temperature measurement, this is normally
>relatively trivial if a computing element (microcontroller, microprocessor,
>PC, PLC, or whatever) is present in the system.  In other cases, multiple
>thermistors can be combined with standard fixed resistors into a network
>which has a highly linear voltage output vs temperature characteristic over
>a reasonably wide range of temperatures.  For instance, Betatherm offers
>several "BetaTrack" devices covering different 100 degree C ranges and
>having linearities of better than +/-0.25 degrees C and accuracies of
better
>than +/-0.15 degrees C (when the known residual non-linearity
characteristic
>is accounted for).
>
>The high level of the output signal from a thermistor sensor means that it
>can often be used in electrically noisy situations where a thermocouple or
>platinum RTD device would be difficult to use or require expensive signal
>conditioning (AC drive with synchronous demodulation).
>
>Thermistors are also produced in a huge range of physical forms which offer
>distinct advantages over other temperature measurement technologies.  These
>include thin films, flexible elements, fine tubes, microdot arrays (arrays
>of thermistors on a Kapton tape substrate), tabs, washers, bolts, etc.
>
>With a proper understanding of the thermodynamics involved, and for fixed
>applications (common in industrial control systems for kilns, forging,
etc.)
>arrangements of several thermistors in a servo type circuit can even be
used
>for medium precision non-contact temperature measurement as a much cheaper
>alternative to the thermovoltaic pile sensors.
>
>As you can probably tell I am something of a fan of thermistors, ...  Other
>temperature measurement technologies also have their place of course but
you
>should never overlook the thermistor.

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email listserv@mitvma.mit.edu with SET PICList DIGEST in the body