>A colleage is considering the AVR for a product line.
> He's concerned about the availability of the parts firstly.

Look at www.findchips.com . It gives some guide as to what you can 
expect. Anything can dry up on you suddenly. The volumes and time span 
of the project influence this of course.

> Second issue is the A2D converter built in reference.  The spec 
> seems a
> bit odd. +- 0.1 V from nominal from a 1.1 V reference is around 
> +-10%
> seems rather strange for that to be called a voltage reference too 
> me.
> Can anyone shed light onto weather this is indeed how bad the 
> reference
> is? Is it really that pointless to use the internal reference or 
> not?

Stability of a band gap reference over all relevant variables is more 
important than absolute initial accuracy. In many cases a reference 
can be expected to be more stable than absolutely accurate. This is 
easily accomodated with digital processing of the result. A key 
feature of a band gap reference is it's relative temeperature 
stability (achieved by summing two voltages which counter track with 
temperature).

Odds are that you can calibrate the ADC to far better than 10% with a 
single reading during setup.

I just checked the ATMega48 datasheet. Reference is 1.1V +/- 0.1V and 
there is no spec on stability. Some testing needed to see what can 
actually be expected BUT it will be far better than 10%.



        RM

-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist