• TalkBass has been independent since 1998. Add your voice.
    Create a free account to reply to discussions, view embedded media, and browse with fewer display ads.
    Join free Log in
    Want zero display ads or expanded classifieds tools? Compare plans.

Class D or AB

#42
Bob Lee (QSC) said:
No, that's not digital. A class D amp doesn't switch between 0 and 1, but between a positive voltage and a negative one.
Click to expand...

Your distinction is arbitrary. It is switching between discrete
levels - the definition of digital.

When I'm working with digital electronics, "on" is generally
referred to as a '1'; "off" is generally referred to as a '0'.
It doesn't matter what the voltage is or if I have multiple
levels of voltage (e.g. positive, zero, negative).

It's still digital.
 
#43
warnergt said:
I agree. This "pulse duty factor" is another way of saying how
long we have a '1' versus how long we have a '0'. There are no
duty factors in an analog system. The level of the pulses ('0' or '1')
is a defining characteristic. These 0's and 1's are needed to
define the output signal.
Click to expand...

So an analog synth generating a square wave with adjustable pulse duty factor is digital? I think not.
 
#45
"Discrete levels" is not "the definition of digital."

Class D is not switching between 0 and 1 or on and off. It is switching between positive voltage and negative voltage, apportioning the output voltage through varying time duration instead of varying voltage division.
 
#47
Here's a class-ab output from one side

amp49.gif


Clearly that side is off at the bottom, and then on to some level the rest of the time.

Putting the other side together, it shows just the opposite, it's off when the other is on.

Figure9D.gif


Off, then some state of On.
Each side, off/on, off/on ...

But this isn't digital and nobody would argue that it is.

Class-d is On + some period of time, integration occurs in the output filter.

400px-Pwm_amp.svg.png


In either case, class-ab or class-d the speakers see the same signal.

I think we're stuck with class-d being called "Digital" because of marketing, but when digital amps come along they will be stuck with how to market them. "Truly Digital"

Guess what though, class-ab, class-d can all exceed the threshold of human hearing, "Truly Digital" won't sound any better. Audiophiles and golden ears will think so still.
 
#48
I think that many people would describe a digital signal as something with a certain "resolution", a quantized signal, something with a sample rate. There is nothing like that going on in a class D amp, therefore most people wouldn't consider it a digital amp.
 
#49
JimmyM said:
I guess you missed this question...

How is a class D amp with an SMPS any more or less digital than a class AB amp with an SMPS?
Click to expand...

Do they have anything digital in their operation?

SMPS is not digital in itself. You could have some control circuitry about it that is digital.

Class AB isn't digital but it also could conceivably be designed with some digital support circuitry. Class D can be done digitally but almost always is done with analog modulation.
 
#50
Bob Lee (QSC) said:
"Discrete levels" is not "the definition of digital."
Click to expand...

The first sentence in the definition from Wikipedia:
A digital system is a data technology that uses discrete (discontinuous) values.

Bob Lee (QSC) said:
Class D is not switching between 0 and 1 or on and off. It is switching between positive voltage and negative voltage, apportioning the output voltage through varying time duration instead of varying voltage division.
Click to expand...

As I said, the voltages are arbitrary. They are switched on or
off. They are never partially on as they would be in an analog
system.
 
#51
warnergt said:
The first sentence in the definition from Wikipedia:
A digital system is a data technology that uses discrete (discontinuous) values.
Click to expand...

A data technology, you say? Data values (i.e., numerical encoding) are not the same as voltage values.

As I said, the voltages are arbitrary. They are switched on or
off. They are never partially on as they would be in an analog
system.
Click to expand...

The rail voltages determine the output. That is not the case in a digital system. Noise on the rail voltages will appear directly as noise on the output. Again, unlike a digital system.
 
#52
dincz said:
So an analog synth generating a square wave with adjustable pulse duty factor is digital? I think not.
Click to expand...

You're approaching a gray area where analog meets digital but
I would say you are correct here.

The output voltage of a your square wave synth generator
is critical; otherwise, it affects the output sound. While it is a
square wave, the level of the signal still contributes to the
output signal. That is, you are still in the analog domain with
your synthesized square wave.

In digital, as in the class D amplifiers, that signal is merely "on"
or "off." For example, it wouldn't matter if the voltage varied
between 4 and 5 volts, it would still produce the same output.
All that matters is if it is '1' or '0' because it is truly in the
digital domain.
 
