But the PWM is a continuous varying analog signal.
The digital signal comes in at byte- wide (or word or whatever) number chunks clocked in one at the time.
So if the input is digital and what results is analog, it must be DA conversion right there.
Jan
The digital signal comes in at byte- wide (or word or whatever) number chunks clocked in one at the time.
So if the input is digital and what results is analog, it must be DA conversion right there.
Jan
Regarding the place of conversion: The conversion actually takes place in the PWM stage. From the information theory point-of-view the reconstruction filter is just cosmetics.
Regards
Charles
Regards
Charles
Last edited:
This is true. However from the information theory point-of-view DA conversion is impossible, since information cannot be generated. So this point of view can not give an answer about the place of the DA conversion agreed by usual terminology. Once a signal converted to digital, it remains digital forever no matter how you convert it.
But the most common sense has in the answer given by Kjeldsen: reconstruction filter can be called DAC in most of the aspects.
Last edited:
There is nothing analog in a digital PWM module in any sense. Neither on inputs nor on output. Neither time, nor voltage, nothing. There are discrete time steps on output, inbetween values are impossible/invalid.
Last edited:
Didn't we agree that this kind of amplifier is a power DAC ? Digital to Analog Converter ? So from some point the signal must be regarded as analog -even though the voltage steps and the temporal resolution are more or less discrete ? Or would you call a reproduction chain consisting of CD player, amplifier and speaker digital ?
So wouldn't it be logical to define the output stage and modulator combination as the place where the conversion happens ? I.e. the converision of numbers into voltage and current (including analog errors like voltage sag, rdson, nonlinearities, jitter ......).
Regards
Charles
So wouldn't it be logical to define the output stage and modulator combination as the place where the conversion happens ? I.e. the converision of numbers into voltage and current (including analog errors like voltage sag, rdson, nonlinearities, jitter ......).
Regards
Charles
PWM may just be one of those fascinating edge cases that keeps things interesting. The protocol is perfectly capable of 'transferring numeric information' and operating purely in the analog domain no matter how strenuously some try to stuff it into a single box.
The classic class-d amp IS definitely analog. But the last few posts were about amps like the TI variant that is using a digital modulator.
Regards
Charles
Regards
Charles
Rdf,
You are right.
Analog signal and analog device may be different. Is a cable digital or analog? What if I transmitt S/PDIF on that cable? Or through a fast OPA? Making strong statements requires full understanding the terms we use and specifying what is the object of the statement. Signal alone is never analog nor digital. Sometimes devices are also cannot be categorized. It depends on the point of view. All digital signals can be considered analog also, and vica-verse. This doesn'mean distinguishing is impossible, just need a specified point of view.
If somebody can tell why he thinks a special arrangement is not digital, than its fine. Saying everything is analog is ignorance.
I stated a definition that gives the less number of cases left ambiguous, but it was ignored.
You are right.
Analog signal and analog device may be different. Is a cable digital or analog? What if I transmitt S/PDIF on that cable? Or through a fast OPA? Making strong statements requires full understanding the terms we use and specifying what is the object of the statement. Signal alone is never analog nor digital. Sometimes devices are also cannot be categorized. It depends on the point of view. All digital signals can be considered analog also, and vica-verse. This doesn'mean distinguishing is impossible, just need a specified point of view.
If somebody can tell why he thinks a special arrangement is not digital, than its fine. Saying everything is analog is ignorance.
I stated a definition that gives the less number of cases left ambiguous, but it was ignored.
Some years ago I read about a digital spring, that could be driven directly by a PWM signal. I'm not talking about using a voicecoil as reconstruction filter, but that the "spring" would trigger directly on the PWM signal.
I'm not sure if its this one from 1...Limited.
Ceramic Piezoelectrics - Springs and Loudspeakers
I'm not sure if its this one from 1...Limited.
Ceramic Piezoelectrics - Springs and Loudspeakers
Yes, it is a very close definition. However have its own error, and not exclusive.
If we overlook the fact that the term DAC is questionable by itself - yes, at some point you can call it analogish. But not at a point where it is obviously discrete. Analog vs. discrete, these terms are antagonists. If you call a discrete signal analog, then the essence of the categorisation is lost.
