Tiny 6x300W amp

[berni8k]

Yes but the square wave frequency will create a broad spectrum of noise that some of it will get trough the filter. At very low powers this noise shows up a lot but its too small to be audible as noise.

 

[AndrewT]

This D-class has pretty much no THD distortion, what your seeing on that graph is the noise of the amplifier. Because of the switching frequency they create inaudible noise at 300Khz that you see here it can be seen very well because the THD is near nothing.

how does a music or other natural audio waveform that is reconstructed from an ~300kHz square wave with variable mark/space ratio have no distortion?

 

[kenpeter]

Because the pulse width is still wholly an analogue signal...
Obviously current mode output when no feedback is applied.
But still you can vary the tiniest details in the pulse width.

You are switching two signals that largely cancel each other.
Only the bridged difference is apparent to the load, and that
difference can appear to switch on/off in much narrower
pulses than the dwell time might seem to imply.

 

[berni8k]

Digital being eater 1 or 0 takes awey the THD effect of some elements in a amplifier. Also there is all sorts of compensation circuity inside the chip to reduce THD (It has internal feedback loops and so on). As far as THD+N goes as long as its under 0,1% its pretty much perfect for our human hearing.

 

[kenpeter]

No, we have merely changed the domain of error to pulse widths.
Which will eventually be integrated back into a voltage waveform
with those same errors, when it must inevitably drive the load...

And unfortunately, when driving a pure current, the voltage error
is very much related to load... And no loudspeaker is ever 8 ohms,
especially not while its moving...

I had forgotten that one had an integrated feedback.

 

[wg_ski]

how does a music or other natural audio waveform that is reconstructed from an ~300kHz square wave with variable mark/space ratio have no distortion?

When the distortion that you do get isn't where you "expect" it to be (at 2f0, 3f0,..... nf0). At least not at the same levels and spectral distribution as you're used to hearing from typical amps. What you get more of is noise, which is interpreted differently.

 

[berni8k]

Some kind of noise can sound good tho, just like the 2nd harmonic of tubes a lot of people absolutely love.

I seen a lot of cases where people take LM3886 amps and suddenly replace the usual transformer PSU with a swicthmode and suddenly they say it sounds a lot better. Now the only thing the switchmode is probably doing is adding noise in to the supply that created the nicer sounding sound by a good kind of distortion.

I dont relay care for perfect amps with 0,001% THD and stuff. If it sounds good it is good. Audio inset about reproducing the sound in true perfection its about sounding good to you.I dont relay have any super trained ears but this D-class sounded nice to me, and thats what matters.