[sdiy] Tempco 2k +3350ppm Group buy

[René Schmitz]

Ian Fritz wrote:
> The problem comes if an inappropriate -- or misunderstood --  reference 
> temperature is used.  For example, if they use
> [R(50) - R(0)] / R(0)  or
> [R(100) - R(25)] / R(62.5)
> then the ideal tempcos would be 1/273 or 1/336.  In other words, the 
> slopes would be correct (as you point out), but tempco depends on the 
> slope divided by the reference resistance.

Agreed, thats nonsense indeed.
The slope you read off a R vs. T diagram would actually be something 
like (R(T2)-R(T1))/(T2-T1) or so. I.e. ~dR/dT. This you can then 
extrapolate to other temperatures. And factor out the resistance at any 
desired reference temperature.

> Yes, I agree.  It would be best if a clear notation was always used.  
> Including the nonlinear correction.  I have a feeling that some devices 
> are more linear than others, but this is confusing to figure out.

Higher order terms are a little too much to hope for, I guess. A clear 
anouncement of reference temperature would be nice for a start. :-P

Cheers,
  René

-- 
uzs159 at uni-bonn.de
http://www.uni-bonn.de/~uzs159