🚀 go-pugleaf

   Hello folks.  I am wondering if someone can offer advice on 
thermocouples.
   I need to control the temperature of a small infrared window to 
0.1 degrees C or better.  The actual value need not be so precise, but
changes in temperature must be kept to a minumum.  At present, the
window is heated by a ring of brass wrapped with heater wire. Adjusting
the current changes the temperature, which is typically 100-200 C.
The temperature is measured with a thermocouple (5 mil Chromel-Alumel wire,
"type K").  Heater current is about 200 mA.
   I would like to build a simple analog circuit to better regulate the 
temperature and be able to easily adjust the set-point.  I envision an
op-amp to amplify the 4mV of the TC to about 1.00 V at 100 C.  Another
op-amp can compare the output to a 0-2V reference voltage and provide
an error output to run a current stage that drives the heater.  I can
read the output of the first op amp on a DVM to obtain a temperature.
The size (and heat capacity) of the window is fairly small and the 
sample (on the window) is fairly heat sensitive, so the stability
must be good (can't quote a spec here, I really don't know).
   Less important, but also desireable, might be a temperature compensation
scheme, since the "room temperature" in this building can vary from 10 C
to 40 C ("Room temperature physics" takes on a whole new meaning here
at Stony Brook).  Also, I could conceivably increase the sample temp. 
up to 600 C, so a way of linearizing the TC might prove useful.
   If anybody has existing circuits for this purpose they are willing
to share, I'd love to see it.  Simple is best, since the simpler it is,
the faster I can build it.  If not, any advice on the best op-amps to
use would also be helpful.
   Finally, though TC's are prefered, I can also use a silicon diode
to measure temp (though not to 600 C).  A similar circuit for a diode
would also be welcome as I have some nicely calibrated 1N914's.  
   Thanks for the help.
							Danube
							(Daniel Koller)