Today many run their DACs directly connected to their class-d amps. I do. Amps with several 100 watts output and over 10A amps out.
I would guess that the signal ground in these amps are not the cleanest one could see. And here comes my culprit... the delicate ground in DACs where both clock ground and Vref grounds are tied up to the dirty class-d ground. Probably not ideal.
Why not make the PWM signal interface (red marking) after the "C" optical and let the input part float compared to the output stage. The only harmful part left sharing ground would be the triangle generator.
I bet such a combo of DAC and amp would be quite better sounding... what do You think?
//
I would guess that the signal ground in these amps are not the cleanest one could see. And here comes my culprit... the delicate ground in DACs where both clock ground and Vref grounds are tied up to the dirty class-d ground. Probably not ideal.
Why not make the PWM signal interface (red marking) after the "C" optical and let the input part float compared to the output stage. The only harmful part left sharing ground would be the triangle generator.
I bet such a combo of DAC and amp would be quite better sounding... what do You think?
//
Yes. When I have done this, I got slight hum and was in the end required to hook up the ground between them to have it quiet - don't know what the problem was. This was between a Soerkris DAC run as differential / balanced, a Lundahl line trafo and a Hypex Nc400...
Lets check with this guy as hes amp is claimed to use no global FB... so paging @flmhhh to see what he says - would this concept work in the SGD 801-M with retained measured distortion figures?
//
Lets check with this guy as hes amp is claimed to use no global FB... so paging @flmhhh to see what he says - would this concept work in the SGD 801-M with retained measured distortion figures?
//
Balanced inputs (when implemented right) combined with internal volume control in the power amplifier normally help a good deal. Input transformers also may be of good use too but may result in side effects. In only a few cases with very cheap and unbalanced stuff I noticed such issues, in general it is not an issue to worry about with well designed devices. Leakage of internal SMPS, RF stray in and stray out (!) and correct grounding are issues today.
"Open" power amplifiers so amplifiers without internal volume control connected directly to sources are way more sensitive to this. Besides that there are more good reasons to use internal volume control. It is a simple test too, taking not much time.
"Open" power amplifiers so amplifiers without internal volume control connected directly to sources are way more sensitive to this. Besides that there are more good reasons to use internal volume control. It is a simple test too, taking not much time.
Last edited:
Maybe you could try to be more friendly and make stuff clear in the first post describing the issue (as common).
Last edited:
Did you measure the ground in the DAC wrt to a well know stable reference, like a an earth rod etc, while playing 50 watt bass transients in the speakers? Or how do you know it is not an "issue"?
//
//
Haha - coming from U 😉 ... I misinterpreted your answer to comment on my trials with Nc400 - I didn't see it was a general answer to the topic.
On the other hand, my experience was actually that a trafo didn't do it in my case... but I'm sure it can.
//
Two wrongs don't make a right. I also think I reacted to the point but was met with a hostile approach straight away. If it makes you feel good that is OK 🙂
On topic: I am pretty sure you will conclude connecting a high power class D amplifier without internal volume control directly to a source is pretty close to a system/user error.
The point maybe is that recent electronic developments solve some issues like high cost/reducing weight but also result in side effects and challenges to get stuff 100 % silent. Issues both hardly avoidable and solvable by the average DIYer.
On topic: I am pretty sure you will conclude connecting a high power class D amplifier without internal volume control directly to a source is pretty close to a system/user error.
The point maybe is that recent electronic developments solve some issues like high cost/reducing weight but also result in side effects and challenges to get stuff 100 % silent. Issues both hardly avoidable and solvable by the average DIYer.
Last edited:
FDA's also solve such possible issues quite adequately. No intermediate conversion to analog anymore so no more analog input stage in the class D amplifier. It only becomes analog at the very output stage. Good experiences with various FDA's I must say. If these have SPDIF transformers and/or Toslink there either is no galvanic connection either with the source. Trouble usually starts with USB.
So the wait is for a serious digital audio player with built in FDA and no need for external interfaces. Like NAD M33 but then affordable. The future for digital audio seems to be in all-in-ones for various reasons.
So the wait is for a serious digital audio player with built in FDA and no need for external interfaces. Like NAD M33 but then affordable. The future for digital audio seems to be in all-in-ones for various reasons.
Not the "problem" I presented as the DAC in these chips share ground with the output stage. So rocky references also in these chips. To put it all in one box doesn't help of course...
//
//
My basis here is the violent jolting of ground in an unsymmetrical power system when playing the big drum. Think about it for a while...
//
//
Of course I did. It can be mitigated in various ways but having normal output power class D amplifiers is a good first step. In normal homes and having normal to high efficiency loudspeakers a few Watts are usually enough. Like with linear stuff having stuff in extremes poses extra challenges.
You should think about your suggestion from the power supply angle... signal isolation means nothing if you do not isolate the power supply rails' returns. How are you going to do that? And while thinking about it, consider the full supply chain length starting from the (ending at the) mains ground return. See what you come up with and then post another block diagram containing the above...
Well having isolated power feeds to the input section and the output stage is not so hard - just use two smps units - trivial. And make it dubbel isolated class 2 an skip the mains safety ground - QED.
But, it is in fact the feedback loop that is the party pooper as it needs to be from the points where things are analog and that's the input and the output. So for that reason, it is a no go I'm afraid.
//
But, it is in fact the feedback loop that is the party pooper as it needs to be from the points where things are analog and that's the input and the output. So for that reason, it is a no go I'm afraid.
//
Isolated feedback with ADCs....to DACs....It will be complicated with more drawbacks to solve than things solved.
Better keep the signal path digital without any DAC or intermediate conversion. No DAC = no DAC issues. No analog in between = no analog issues. FDA will be the future.
Better keep the signal path digital without any DAC or intermediate conversion. No DAC = no DAC issues. No analog in between = no analog issues. FDA will be the future.
Last edited: