The title “a pure digital amplifier” is perhaps open to misinterpretation. A digital signal is simply a series of numbers. The numbers can be PCM or PWM or PDM or whatever. It’s not possible to amplify a series of numbers, unless we consider that a digital amplifier is a digital gain function that (for example) multiplies all of the PCM samples by a fixed factor. However, a practical amplifier has to provide an analog output to drive loudspeakers, and so it cannot be a pure digital amplifier. If we had digital speakers, we wouldn’t need an amplifier.
I think the question was really about merging the function of a DAC and PWM for Class D. Suppose we built a pulse width modulator running at 44.1 kHz with a 16-bit digital input. We could simplistically modulate the pulse width by counting cycles of a 44,100 * 2^16 = 2.89 GHz clock. That seems possible, but perhaps not ideal. If we apply the same idea to 96 kHz and 24 bits, we would need a 1.6 THz clock, which is certainly not possible. I wonder if there are other more sophisticated techniques that could give us a digital PWM modulator?
Also, consider that Class D amplifiers for audio normally use large amounts of negative feedback. It’s not obvious how we could provide feedback from an analog output to a digital input without including an additional ADC in the system.
Well, this has been an education. Digital PWM modulators are in fact readily available. As an example, I was looking at Texas Instruments TAS5508, which provides eight channels of PCM to PWM conversion for home theatre applications. Digital PWM modulators use up-sampling and noise shaping to achieve reasonable resolution over 20 Hz to 20 kHz with a practical switching frequency. A designer can add TAS5121 amplifier modules to create a complete amplifier. The parts count is low, and I imagine it’s a cost-effective solution. The PWM output module operates open loop, meaning that there is no overall negative feedback. THD is relatively high at 0.1%. This doesn’t seem like an audiophile solution to me.
Warning: array_search() expects parameter 2 to be array, null given in /var/www/www.psaudio.com/html/wp-content/plugins/styles-layouts-gf-grid-layouts/styles-layouts-gf-grid-layouts.php on line 451
The title “a pure digital amplifier” is perhaps open to misinterpretation. A digital signal is simply a series of numbers. The numbers can be PCM or PWM or PDM or whatever. It’s not possible to amplify a series of numbers, unless we consider that a digital amplifier is a digital gain function that (for example) multiplies all of the PCM samples by a fixed factor. However, a practical amplifier has to provide an analog output to drive loudspeakers, and so it cannot be a pure digital amplifier. If we had digital speakers, we wouldn’t need an amplifier.
I think the question was really about merging the function of a DAC and PWM for Class D. Suppose we built a pulse width modulator running at 44.1 kHz with a 16-bit digital input. We could simplistically modulate the pulse width by counting cycles of a 44,100 * 2^16 = 2.89 GHz clock. That seems possible, but perhaps not ideal. If we apply the same idea to 96 kHz and 24 bits, we would need a 1.6 THz clock, which is certainly not possible. I wonder if there are other more sophisticated techniques that could give us a digital PWM modulator?
Also, consider that Class D amplifiers for audio normally use large amounts of negative feedback. It’s not obvious how we could provide feedback from an analog output to a digital input without including an additional ADC in the system.
Well, this has been an education. Digital PWM modulators are in fact readily available. As an example, I was looking at Texas Instruments TAS5508, which provides eight channels of PCM to PWM conversion for home theatre applications. Digital PWM modulators use up-sampling and noise shaping to achieve reasonable resolution over 20 Hz to 20 kHz with a practical switching frequency. A designer can add TAS5121 amplifier modules to create a complete amplifier. The parts count is low, and I imagine it’s a cost-effective solution. The PWM output module operates open loop, meaning that there is no overall negative feedback. THD is relatively high at 0.1%. This doesn’t seem like an audiophile solution to me.
Yes!
Hi Paul,
I wonder what Jim Kinne would say about this.