The cryogenically cooled (77K) IR detectors that I worked around used a PN junction -- usually the BC junction of an NPN transistor (I think it was the BC junction -- maybe the BE junction? The one that works more like an ideal diode than a diode, at any rate).
They required calibration, but a "simple" dunking in LN2 provided the temperature reference for that.
No, very linear means that the compensation is very close to 1:1.
13.591 +/-3uV for any reference junctionOr to put it another way, a temperature change of the reference
reading.
--sp
OK all I need is a room temperature superconductor.
I thought we were supposed to have those by now :^)
George H.
What does that have to do with linearity?
Free beer "tomorrow"!
Probably collector-base shorted. n ~= 1
There are calibrated diodes sold for that purpose- not cheap, but much cheaper than the Ge RTDs.
nice healthy 5mV+ of voltage.
--sp
If you plot uV/K vs. temperature the line is not flat- it approaches zero near room temperature.
If you plot uV vs. temperature, the plot has a hook around room temperature (almost flat). It is not straight i.e. it does not appear like a (straight) 'line' and hence can be considered 'non-linear'.
The particular **type** of nonlinearity exhibited here (a hook to near zero slope around room temperature) means that cold-junction compensation is not important.
Sorry, I don't know how to say it any more clearly.
--sp
That's because the non-linearity is not significant. The reason why you see little error at room temperature is because the error is near zero at room temperature. If you do the same measurement with the cold junction significantly away from room temperature the error increases significantly, no?
I was asking about this statement... "If you really don't like cold-junction compensation you can always use type B- it's so nonlinear around room temperature that it doesn't really need compensation"
The reason why it doesn't need compensation near room temp is not
*because* it is non-linear. The low error part of the curve just happens to fall at room temperature. No?On Fri, 11 Sep 2015 18:54:27 -0400, the renowned rickman wrote: lly need compensation"
Well if it wasn't nonlinear and had low output around room temperature (and thus everywhere else) then it would not be a very useful thermocouple, would it?
--sp
Are you really that poor at logic?
Hmm, a lot of heat would need to travel down the wire. The temp probe is 3 ft long, I have about 1ft in the oven, leaving 2 ft going to the meter. The toaster oven is mounted under cupboard, and the meter is below on the counter.
I don't understand Vm = Vtc-(Vcj1 + Vcj2) = Vtc-Vjc Is Vcj = to mv/t from the table? ie. both junction are included in the in the reference temperature.
The transition to copper, ahh, there is a plated connector on the probe, this plugs into an adapter that plugs into the meter. The adapter came with the original meter, for this purpose.
I'm seeing a problem with actually knowing the temperature at the probe connector/meter connection. I assumed it would just be room temperature. Now I'm concerned heat migration down the wire will upset that. It was very sensitive to have my hand near the meter connections. Mikek
That's why a decent thermometer measures the cold junction temperature.
Can you solve this ?
--sp
Oh boy, how long is it going to take me?
George H.
Have something to add? Share your thoughts — no account required.
Ask the community — no account required