Jump to content
IGNORED

Do I need clean USB power when using DAC with its own power supply?


Recommended Posts

42 minutes ago, botrytis said:

There is a thread in the objective forum her that shows a well designed DAC should not allow noise from a PC go through to the analog output.

 

Those measurements were made in the normal mode. But common-mode is where the noise over USB shows up.

Link to comment
2 hours ago, Archimago said:

I haven't looked at the objective forum thread but all these years of testing and reading comments by people that say a linear power supply makes a difference (or silly USB filters like the JitterBug), I'm left scratching my head.

 

There's  no need for head scratching, its well established in the EE world that switching supplies are common-mode noise aggressors.

 

https://www.eetimes.com/power-tip-47-tame-conducted-common-mode-emissions-in-isolated-switchers-part-1/#

Link to comment
10 minutes ago, Archimago said:

 

How do you know that these are issues affecting modern USB DACs?

 

 

Since its not an issue of the DAC, rather the power supply, it affects not just USB DACs and not just 'modern' DACs. S/PDIF fed DACs can have common-mode noise susceptibility too.

 

12 minutes ago, Archimago said:

Engineers who make DACs these days are also aware of potential noise so will address them in the designs whether common mode or otherwise. So unless we can demonstrate an issue with the DAC output using the stock power supply that came with the device, no need being concerned, right?

 

 

Do you have evidence of engineers awareness of potential noise issues? I think Schiit engineers may be aware (that's just a guess though) but are engineers in general? No, it isn't just about 'the DAC output' because typically that gets measured normal-mode.

Link to comment
18 minutes ago, Archimago said:

 

Sure, but the audio signal is characterized by the electrical output from the DAC to your amplifier. What do you think is wrong with the output signal to suggest power supply related issues are present (whether it's USB or S/PDIF)? Do you have a specific example of this issue?

 

That is the usual way of defining the audio signal I agree. But amps interconnected with DACs via single-ended cabling don't have very good common-mode rejection (CMR) in many cases. Nothing need be wrong with the output signal when common-mode noise resides on both conductors (which is the very definition of common-mode) yet the amp typically gets influenced by it and it doesn't benefit the SQ. An AP won't see CM noise because its designed to have very good CMR. No I don't have an example to share right now, but I'll keep my eyes peeled. 

 

26 minutes ago, Archimago said:

Well, over the years I have spoken to a few DAC "audiophile industry" designers. People you would know if I named them but that's not really relevant here since I don't think they generally want to participate on consumer forums ;-). And I have seen some of the internal testing documents (that cannot be released of course). Let's just say that noise level testing is a very important parameter for everyone and it's done on multiple levels from the power supply output noise to of course balanced and unbalanced outputs.

 

If there is a noise issue that can be addressed, I cannot imagine that they would not fix it. (To be clear, the DACs these guys I've spoken to make are not cheap!)

 

So when I hear that aftermarket companies are making power supplies or noise filters that somehow improve performance, I'd certainly like to see what the deficiencies are and then how the products have fixed the issues.

 

I recall from reading his posts here and on AudioCircle that Charlie Hansen (Ayre) was aware of the CM noise issue. Which is presumably why he went to the lengths he did to introduce isolation between source and DAC. But he said even that didn't render his DACs completely immune to source noise.

 

'Nose level testing' does not sound to me like the same thing as CMR. I'd be very interested if you had evidence of companies doing measurements of CMR on their single-ended inputs. Its quite common on balanced inputs but then the test signal is only applied to the signal conductors (pins 2 and 3 on an XLR) and not between pin1 and mains earth. Not every manufacturer I've opened boxes of is aware of 'the pin1 problem' but I do admit awareness of it is gradually seeping out. Bruno's doing a sterling job in that regard I think.

 

As regards the aftermarket offerings - you and me both.

Link to comment
4 hours ago, audiobomber said:

If I were looking for a measurement to correlate with PRaT, the first place I would look is transient performance. How closely does the DUT follow transient swings, i.e. attack sustain decay release? 

 

If I were looking for a measurement to correlate with PRaT I'd be looking in the realm of noise and noise modulation. I have a Xindak amp which I modified the power supplies of and it greatly improved the PRaT - I figured noise on the supply lines (which may well have been signal-modulated) wasn't being rejected sufficiently by the opamp front-end of this amp in stock form.

Link to comment
1 minute ago, March Audio said:

I don't see what the disagreement is, it is just stating that the audibility of jitter is dependant upon its nature (frequency etc).  All very well known and understood.

 

Audibility of jitter is also dependent on DAC architecture (noise-shaping etc.) but I don't see that stated in many places.

Link to comment
7 minutes ago, March Audio said:

No, I was differentiating between audibility and the fundamental cause for the jitter.  The fundamental cause is not relevant to audibility, the nature and level of the jitter is.  You were conflating the two.

 

Your last sentence is nonsense.

 

Seeing as your answer to my earlier question is 'no' then with those studies into jitter audibility you've been waving your hand towards, did any of the authors mention the architecture of the DAC they were listening to? And if so, what kinds have been used in those audibility tests?

Link to comment
Just now, audiobomber said:

Or maybe you improved current delivery, which is a big determinant of PRaT, IME,

 

If by 'improved current delivery' you mean I lowered the PSU impedance then yes, I did. In another sense I restricted current delivery because I changed over from series regulation to shunt regulation. Shunt regulators tend to have a lower maximum current capability than series.

 

There does tend to be rather a mis-match in vocabulary between those with EE/tech backgrounds and those without. I am curious about finding ways to link up the two domains.

Link to comment
6 minutes ago, audiobomber said:

What I mean by current delivery is that the amp needs to provide the required amount of current to satisfy the load, in the proper amount, and on time. If it fails to do so, the result is current clipping, which can occur on transients even when current demand is well below maximum rated power. IMO, current delivery or current sourcing, is the reason high-end amps sound better than mid-fi amps.

 

 

I'd agree, the basics haven't changed. Here 'current delivery' only applies to the power stage of an amp where the load impedance (the speaker) isn't precisely known. In the case of my particular mods I didn't change anything in the power stage, rather I modified the power supplies to the signal stage. In signal stages the load impedance is known and opamps don't go into current limit ('current clipping') unless the designer's made a grave error.

 

What I suspect was happening in the stock amp was noise from the power stage supplies was 'bleeding through' to the signal stage supply, due to the regulators not being up to the job (insufficient line rejection). They also had too high an output impedance due to them feeding just a pair of opamps (about 8mA) - regulator output impedance is quite a strong function of output current and 8mA from a TO-220 reg is almost a negligible amount where the max rating is 1.5A.

Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...