HomeRoast Digest

Topic: Thermocouple Correction (5 msgs / 109 lines)
1) From: David Iverson
For all you thermocouple users, how are you dealing
with the non-linear output curve?  It looks like over
the range of typical roast temperatures there's about
a 30 deg. F error band for a type J thermocouple.  Is
it significant?  
So far I've just been subtracting a constant to get a
close match with the thermometer around 400 - 450, but
that makes the cooler temperatures read low.  I'm
looking at adding a software correction curve routine,
but wondered how/if others deal with this.
Do you Yahoo!?
Friends.  Fun.  Try the all-new Yahoo! Messenger.http://messenger.yahoo.com/

2) From: Bob Yellin
I'm not sure what you mean. For thermocouple wire only, with no =
connector, my
table says the following for type J and K wire:
2.0F (1.1C), or 0.4% of reading above 32F (0C); =
2.0% of reading below
32F (0C)
What am I missing? Look here (for example):http://www.coleparmer.com/techinfo/techinfo.asp?openlist=D,E,C,A&htmlfi=le=SelectingThermocouples.htm&Title=Thermocouple+Selection+Guide
Bob Yellin

3) From: Jeffrey A. Bertoia
This is a multi-part message in MIME format.
I would bet that Dave does not have cold junction compensation...
What are you using to read the TC before your A2D?
Bob Yellin wrote:

4) From: Matt Hoffman
Bob Yellin wrote:
I believe he's referring to the fact that thermocouple's mV reading 
isn't linear with temperature. It's not error, because it's expected and 
a function of the metals involved in the TC.  The equations to take care 
of that nonlinearity can get rather complex (if i remember right, 
properly correcting a type J TC requires a 9th order polynomial...I 
don't remember for sure, though).  But that's usually the responsibility 
of the device reading the TC.  Dave, what are you hooking the TC up to?  
The only way I know that you'd need to do corrections by hand would be 
if you're hooking it straight up to a voltmeter that isn't supposed to 
handle type J's.
AFAIK, a lot of multimeters only claim to handle K's because a type K is 
almost linear for the first 800 deg.F or so (wildly guessing...I haven't 
seen the graph in a while), so it's easier to correct.  Or maybe they 
don't correct for the curve at all, and figure you won't miss those few 
degrees of accuracy.  But the graph of a type J's mV output is visibly 
curved, as dave points out.

5) From: David Iverson
Thanks to all who gave interesting and informative
replies about thermocouple correction factors.  I
plugged a few numbers into the polynomial correction
equation and found with the way I've calibrated the
A/D converter and OpAmp, the readings are about 10
degrees F high in the neighborhood of 400 deg.  I
built a little lookup table function to provide
appropriate conversions.  Still have to give it a try.
Thanks again and happy roasting,
  Dave I.
Do you Yahoo!?
Friends.  Fun.  Try the all-new Yahoo! Messenger.http://messenger.yahoo.com/

HomeRoast Digest