> I use this method in my products and it works OK, but you must low-pass
> filter the PWM signal before sending it to the amplifier (a passive two
> poles RC filter is OK).
>
> The LM386 is OK, I use it myself, but your circuit with one transistor
> is not going to work well.
> You need an amplifier that can force current in both directions into the
> speaker, something with high- and low-side active components, plus some
> means to block DC through the speaker (a capacitor or a well balanced
> BTL amplifier - the BTL amp. is better for low voltage applications).
>
> A very important fact I learned: The processor's digital noise in the
> audible range passes easily through the LP filter and gets amplified. I
> solved this by passing the digital PWM (before LP filter) signal through
> a logic gate (buffer or inverter) powered by its own well filtered and
> regulated power circuit.
>
> If you do things correctly, it is not needed to pump the gain up to
> 200x, the default 20x is enough.
>
> Best regards,
>
> Isaac

Thank you all for helping me
I think the push-pull schematic which was suggested earlier is doing
just that : forcing current in both directions into the speaker. I
will try to implement all of the solutions and figure out which one
works best.
One question abot the fet : is it a much better choice than the
bipolar transistor for this kind of job ?
Jinx : in that transistor configuration i don't think i get current
amplification. I might be mistaken though.
In this weekend i will try to implement what i learned so far. I will
try to make a youtube video with my findings.
Thank you again for all your help, guys

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