In info point of view - yes. But this is a stupid way of categorisation. As you know CDs are already categorized by different parameters. You know, DDD, ADD, etc. The force to categorize heterogenous systems to only 2 catagories implies inherent contradictions.
No. What reason do you see? It doesn't happen anywhere in strict meaning. If you are keen on defining where it happens, then you must first define a distinquishing feature.
Numbers? Numbers are only logical assignments. And there are numbers as logical assignments on both sides. Why do you think 10010110b is more a number than the equivalent 150 (width of pulse defined in clock time)?
And pleas try to forget about that @*'$$;:::&-😡 habit of thinking rational, irrational, etc... numbers are not numbers. We are not in 1000 Before Christ. 2.5654424463544552 or pi are also numbers. And every voltage can be described by a number and a unit. A digital voltage can be for example 0 or 1 (set of numbers) times 3.3V (unit), an analog for example [-3.3..+3.3] (continuous domain of numbers) times 1V (SI unit in this case). In both cases there is a phisical and a logical signal. These are not converted by any electronics, only by confused minds.
The comparator converts several time dependant functions to a different time dependant function. If the resulting function is quantized, then digital by definition. It is this easy in this case. No need to make everything fuzzy by mixing other signals into it.
You are confusing binary and digital. Binary is some example of the digital number systems, but not the only one. Do you think 1, 2, 3, 4, 5, is not digital? It is so digital that the name is originated from here. Digitus= fingers and/or thumb. Digital: the way you counting with your fingers. Basic stuff.
This argument keeps cropping up. To me the issue is quite simple: an analogue system uses analogues (e.g. voltage, pulse width) to represent the signal; a digital system uses some discrete representation of numbers to represent the signal. Note that a digital system could (but rarely does) use an analogue representation of the numbers; more commonly, some analogue systems may use digital techniques to prepare or process an analogue signal. A Class D amplifier which has a 'digital modulator' is still an analogue system, even if the makers want to call it digital.
I suspect the confusion arises because most digital systems use binary, so people assume that every system with just two signal levels must be digital. Class D is an example of a switch-mode analogue system. Accepting a digital input does not make an amplifier 'digital' (except for marketing purposes); somewhere between the digital input and the analogue output there will be a DAC. If the DAC is part of the output stage so just before the output filter there are discrete levels then, and only then, do we have a real digital amplifier - but this would not be Class D.
I suspect the confusion arises because most digital systems use binary, so people assume that every system with just two signal levels must be digital. Class D is an example of a switch-mode analogue system. Accepting a digital input does not make an amplifier 'digital' (except for marketing purposes); somewhere between the digital input and the analogue output there will be a DAC. If the DAC is part of the output stage so just before the output filter there are discrete levels then, and only then, do we have a real digital amplifier - but this would not be Class D.
DF96,
I dont understand what do you mean
Nobody said it does.
Always we are referring to such circuit as digital amp, and nothing else.
Tha power stage of it is made of 2 controlled switches. ClassD by definition. What else Class could have been?
I dont understand what do you mean
do you think voltage or pulse with was inherently analog? And can you show an electronic digital system that doesn't use voltage (or current) to represent the signal?
Nobody said it does.
Always we are referring to such circuit as digital amp, and nothing else.
Tha power stage of it is made of 2 controlled switches. ClassD by definition. What else Class could have been?
Pulse Width Modulation is not a signal, but a process or circuit. There are analog and digital ones also. I already linked a digital one. The digital one is not continuous. Maybe some days of writing will be enough to pass through this information.
But the result is digital. Discrete. Quantized.
Not necessarily related to the fidelity of the amplified signal compared to the source signal, but I've noticed that I can have two amplifiers, one an old school class AB solid state amp rated at 200 watts and a new class D also rated at 200 watts and all other things being the same, e.g., the same bass guitar, the same cables m, the same 4x10" cab, the same output impedance jack plugged into in the back of each amp (8 ohms), etc....why does the older class AB amp sound noticeably louder and register about 3-4 dB louder on my decibel meter when placed at one meter from the cab?
Sent from my iPhone using Tapatalk
Sent from my iPhone using Tapatalk
- Status
- Not open for further replies.