Hi All,
I found this thread when I doing a Google search for our (Playing With Fusion) website. I have a piece of correction code for the T-type thermocouple based on what a number of things I found regarding this thermocouple. We are going to release an appnote at some point here, but it's sitting behind a couple of other pieces of code that I need to get released! Please contact me (fusionstein@playingwithfusion.com) if you'd be interested in getting a pre-release version of the code to try out.
To go into a slightly deeper discussion, what MAXIM did with this sensor was to linearize around 'room temperature' operating conditions. Basically, the chip assumes that the temperature coefficient of the thermocouples are constant over the entire operating temperature range. Which is fine, if you're running a K-type between 0 and about 500C. Jump to a low-temp application or any other type of thermocouple (T-type is especially 'bad'), and the nonlinearity shows up in a pretty horrible fashion;T-type thermocouple readings are off by over 10C at a hot junction temperature of only 100C. Pg 8 of the MAX31855 datasheet says it all.... if you know what you're looking at (I didn't at first). I contacted MAXIM to discuss the design of their chip a couple of months ago, and only then fully realized what I was working with. A few voltage to temperature tables and learning how the Arduino handles large data sets, and voilà! The beginnings of a nice little appnote.
So, we come to the present. I have the beginnings of an appnote, a quad-channel TC board with 3 different thermocouple types represented, and not much time to test at the moment. Shoot me an email if you'd like to be a guinea pig, and I'll work through the pain with you. And please, don't hesitate to contact us directly if you have issues like this in the future - we have no intention of releasing products that we know nothing about!