But there is of course a DAC also in a FDA. In your view, what is the net save, signal path wise (function block scheme) in a FDA?
//
//
They are PowerDACs.
No analog noise/ripple and what not, no expensive analog signal switching with switches or relays, no analog volume control. Had quite a few of them and they can be quite good. Regardless of what we think it is cheaper so it will gain market. It already does, it is not my view.
See it the other way. Why have digital sources, convert digital to sensitive analog low level signals hampered by many things along the way and then amplify them analog (again hampered along the way) with a high power "open" amplifier (that amplifies everything it sees at it inputs)? Mind you, all this in a predominantly digital world!
Anyway, connecting DACs directly to high power amplifiers without volume control is asking for trouble regardless of personal views. Just the fact that everything at the inputs is amplified with maximum gain is enough trouble to deal with.
No analog noise/ripple and what not, no expensive analog signal switching with switches or relays, no analog volume control. Had quite a few of them and they can be quite good. Regardless of what we think it is cheaper so it will gain market. It already does, it is not my view.
See it the other way. Why have digital sources, convert digital to sensitive analog low level signals hampered by many things along the way and then amplify them analog (again hampered along the way) with a high power "open" amplifier (that amplifies everything it sees at it inputs)? Mind you, all this in a predominantly digital world!
Anyway, connecting DACs directly to high power amplifiers without volume control is asking for trouble regardless of personal views. Just the fact that everything at the inputs is amplified with maximum gain is enough trouble to deal with.
Last edited:
You simply can't go from PCM to PWM without decoding and voltage generation - this is the core of DA conversion and it is dependent on a voltage reference (as are the clocks) - this reference will still share ground with the power bridge also in a FDA so the problem remains even if it has other advantages that you describe. I agree it is probably a good concept that will gain more and more traction - it seems logical.
The problem arises when the shared ground trace has a resistance - if this could be reduced to zero, the problem goes away as now both sides basically has exactly the same potential and no ground current is present. This resistance might be possible to reduce in a FDA and that's good and perhaps a reason of why we like in them!?
//
The problem arises when the shared ground trace has a resistance - if this could be reduced to zero, the problem goes away as now both sides basically has exactly the same potential and no ground current is present. This resistance might be possible to reduce in a FDA and that's good and perhaps a reason of why we like in them!?
//
To be honest, I don’t understand what problem you want to solve in class D using galvanic isolation of PWM, but even if you take FDA without DAC, i.e. take direct digital synthesis of PWM, then the main weak points in such an amplifier will be the power supply, output filter and PWM amplifier.
Yeah, I think you just touched on a subject I tried to point out in my previous post.
Powering the analog and digital sections separately (the red marking separator), with separate SMPSs, still requires a low-impedance ground fill for each section, which you can't establish because you intend to use the SMPS modules with isolated mains earth-to-negative return eyelets at the DC-out side, which prevents you from filtering the noise successfully.
Probably a year ago or so, I was thinking about class D amps, power supplies, and the requirements that will have to be met to make it sound good AND be suitable to integrate with the rest of the analog system, and I pretty much came to the same conclusion as JP: either go digital full-on and use a large, same-potential ground plane, or keep it simple (amplification) in the analog domain and not worry about noise.
Just to be clear - the red mark was to be an optical connection - so a smps feeding each side would not have any galvanic connection thanks to the opto... but again - the FB issue prevents this solution.
//
//
So you don't think that an output stage delivering 15A in a burst effects the delicate ground of an oscillator that's is supposed to do -130dBc/Hz at 0,1 Hz or the Vref of a DAC that is suppose to be accurate to nV... and combine these entities with a ground that has some "substantial" resistance, oppose no "problems"... OK.
//
It is overthinking. That burst of 15A would solve the issue rigorously as you would likely become also instantly deaf.
In simple tests I noticed it was already deafening loud in the mA range! My reference amplifier is 2 x 3..4W on a good day with wind in the back.
When playing with lightning expect thunder 😉
In simple tests I noticed it was already deafening loud in the mA range! My reference amplifier is 2 x 3..4W on a good day with wind in the back.
When playing with lightning expect thunder 😉
Last edited:
I'm not hunting noise here - I'm hunting music signal correlated anomalies like compression, crosstalk and general distortion....
Why only focus on one thing? I never said uncorrelated noise is not a problem....
//
Why only focus on one thing? I never said uncorrelated noise is not a problem....
//
In order to introduce such optical isolation, you will have to break the feedback between the amplifier's power output and the input analog signal, while the optical module itself will add its distortions to the PWM signal. The lack of feedback will also entail a strong influence of the instability of the power supply voltage on the output signal.
Regarding FDA without feedback, the lack of feedback leads to the same problems that I described above.
In my FDA there is PWM feedback before the LPF, I can disable it at the software level, and I did this, in the absence of feedback, the THD level immediately increases, the noise increases and all the garbage from the power supply goes to the LPF output, the difference in sound is clearly audible without devices, the sound becomes dirtier, and the low frequencies are smeared.