At 01:41 PM 9/10/99 -0400, you wrote:
>> The problem here as you note is the 8 bit AD, which isn't enough for my
>>application.
>
>If you have time during operation, you can increase the significant bits by
>multiple samples, with a known noise factor added in.  256 samples will
>yield 12 bits resolution easily.
>
>                 R1
>PORTx.y -----/\/\/\/\--o---Analog signal
>                       |
>PORTA.b ---------------o
>                       |
>                       -
>                       -C1
>                       |
>                      GND
>
>By toggling PORTx.y pin high/low every n samples (I used 16 in my project)
>and selecting the right values for R1 and C1 (C1 may not be necessary), it
>works.
>
>Somebody else can explain how/why - I'm busy!  Wagner would be a good
>candidate...
>

Seems like the toggle charges the cap and you are using the known decay
rate to more accurately measure the Analog signal.  Yikes... Sounds a bit
hairy to me, although a very interesting idea.  I suppose the oversampling
takes care of temp variations in the cap and resistor through averaging or
something.  Let me get a couple more cups of coffee in me and I may be able
to figure out how that works.

-E


>------------------------------------------------------------------
>andy@rc-hydros.com      http://www.rc-hydros.com     - Race Boats
>andy@montanadesign.com  http://www.montanadesign.com - Electronics
>==================================================================