AVR-Chat

At 08:46 AM 12/30/2003 -0800, Larry Barello wrote:

>I don't see anything wrong with the code, but the documentation is off.
>
>Just a couple notes: you can enable, set the clock and start the first
>conversion in one write.  There is no need to delay 15us, etc. 

Yeah, that was just "bulletproofing".

> Just let er
>rip and discard the first conversion  - unless that delay you have is to
>allow the Aref cap to charge... 

No, I saw a note about not enabling until one ADC clock after the ref/mux switch.
A foof-loop is about twice that.

> In any case the first, dummy conversion is
>something like 24 cycles long and at 2mhz clock (your doco says 125khz, but
>your code selects 2mhz) that should be sufficient.

AH! You got me :)  Three bits does not equal 3. :-P 
87!

>As to the problem?  Well, is ADCH declared correctly in the header?  I use
>the left adjusted, high speed clock ADC on an mega16 and it works like a
>champ.  Since you said the normal ADC routines work, I am assuming the Aref
>pin is good and not tied to +5v or something like that.

It was simpler than that, the voltage that I was trying to measure wasn't turned on yet... 0V.. @#%@^