#53
Alban said:
I think that many people would describe a digital signal as something with a certain "resolution", a quantized signal, something with a sample rate. There is nothing like that going on in a class D amp, therefore most people wouldn't consider it a digital amp.
Click to expand...

Those people would be describing a subset of what constitutes
a digital signal.

Quantization and sample rates are not requirements for digital
systems. I could have a simple Invalid Link Removed nand gate that turns on
an L.E.D. when sensors detect that somebody opened my
front or back doors. That circuit has no quantization beyond
"on" and "off" and it has no sample rate; it operates continuously.
There is nobody who can argue that the 7400 is an analog
device. It is unequivocally a digital device.
 
#54
Bob Lee (QSC) said:
A data technology, you say? Data values (i.e., numerical encoding) are not the same as voltage values.
Click to expand...

OK. What's your point?

Bob Lee (QSC) said:
The rail voltages determine the output. That is not the case in a digital system. Noise on the rail voltages will appear directly as noise on the output. Again, unlike a digital system.
Click to expand...

Possibly, but you've gone a little too far in the signal path in
an attempt to prove your point. Let's go back to where
the signal is truly digital.

What is the signal driving those MOSFETS which switch
the rail voltages?

It is either "on" or "off". Does noise on this switching signal
appear on the output? If designed properly, it shouldn't.
Those MOSFETS should be full on or full off. Otherwise, they
get hot and you have defeated the whole purpose of using class D.
 
#55
The real important consideration here is how the analog signal is being represented.
It is not being represented by a quantized or discrete value. It is being represented
by another analog quantity, the proportional output of the power switching circuitry.

At no point is the signal voltage being represented by some digital code. You could
convert the pulse width into a digital quantity if you wanted to (just measure it's time
duration), and then use that digital code to determine the output width of the pulse.

But none of that is being done. The signal is an analog, continuously variable quantity
and the modulated pulse width is also an analog quantity.

If it was digital, you should be able to store the information as a sequence of codes.
But how to you do that without first converting the pulse width into a digital numeric
value?
Again, there is no representation of the signal as a digital code.
 
#57
megafiddle, you make many good, accurate points. And I'm
with you until your fourth paragraph.

megafiddle said:
If it was digital, you should be able to store the information as a sequence of codes.
But how to you do that without first converting the pulse width into a digital numeric
value?
Again, there is no representation of the signal as a digital code.
Click to expand...

Let's go back to my door sensor example. How do I store
that information as a sequence of codes? You can do that no
more easily than you can store the class D amp data. But the
door sensor uses a digital IC at its core. Can you say that is
not digital? It is certainly not an analog circuit.

Let's revisit Wikipedia and see if we can get some
arbitration on this disagreement.

Wikipedia recognizes Morse Code as a digital system.
Morse Code uses mere "ons" and "offs" modulating the width
of the "ons" and controlling the timing between each. I'll be
damned if that doesn't sound exactly like the controlling
signal in a class D amp.
 
#58
warnergt said:
OK. What's your point?
Click to expand...

That voltage amplitude is not a digital quantity.

Possibly, but you've gone a little too far in the signal path in
an attempt to prove your point. Let's go back to where
the signal is truly digital.
Click to expand...

There is no such point.

What is the signal driving those MOSFETS which switch
the rail voltages?

It is either "on" or "off". Does noise on this switching signal
appear on the output? If designed properly, it shouldn't.
Those MOSFETS should be full on or full off. Otherwise, they
get hot and you have defeated the whole purpose of using class D.
Click to expand...

The pulse train signal is positive and negative, not "on and off." Positive turns on the positive rail voltage for a time. Negative turns on the negative rail voltage for a time. These rail voltages don't represent anything abstract, as in a digital circuit. They are the voltage presented to the integration filter. The MOSFETs are indeed turned full on and full off, but that would have to be in a digital context to be considered digital. Analog voltages can be switched on and off, too, and that's what's actually happening in a class D amp.
 
#59
First off, Wikipedia isn't the be-all-end all source, especially when dealing with the finer details.

Second, we've had this discussion before, and again, the engineers among us agree that class D isn't digital. When dealing with varying voltage as class D isn't just using on/off but also time as a parameter of how it works.
 
#60
seamonkey said:
Class doesn't matter at all.
They both sound the same.
The difference is in features of one amp over the other.

How does it matter to you?
Click to expand...
I've heard some claim that the LittleMark lll and LM 800 sound different. I'm not sure why other than the different power supplies??
 

Latest posts

Top Bottom
Back