Author Topic: Class D explained  (Read 18274 times)

Offline Simango4

  • Trade Count: (+14)
  • AVForums Veteran
  • ***
  • Posts: 229
Re: Class D explained
« Reply #15 on: April 26, 2018, 09:17:46 AM »
Awesome explanation  :dop:
Music is What Feelings Sound Like...
You Don't Stop Gaming Because you Grow Old, You Grow Old Because you Stop Gaming...

Offline Ampdog

  • Trade Count: (+1)
  • AVForums Veteran
  • *****
  • Posts: 7,058
Re: Class D explained
« Reply #16 on: April 29, 2018, 12:22:08 AM »
And here I am, noticing it only now!!

I could agree with the quality of the explanation, particularly in the absence of drawings. And may I add: Of particular 'nostalgic' interest to me: I wrote my degree thesis on this, all of 52 years ago!  (Believe Mr Watts was still..... er, well, not quite there (or is it here) yet.)

Seriously, one matter is evident: In this case it is a matter of the more components, the better - not the analogue way (component count in well-designed circuits of course). All should understand the 'digit-using' (to try generate a general term) way better.  And the all-underlying dual advantage often overlooked: Semiconductor linearity is not important, and dissipation low: One either has high-voltage-zero-current conditions, or high-current-zero-voltage ones.

Possibly horribly superfluous by now, but a.o. I used a ball/weight attached to a suitable length of elastic for demo purposes. If you move the holding hand up an down at equal intervals, if fast enough, at a stage the weight will be at rest at the half-way position, the elastic support hand motion simply too fast to be followed - one can demo this in various ways.

But here we are in 2018, and the basic technology utilised increasingly because of its many advantages over pure analogue.

[Again NOT to distract form Mr Watts' description, just one more surprising phenomenon encountered on the way is the effect of the (mandatory) output filter. For this illustration a sketch is essential, but I will try.] This is the phenomenon that any wave train emerging after the D/A filter, is not an average of the input pulses kind of smoothed out, It is a perfect sine wave of the input repetition rate!  One gets this by virtue of a very rapid cut-off output filter (not necesarily rapid rate!). One of the favourite demos is that at a pulse rate of only two pulses (never mind how short) per highest audio freuqency at input, a very smooth sine pulse emerges.  In simpler terms, it has been illustrated that at a sampling rate of only two pulses per highest audio frequency, quite understandable speech can be reproduced!  This is far away from 192 Hz sampling rate and not very hi-fi, but to illustrate the basics.

For those still not au-faix with matters 'pulsating', pulse technology can be quite rewarding. More so because increasingly ICs combining many of the basic functions are made available.  (See how carried away I just became!)

Thans to Shovner for initially posting Mr Watts' explanations.
Audio must be the only branch of engineering where lack of basics' knowledge is considered a superior form of wisdom. (Anon)