Add your comments about the Circuit Note found at http://www.analog.com/en/circuits-from-the-lab/CN0214/vc.html.
Questions or comments about CN0214, leave us your feedback here.
Possible flaw in the source code:
Since the thermocouple only "sees" the difference between the "hot" junction and the cold juction it will generate a voltage
V_Meas = V_TC(T_Hot - T_Cold) = V_TC(T_Hot) - V_TC(T_Cold)
To get the temperature at the hot junction the temperature at the cold juction has to be converted into the corresponding thermocouple voltage first and then added to the measured voltage.
V_Meas + V_TC(T_Cold) = V_TC(T_Hot) - V_TC(T_Cold) + V_TC(T_Cold) = V_TC(T_Hot)
Because the cold junction temperature measurement is calculated using the measured RTD voltage, I don't see why this needs to be converted to a thermocouple voltage?
The cold junction temperature measurement is intended to calculate the temperature at where teh thermocouple meets the PCB.
By converting this temperature reading to a thermocouple voltage, isn't this unnecessary and potentially introducing an error?
Actually, I see your point now.
We will revise our source code in the coming weeks and update the document also.
We will add the following changes:
1) Add linearization of the RTD measurement to get improved RTD accuracy.
2) Convert the RTD (Cold Junction) temperature to its equivalent thermocouple voltage - this will involve adding another look-up table to go from voltage to temeprature.
3) Use the voltage in 2) to compensate the overall thermocouple voltage and thus the overall temperature.
Does this mean the original calculation is incorrect?
For a type K thermocouple and 100 °C <= T_Hot <= 1000 °C the errors are for
T_ColdJunction = 0 °C -> 0 °CT_Cold = 25 °C -> -0.02 °C to 1.56 °CT_Cold = 50 °C -> -0.24 °C to 2.51 °CT_Cold = 75 °C -> 0.17 °C to 3.20 °C
(not including the errors from the inverse function (max -0.05 °C to 0.06 °C) http://srdata.nist.gov/its90/download/type_k.tab)
Just a quick question about this circuit, which type of thermocouple can be used. I have an application that use the S type which are difficult to handle due to its low voltage characteristic. (16mV @1550°C) this is 4 time lower then the K type.
Kind regards, Sasa
It depends on what your requirements (accuracy, resolution , environment, measurements/sec, ambient temperature, drift etc) are.
The calculations are the same except for the different coefficients.
Retrieving data ...