Digital audio input for a a digital class D amplifier

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
AX tech editor
Joined 2002
Paid Member
Its Digital because its divided into fixed parts. 1.33 being Digital when meaning 1 1/3.

No. It does represent an analog value thus cannot be digital. Digital is about numbers.
In fact when a processor sends out a PWM signal it is converting the internal number representation into an analog value (most often a voltage derived from the supply voltage) - it's doing the DAC thing.

Jan
 
It is a 100% digital signal, i.e. a number - either 1 or 0. It can just stay 1 longer than a digital signal you have stuck in your mind.

No, it is not 1 or 0. It is some physical voltage. Power supply output impedance. Switch on-resistance. Switching speed. Their non-linearities. All of that is not numbers. All of that is analog. All of that is not constant. All of that is significant but nothing of that is ideal. Under high currents it is extra significant (it is why a "power" DAC is always a poor DAC, much worse than the low-current one). All of that seen by the receiver (the speaker, amplifier stages, driver, power stage, etc)... None of them "reading" strings of numbers. It doesn't matter for them if the voltage is + ~60V or - ~60V. It matters for them if it is 59.8V or 60.7V while for digital signal it wouldn't really matter. Whenever is it digital or analog decided clearly by how the receiver interprets them. If it recovers the original symbol sequence then it is digital. If it relies on some integrated value of an output voltage or current - it is pure analog. PWM in class D is always an analog signal because the receiver interprets it as such. This is not something to argue about except of purely religious reasons.
 
You are claiming PWM can only be analog when others claim it can also be digital.

That's what I don't get either. Revisiting this graphic it appears to be a perfectly valid time multiplexed representation of 1100010 = 98.
 

Attachments

  • pwmmodulation.jpg
    pwmmodulation.jpg
    12.1 KB · Views: 69
That's what I don't get either.
You also prefer to fight straw man? Where your opponent saw that PWM can't be digital signal in general?

Revisiting this graphic it appears to be a perfectly valid time multiplexed representation of 1100010 = 98.
How this 98 is related to the PWM output of a switching power amplifier or a topic of the discussion?
 
Mark Whitney said:
1.33 could in reality be anywhere between 1.325 and 1.334 which is not the same as 1 1/3
1.33, 1.325, 1.334 and 1 1/3 are all numbers. They can all be represented exactly in an appropriate electronic circuit.

Extreme_Boky said:
It is a 100% digital signal, i.e. a number - either 1 or 0.
No. It is either 'up' or 'down', '+' or '-'. It is not 0 or 1.
 
rdf said:
That's what I don't get either. Revisiting this graphic it appears to be a perfectly valid time multiplexed representation of 1100010 = 98.
Using pulse width (an analogue quantity) to indicate a binary value is not fundamentally different from using voltage (an analogue quantity) to do the same. In either case if the width/voltage is greater than some threshold then 1 is read; if less than some other threshold then 0 is read. Using an analogue quantity to indicate a binary value is how digital works. This does not mean that anything which uses a pulse width or a voltage is necessarily digital. On the contrary, it is only when there are thresholds etc. that it is digital; everywhere else (including Class D) it is analogue.
 
AX tech editor
Joined 2002
Paid Member
No, it is not 1 or 0. It is some physical voltage. Power supply output impedance. Switch on-resistance. Switching speed. Their non-linearities. All of that is not numbers. All of that is analog. All of that is not constant. All of that is significant but nothing of that is ideal. Under high currents it is extra significant (it is why a "power" DAC is always a poor DAC, much worse than the low-current one). All of that seen by the receiver (the speaker, amplifier stages, driver, power stage, etc)... None of them "reading" strings of numbers. It doesn't matter for them if the voltage is + ~60V or - ~60V. It matters for them if it is 59.8V or 60.7V while for digital signal it wouldn't really matter. Whenever is it digital or analog decided clearly by how the receiver interprets them. If it recovers the original symbol sequence then it is digital. If it relies on some integrated value of an output voltage or current - it is pure analog. PWM in class D is always an analog signal because the receiver interprets it as such. This is not something to argue about except of purely religious reasons.

Very good, thanks for that explanation.

Jan
 
Extreme_Boky said:
PWM comparator uses a threshold to encode incoming analog signal into PWM digital signal of 1 or 0. According to your understanding, the class D is digital.
I was talking about using a threshold to read the signal, not a threshold to generate the signal.

If you say 'anything over 2.2V (or 0.8us pulse width) is 1' then you are doing digital.

If you say 'X volts at the input is represented by a pulse width of 0.5X us' then you are doing analogue.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.