Need some help with thermocouple's and calculating temperature from different calibrations

[MrBond]

Ok, so I'm doing this experiment at work that takes me about 2 hours to set up and 2 hours to run using thermocouples and a pc logging device. In our software, there's an option for the type of thermocouple you used.

Well, I must have been still sleepy this morning when I set it up, because I used type J in the software, and a type T thermocouple in the actual test.

Since the all the logging device does is measure the voltage, which the software uses to calculate temperature, there has to be some sort of formula I can use to go from Type J temperature reading to voltage, to type T temperature reading. I've got about 20,000 data points, so I'll be doing it in excel, of course.

I found some calculators online for it, but no actual formulas. Does anyone know the formulas for it offhand? This would be a HUGE help and save me about 4 hours tomorrow at work , so thanks in advance

 

[NewSc2]

another reason why i don't like chem labs

 

[MrBond]

Originally posted by: NewSc2


another reason why i don't like chem labs

It's not even a chem lab, I'm out on Co-op. The experiment itself ran just fine, and my data is slightly useable for what I want to do with it, but I'd MUCH rather have the correct numbers.

I know there's got to be some simple forumula from taking a temperature and figuring out the voltage that cooresponds to that, then converting it using another formula to the right value, but damned if I can find t hem

 

[KMurphy]

http://zone.ni.com/devzone/conceptd.nsf/webmain/B886ECA5994AA782862568650068FCEA/$File/AN043.pdf

That is the best link I could find. The best way is to use the temperature scale of each type and use their outputs (4-20mA or 0-5VDC) to get an approximate ratio. These devices are non-linear however and you should probably redo the test for accurate results.

 

[MrBond]

That might just work, I'll have to wait until tomorrow am to see for sure.

 

[RossGr]

If, with the software that you have found on the web you can input at voltage and get a temp simply enter a voltage, get the corresponding temp for a J type thermocoulpe, and at the same voltage get the temp of a T type. Now compute the ration T/J that is the temp for the T type divided by the temp for J type, now multiply each of your data points by this number. I would be best to use a temperature value in the middle of your measurement range.

 

[Heisenberg]

Thermocouples are nonlinear so any type of simple ratio or shifting of the data isn't really going to work. Unless you want to try power fitting (which would give questionable results), it would be best to simply repeat the experiment. Four hours for an experiment isn't really that long anyway.

 

[MrBond]

I may end up just repeating the experiment. The actual data points themselves aren't important, but the overall curve they make is what we're interested in. The curve should be the same regardless of the thermocouple type. so it may not make a difference. This is actually the second time I've ran this experiment, so I'll compare the two curves and see how similar they and and go from there

 

[etech]

The curves for different types of thermocouples are not the same.

omega