• Please note that Audiokarma will be offline briefly for updates early on Monday morning, September 28th.

Class D amps and 1-bit DACs

L

Late Boomer

Super Member
Frankly, I had no idea what a Class D amp was. Today I googled "digital amplifier", and turns out that this is it :) I am reading a very down-to-earth article on what Class D amp is, and it says:
Basically, a digital (Class D) amplifier takes an incoming analog signal and converts it into a digital representation comprised of pulse widths. Although there are a number of different design variations, Class D amplifiers are essentially switching amplifiers or Pulse Width Modulator (PWM) designs. The incoming analog audio signal is used to modulate a very high frequency Pulse-Width Modulated (PWM) carrier that works the output stage either fully on or off. Later on, this ultra-high-frequency carrier must be removed from the audio output with a reconstruction filter so that no ultra-high frequency switching components remain to corrupt the audio signals.
Click to expand...
Um, but PWM (and PDM) are implementations of delta-sigma 1-bit (and, later, multiple-bit but still serial) DAC? So, this is a natural implementation of a built-in amp for CD player or a DAP with a 1-bit DAC?

I have two models of Panasonic portable CD players, they look almost the same, only the newer model says "D-Sound". Reading other forums, it seems that D-Sound includes, beside other things, a digital amp (Class D?), while earlier models, supposedly, used old-school analog amps? People "in the know" say that the earlier model, without D-Sound, was the last good portable CD player made by Panasonic, and everything made later is digital crap :) Amazing how people latch on anything analog even when playing CDs.

It just seems natural to me to use existing bitstream as a source of amplification, and to use the feedback loop to compensate for inconsistencies in the discrete parts. So, I guess, Class D amp is a smart thing.

For analog sources, Class D amp is, basically, a delta-sigma ADC/DAC plus the analog amplification stage?
 
Register to hide this ad
mhardy6647 said:
Class D amplifiers are not digital per se. They work by entirely analog means, essentially chopping the signal.
Click to expand...

The output drive is digital, though, since it is in only one of two states: on or off.

Amplifier classes are described according to their output drive mechanism (e.g. Class A, B, AB, etc).
 
cpt_paranoia said:
The output drive is digital, though, since it is in only one of two states: on or off.

Amplifier classes are described according to their output drive mechanism (e.g. Class A, B, AB, etc).
Click to expand...

I think it's clear that the D does not stand for digital.
 
Last edited:
spark1 said:
What does "output drive" mean?
Click to expand...

See the handy summary quoted by mhardy: it discusses amplifier classes in terms of the operation of the output devices: that's what I meant by 'output drive' (I think I intended to type 'drivers').

The 'D' in Class D does not stand for 'Digital'. That is true. However, its output devices do not operate in the linear region, unlike the output devices in the other Classes mentioned. In that respect, they act in the same way as a CMOS output stage; pulling hard to one rail or the other. Not operating in the linear region means that they drop very little voltage whilst carrying current, so dissipate very little power, which is why they are so efficient.

The similarity to a CMOS output, and the on-off, switching nature of their operation, is why they are often called 'digital amplifiers'.

My point about a 'power DAC' was that one might envisage a Class D amplifier where the output power switching stage was fed directly by a PPM or PDM bitstream generated from a PCM sample, rather than from an analogue input and a ramp comparator.
 
Last edited:
I saw the summary (and have seen it before). Your unique terminology is what threw me off.
 
spark1 said:
Your unique terminology is what threw me off
Click to expand...

'Drive' or 'driver' is not my unique terminology; it is a term widely used to describe power devices, in a generic sense, whether that be an amplifier, motor controller, logic device, I/O device in a processor, etc.
 
gvl said:
It’s binary at best, but since it doesn’t actually encode numbers in base2 it’s not digital.
Click to expand...

The output of a Class D driver stage is a series of 1s and 0s, providing a PWM digital representation of the signal. It is not an analogue driver stage. Once filtered, the output is analogue.

PWM is a means of digital representation of an analogue signal. PPM and PDM are other digital representations. These are digital conversion methods that don't use discrete samples like PCM. These have a long history in communications, such as CVSD. They are all digital representation methods
 
I have seen a few articles on class D amp modules being used to salvage receivers with STK IC's that are no longer available and maybe not compatible with the STK-050 replacement developed here in AK, such as the STK-039 and SKT-016. The articles make it sound like some Class D modules are sort of plug and play, but I have not been able to decipher the jargon enough yet to understand all the details.
 
cpt_paranoia said:
The output of a Class D driver stage is a series of 1s and 0s, providing a PWM digital representation of the signal. It is not an analogue driver stage. Once filtered, the output is analogue.
Click to expand...

The 0s and 1s here do not encode the signal, their duration does, the duration value is continuous within its allowed range. I realize PWM is categorized as digital everywhere but this is really a misnomer as the carrier of audio data is analog (or non discrete) in nature.
 
gvl said:
The 0s and 1s here do not encode the signal, their duration does
Click to expand...
Just like in a 1-bit DAC. I am having hard time separating a DAC from an amp, but it seems to me that this stage still leans more into representation via digital (a bunch of pulses, a number of which - not voltage or current - represent the amplitude) rather than into amplification of actual voltage and/or current. Just because this approach was invented before 1-bit DAC became a thing does not make it non-digital, IMHO. The letter "D" also being the first one in "digital" is just an [un]happy coincidence.
 
You must log in or register to reply here.
Back
Top Bottom