Hi Phil,

On 8/9/05, Phil Keller <Phil@pkeller.net> wrote:
> I am trying to take an analog signal (out of an MP3 player) and
> determine its signal amplitude with the PIC's A2D input.  I would like
> to keep the sampling frequency down (to reduce the clock speed and save
> power) so I thought of adding a large cap outside the PIC in parallel
> with a resistor to ground to hold the level.

The sampling frequency must be in agreement with the Nyquist theorem,
you must have a bigger sampling frequency than 2x of your maximum
analog signal frequency. Adding a cap will create a crude low pass
filter at the AD input, egual with loosing your frequency bandwith of
the analogic signal (the telephone effect)


> 
> The problem is that the manufacture spec says that I must never connect
> Audio ground to Digital Ground.

An analogic or digital ground must be viewed by the user perspective
and not by the manufacturer side. PIC does not have a separately
analogic ground pin, clearly defined in the datasheet. Why ?

Maybe because the system is so small it never need it, if you may keep
the noise level on the "digital" ground acceptable small. So, if you
know how to do this, your PIC digital ground become an analogic
ground.

> 
> I thought of simply having the analog input across a resistor to analog
> ground and using a blocking cap to couple the signal to the base of a
> PNP (The base is DC biased to Vcc/2 and the Collector to digital ground)
> with a diode on the emitter to the PIC and large Cap & resistor to
> digital ground.

Try a simple ferrite choke with one wire  between your analogic signal
ground and PIC "digital" ground. On both sides of your choke a pair of
10nF and 100nF from Vcc to ground. But much better try a clean PIC
supply comming from a linear regulator and care for
communication/digital path (RS232 converter or USB ground is quite
noisy)

cheers,
Vasile

> 
> I tried this with Linear Technology's LTspice and the results were very
> ugly.
> 
> I am sure that this has been done before but I have been unable to find
> anything via Google or in the PIClist archives.  The right keywords have
> escaped me so far.
> 
> All suggestion, pointers, ideas or answers are gratefully appreciated.
> 
> Phil
> 
> 
> 
> --
> http://www.piclist.com PIC/SX FAQ & list archive
> View/change your membership options at
> http://mailman.mit.edu/mailman/listinfo/piclist
>

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