If you are intent on using thermocouples, then I would highly recomend signal conditioning, as at room temperatures, the millivolt output of almost all thermocouples is very low (depends on the type of thermocuple). Thermocouple-specific signal conditioning will also provide for the required signal characterization (thermocouples are not linear devices). The other problem you mention, calibration, is also critical with thermocouples especially if you are operating in the room temperature and lower ranges. You have to have a method of compensating for the reference junction temperature (assuming that it remains constant, which in an outdoor environment, is questionable). Having said all of the above, I would highly recomend that you elect to use RTD's to measure your temperatures. The first gain is that you can calibrate them with a fixed value resistor substitution. There is no reference junction ambiguity/possible-variable to deal with. You will require signal conditioning with and RTD, but the cost is similar to that for a thermocouple, and you circumvent all of the problems of thermocouples. The second big advantage with RTD's is that you can wire to them with ordinary copper wire. With thermocouples, you must use specific thermocouple wire (read expensive) from the thermocouple terminals to the point at which you measure your reference junction. If there is some distance to your thermocouple, the cost of wire can outstrip the cost of the thermocouple itself! Which ever path you decide to take, NI has good signal conditioners listed on their site/in their catalog. Hope this is of some helpt to you. Dave
