Re: PWM Amp Design

John Woodgate <jmw@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote:

>I read in sci.electronics.design that Nico Coesel <nico@xxxxxxxxxxx>
>wrote (in <4258ea0c.670791226@xxxxxxxxxxxxxx>) about 'PWM Amp Design',
>on Sun, 10 Apr 2005:
>
>>Sounds like something I designed a few years back (with less power
>>output though) for an induction loop amplifier. The compensation filter
>>is a pain in the ass where it comes to dissipation.
>
>You put a Zobel network (series RC) across the loop? There is a much
>better solution described in BS 7594, using a series capacitor and one
>or two resistors. But these days, there is usually no reason not to use
>a current-drive amplifier.

I used a Zobel network indeed (I more or less followed the HIP4080
appnote). But when it comes to class D amplifiers I came to the
conclusion that trying to make a perfect filter network using bessel
or butterworth is a futile excercise. The inductor and capacitor
values will vary too much because of the changing currents and
voltages.

Also, a zobel filter depends highly on the load, so you would need to
fit a different filter for every application or make it adjustable.
In case of a induction loop amplifier, the number of tables and chairs
and their location in the room would enter the equation :-))

In other words, butterworth, bessel and zobel look nice in theory, but
their practical use is very limited when it comes to class D
amplifiers.

I found out that using some inductors and capacitors to smear the
pulses into something that looks like a signal and using a common mode
filter as a final stage to get rid of the HF components works much
better.

--
Reply to nico@nctdevpuntnl (punt=.)
Bedrijven en winkels vindt U op www.adresboekje.nl
.