[sdiy] ot: class d audio amp

[Czech Martin]

well, I think the the output LRC lowpass should have a first stage
with air coils and low high frequency loss capacitors.
These won't be appropriate for 20kHz, but better at 200kHz,
this is where the main frequency is.
Some kind of "guard" filter that protects the following
audio lowpass from beeing driven with too much RF.

Another thing is the diodes that short the first inductance
kick back during switching. I guess Schottky types.

For improvement I have seen "dithering",
pulse duration & pulse frequency modulation
and several feedback mechanisms.
Might be difficult to implement with simple (DIY)
means. Anyway, I'm looking for sub bass amplification,
where efficiency of portable speakers is very bad.
So the bandwidth is perhaps 1000Hz only.

The other thing is a portable noise source, distortion
doesn't really matter in this case. But battery life
does.

Many thanx for all the input I got!

m.c.

> -----Original Message-----
> From: owner-synth-diy at dropmix.xs4all.nl
> [mailto:owner-synth-diy at dropmix.xs4all.nl]On Behalf Of Oakley Sound
> Sent: Donnerstag, 23. Oktober 2003 11:38
> To: synth-diy
> Subject: Re: [sdiy] ot: class d audio amp
> 
> 
> > Class D audio amps are pure switchers, in that the "amplifier" is
> either on or off, resulting in higher efficiency (no 
> crossover) compared
> to Class A, B, AB amps...
> 
> Whilst they do display potential to be more efficient, care 
> needs to be
> taken to ensure that not too much power is wasted in driving the LC
> filter at the output stage. Higher switch frequencies also require
> quality ferrites for the L parts of the filter.
> 
> Things have changed since I worked on a 2KW class-D back in 
> 1987. It was
> mainly intended to produce a 'pure' (1% THD) 31.25KHz sine wave
> oscillator at 2KW. FETs, high voltage caps and Ferrites have all
> improved these days. There's at least one company offering evaluation
> boards for class-D amps. But then, my colleague and I at Marconi
> Instruments were breaking new ground at that time, we even 
> had visits to
> several UK universities talking about our work. Unfortunately, the
> parent project was shelved due to commercial pressures. [This 
> pissed me
> off so much that I left, only to loose my job two months later when my
> new company's directors embezzled the profits!]
> 
> I remember reading that the chap that designed the first 
> class-D amp in
> the 1930s had to work with relays! We had MOSFETs, six of them I think
> in H formation running off rectifed mains. The compound Cgs at nearly
> 10nF meant that the gate drive current was peaking at around 
> 5A. I think
> we used four DS0026 gate drivers chips. But our biggest problem, along
> with the new fangled EMC tests, was with high values of d(Vds)/dt; it
> would break through into the gate and the FET would stay on too long.
> BANG!! Cartridge feed FET systems were imagined.
> 
> Regards,
> 
> Tony Allgood                         www.oakleysound.co.uk
> 
> Oakley Modular Synthesisers      Penrith, Cumbria, England
> 
>