I have an AM25T on which one of the thermocouples seems to be wired in reverse. It is at a remote location and would be difficult and expensive to visit. Even if I could correct it today, I have weeks worth of existing data I would like to recover.
I am recording the AM25T's reference temperature along with all the connected thermocouples. If I understand correctly how the AM25T works, it should be possible to back calculate from the indicated temperature and the reference temperature, what the thermocouple voltage was. I could then apply that voltage to the reference thermistor voltage to get the correct temperature. If there is not a flaw in my logic, how to I get the reference voltage?
I think in short range, we can regard it as linear relationship between the output of delta temperature between the reference temperature to the tip of the thermocouple. So, if your Tref = 25, and reversed Trev= 23 and the real data Treal= 2Tref- Trev= 50-23=27.
The TCDiff and TCSE instructions of the datalogger measure the voltage between the two leads of a thermocouple and then use the standard NIST polynomials (documented in PB93-190338) to convert that voltage to a temperature. If you had the details of the polynomial (a different one for each kind of thermocouple type, K, T, S, B, etc.) then you could in theory "reverse" the effect of the polynomial (use algebraic techniques/MATLAB to calculate the inverse function of the polynomial and apply it to the numbers) to obtain the voltage back. At that point, you could reverse the sign of the voltage and then run it (forward) through the polynomial again.
If it is that important to you to jump through those hoops to obtain temperature, you could do it. However, I would only trust data obtained in that way for general/trend kinds of analysis and wouldn't give it the same weight as data obtained directly.
Here are some links about the NIST polynomials:
http://en.wikipedia.org/wiki/Thermocouple
http://ph.answers.yahoo.com/question/index?qid=20100423214251AABELS8
Accession Number PB93-190338
Title Temperature-Electromotive Force Reference Functions and
Tables for the Letter-Designated Thermocouple Types
Based on the ITS-90.
Publication Date Apr 1993
http://www.ntis.gov/search/product.aspx?ABBR=PB93190338
http://www.home.agilent.com/upload/cmc_upload/All/5306OSKR-MXD-5501-040106_2.htm?&cc=IN&lc=eng
Thank you for the comments. I will check out the sites and give it a try. I know that the resulting value may not be as accurate as it would have been with a properly wire sensor. However, recovering months worth of data trends is important and a loss of accuracy can be tolerated in this case.
This post is under review.