> I've been trying to get a 12f675 to read and linearize
> a thermocouple.


> If that fails, I'll use a bigger pic(16f873) and try using floating
> point arithmetic to fit the curve (And I believe floats are a big
> pain in assebler

Not really.  You do have to wake up to create the floating point subroutines
in the first place, but after that it's just a bunch of calls.  For floating
point I extend the general register model of my development environment to
include three 32 bit registers overlayed onto the existing 8 bit registers.
The floating point routines all act on the low 24 bits of REGA, REGB, and
REGC.  These include float, fix, add, subtract, multiply, and divide, plus a
few useful macros.  The preprocessor also has built in floating point
computations, plus the ability to convert the result into a 24 bit hex value
that MPASM can understand.

Is this a funded project?  If so, I can get you going with floating point on
PICs quickly.  We are also willing to extend the floating point library as
needed for real customers.  Sorry, but this is one part of the PIC
development library we don't give away for free.  Contact me privately if
you're interested.

> Is there something very wrong I'm doing here ?

I can't tell.  It would help if you said what you are actually doing instead
of referring to app notes.  I (and I suspect most people) aren't going to
look up an app note to answer a PIClist question.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.