1-800-PSAUDIO

# The future of Class D

January 24, 2018
by Paul McGowan

### 6 comments on “The future of Class D”

1. Eric M. says:

I’ve listened to your explanation twice now for class D amplification and am still confused. You mention that the pulse width increases (in time?) to make louder (increase amplitude?). I’m thinking of a sine wave which is probably wrong, but if it is, wouldn’t pulse height increase loudness? Maybe a picture the next time you explain would help … although I think you did do a picture last time!

Thanks

1. It is always a square wave. There are only two states, on and off. What is varying is the amount of time it is on. Longer for louder and shorter for softer.

So yes, it increases in on or off time. The pulse height is always the same.

I found this video on YouTube that’s pretty easy to understand. https://www.youtube.com/watch?v=YmPziPfaByw

Note when you watch this how longer duration pulses produces higher voltage, shorter lower voltage. Now all you need to do is realize that a sine wave, what we want to wind up with, is nothing more than rising and falling voltage. If the PWM happens fast enough then its output voltage rises fast enough to make sine waves up to 20kHz.

If the light bulb goes off in your head after watching this video, and I suspect it will, then all you need to know more than that is that what’s left for the designer to do is get rid of all the sharp transitions between pulses so we wind up with a smooth sine wave. That is the job of the class D output filter, the crux of the design challenge faced by all class D designers.

1. Eric M. says:

OK, that makes sense, I understand the PWM. Here’s another that helped:
So, in the case of a Class D amp, pulse width modulation is occurring to EVERY frequency that is present in the music being reproduced. Is that right?

Thanks,
Eric M.

2. hahax says:

I would think that a fast, strong power supply is critical to class D performance. Theoretically , as in any square wave, you want the signal to go from zero to the design height instantly. That’s impossible so you want it to be as fast as possible. Otherwise the peak will be ‘late’ and won’t be decodes completely accurately by the low pass filter decoder, in a sense a form of jitter analogous to jitter in digital decoding.

3. ifih says:

These Dutch guys do wonders with any class D amplifiers using a feedback ASIC they developed. I have heard them at the dutch XFI show in Eindhoven, really something special.

https://www.axign.nl

Founder & CEO