At 09:00 AM 6/19/2007, you wrote: > >>But, radiation and convection may be significant contributors > >>depending on a number of factors which may be poorly understood or > >>controlled here. > > > I think they can be ignored for the purpose of getting an estimate > > of the > > effect of the T/C under the presumed conditions (< 100=B0C). > >Why? >That may be good advice. But, can you provide rough justification for >it. Hi, Russell:- Partly experience, but also I just finished spending an academic year in a post-graduate program learning how to design injection molds ( I was a bit surprised to get the award for the highest marks last night at the convocation ceremony ) and part of that is calculating heat balances and heat flows on molds and hot-runner systems.. conducted heat (even when we resort to relatively exotic materials such as titanium and shapes designed to minimize heat flow) is the most important part of the calculations, and I see some analogies. You still need an estimate of the temperature gradient in the wire to use the thermal resistance. Copper is a rather common lab T/C positive wire (Copper-Constantan) and often used for thermal testing of motors and such like, but others such as K will have much lower thermal conductivity (probably more like stainless steel). >Holding the leads a small distance away from a hot component with thin >leads will expose you to a much reduced temperature. >My intuitive feel is that if you had a small device with thin wires at >approaching 100C then the lead conduction would be smallish but that >if you placed a sensor (finger, tongue or arcane electronics) near but >not touching it that you would sense substantial energy transfer. Right, and we don't much care about that so long as we don't affect it too much. We're trying to measure the temperature of the part without affecting it, so much as possible. Conduction out the leads, heat pads, etc. is really important, as you can see by looking at data sheets for ratings on various part packages. >Leads may be copper or (some may be surprised to find) steel. Plated steel is common, although probably *not* on a part designed for high power dissipation. >Copper has a coefficient of thermal conductivity of about 390 W/m/k. >Steel/iron leads will be about 5+ times lower conductivity than that. >Watts dissipation by lead heatsinking can be worked out from there. >Having copper or steel leads will greatly affect the result. >In the case of a thermocouple the leads will usually be neither copper >nor steel. Never steel, but copper is common. The leads of a type K T/C have thermal conductivity of 17 (+) and 29 (-) vs. 401 for copper (W/mK). Constantan is 23, so the positive lead will dominate in the heat loss. A 0.003" or 0.005" K T/C won't conduct much heat, and won't affect a larger copper wire (say 0.032") significantly. >The fact that PTC devices and other thermistors can have their >thermal/resistance characteristics very substantially affected by eg >adding a sleeve of heat shrink indicates that radiation and/or >convection are significant factors. Yes, they are quite different from thermocouples.. vastly inferior in this application because you can't easily couple the thermistor slug to the leadframe, so any kind of air flow will cause inaccuracies in the reading, plus the response is going to be slow due to the large thermal mass. You can analyze the behavior of thermocouples in thermodynamic terms, using the heat flow down the wires.. the temperature gradient, which creates a voltage gradient, which is kind of interesting. I've got at least one book that looks at it from that angle. Best regards, Spehro Pefhany --"it's the network..." "The Journey is the rewar= d" speff@interlog.com Info for manufacturers: http://www.trexon.com Embedded software/hardware/analog Info for designers: http://www.speff.com -- = http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist