[johnjen]

CPUs operate in GHz region. You can hear GHz?

Obviously not.
But there are beat frequencies and their interactive harmonics that can be, and when absent, or minimized these differences will become apparent.

Not to mention that the PSU will 'react' to the varying demands for power and there are other subsystems besides the CPU that will also interact and modulate ground as well.

And our systems are designed to be sensitive to small analog signals and if these are introduced into the heart of the active circuits, directly, via ground, then 'noise' is a result.

And as I mentioned EVERY chip, and circuit does 'stand' on that very same ground, not only as part of the signal return path but also it IS the foundation or reference that the circuits 'define' as zero. So when zero is modulated so is every active portion of the circuit.

JJ

 

[tincanear]

That's not quite how noise works... any conductor or resistor creates noise. And there are plenty of those in any electronic system.

there are multiple noise sources:

resistors have thermal (Johnson) noise, but it has a wideband characteristic (like white noise).

in non-linear (real world) systems, two frequencies f1 and f2, for instance, will be modulated to create two more frequencies, namely f1+f2 and f1-f2, so its possible that GHz
CPU's and their associated circuits can produce audible noise.

some switching power supplies still operate in the tens to hundreds of kHz range.

the "12V DC" cooling fans in PC's actually chop the DC into multiple phases with frequencies in the audible range.

 

[Jvincent]

there are multiple noise sources:

resistors have thermal (Johnson) noise, but it has a wideband characteristic (like white noise).

in non-linear (real world) systems, two frequencies f1 and f2, for instance, will be modulated to create two more frequencies, namely f1+f2 and f1-f2, so its possible that GHz
CPU's and their associated circuits can produce audible noise.

some switching power supplies still operate in the tens to hundreds of kHz range.

the "12V DC" cooling fans in PC's actually chop the DC into multiple phases with frequencies in the audible range.

I was only trying to convey one very basic principle of noise. Thank you for making a more detailed explanation as well!

 

[471724]

Just a couple of thoughts to consider.
#1 the tablet is still a digital device and it is it's own ground, which also means the operational noise of the computer operating IS on its ground in one form or another. It's true it doesn't have ac to dc power supply noise, but in my tests the digital operational noise just swamps the power supply generated noise.

#2 These USB boards starting with the 5 gen boards use isolation input/output transformers to electrically isolate the incoming and outgoing USB signals from the dac.
So it doesn't matter which device is the source of the USB power because they both are.
Because they are not connected to each other, electrically.
And so from a certain perspective each 'side' of the USB cable uses it's own power. The Schitt dac powers itself and the tablet powers itself and they share data but not ground nor volts.
This is the way that ethernet network cables operate as well, which is what they did, they used ethernet network isolation transformers and isolated the internals of the USB card from the external source voltages.

This isolation was introduced with the gen 5 cards and it was and remains very effective.

Just my 2¢ worth

JJ

I was just going by the Schiit Yggdrasil Unison USB upgrade website page which contains this capsule description of one of the advanced features of the upgrade: "Self-power by the DAC for the critical low-noise re-clocking and latching sections." Supposedly this wouldn't be described as a feature unless not using USB cable-supplied power for this critical function is better than using cable-supplied power. Perhaps this design feature was incorporated for the more common case where the digital source is a PC or other AC powered unit.

 

[alpovs]

That's not quite how noise works... any conductor or resistor creates noise. And there are plenty of those in any electronic system.

I am talking about our ears. Who cares if the noise is in MHz or Ghz as long as we don't hear it. Cosmic radiation, neutrinos, all kinds of waves are passing through you right now. Do you care?

 

[Jvincent]

I am talking about our ears. Who cares if the noise is in MHz or Ghz as long as we don't hear it. Cosmic radiation, neutrinos, all kinds of waves are passing through you right now. Do you care?

If I couldn't hear it then I wouldn't care. But depending on the rest of your chain (if youre buying a $2k+ DAC then im guessing its pretty decent) you can definitely hear noise. It's very apparent.

 

[earnmyturns]

I am talking about the same. What are the frequencies? What electronics operate below 20 kHz in 2020?

Aliasing through nonlinearities folds higher-frequency signals into the audible range. Digital electronics are highly nonlinear.

 

[johnjen]

Another perspective to consider is…
Square waves are 'harmonically rich', indeed one way of 'describing' a square wave is to use a series of sine waves over a broad bandwidth.
This very technique is used in a variety of ways.

IOW square waves (the very basis of the digital signal) can all to easily react with each other to create beat frequency signals as a result, and digital circuits themselves are really just biggley modified, very fast acting analog circuits to begin with.

And analog circuits can react to beat frequencies all to easily…

JJ

 

[SilverEars]

Another perspective to consider is…
Square waves are 'harmonically rich', indeed one way of 'creating' a square wave is to use a series of sine waves over a broad bandwidth.
This very technique is used in dac chips themselves.

That's not how square waves are formed. I think you are mixing Fourier series of how square waves are represented vs how they are actually formed with transistors, which are pulses. We can refer to transistor switching frequencies.

Anyway, all this talk is just talk. Any of you design stuff for a living and deal with circuit noises? If not, please keep within what you know.

 

[jp11801]

Just a quick note on the Yggy GS if anyone is on the fence. Disclaimer - I have owned the A2 with Gen 5 but that was 6 months ago so I am not comparing the GS to that as it would be impossible for me to accurately recall and my digital source has changed.

First I am running Roon Rock from a NUC 8i3 powered by an Sbooster linear power supply. Build is complete but is going to be re-cased in an Akasa Turing.

Gen 5 usb from the NUC somewhat sucked in that the upper mids were slightly harsh and bass was good not great. I have a Matrix X-SPDIF 2 that I just for kicks plugged in and used the AES out to the Yggy GS - BAM effortless sound smooth yet detailed great depth and width for the stage just still upfront presentation. The unison would have to be a significant upgrade to the Gen 5 to supplant the current sound I am getting.

If you are on the fence on a GS my take is they sound great and you get the rom upgrade on the sound. As long as you have a great USB convertor you are in business.

 

[dieslemat]

Just a quick note on the Yggy GS if anyone is on the fence. Disclaimer - I have owned the A2 with Gen 5 but that was 6 months ago so I am not comparing the GS to that as it would be impossible for me to accurately recall and my digital source has changed.

First I am running Roon Rock from a NUC 8i3 powered by an Sbooster linear power supply. Build is complete but is going to be re-cased in an Akasa Turing.

Gen 5 usb from the NUC somewhat sucked in that the upper mids were slightly harsh and bass was good not great. I have a Matrix X-SPDIF 2 that I just for kicks plugged in and used the AES out to the Yggy GS - BAM effortless sound smooth yet detailed great depth and width for the stage just still upfront presentation. The unison would have to be a significant upgrade to the Gen 5 to supplant the current sound I am getting.

If you are on the fence on a GS my take is they sound great and you get the rom upgrade on the sound. As long as you have a great USB convertor you are in business.

Interesting. I too have the Matrix X-SPDIF 2 that I use for my Yggdrasil A2 in my living room.I have another Yggdrasil A2 that I have been contemplating on upgrading to unison. I think with a lot of comparison AES EBU still is better so I may be leaning towards the Raspberry piaes solution. Rather than upgrading. But will see what happens

 

[johnjen]

That's not how square waves are formed. I think you are mixing Fourier series of how square waves are represented vs how they are actually formed with transistors, which are pulses. We can refer to transistor switching frequencies.

Anyway, all this talk is just talk. Any of you design stuff for a living and deal with circuit noises? If not, please keep within what you know.

Yeah I should have taken more care in my choice of words.
I'll go back and edit that post to more accurately state what I wanted to say

Thanks for the correction.

JJ

 

[Alexnova]

This is easily the best USB implementation I've ever heard. More clarity, detail, less grayness and hardness from Gen 5. Less one note bass too. Great job Schiit.

 

[acguitar84]

I wonder if the new USB implementation would be better than using the lynx soundcard/aes that I'm using now, or if it would be similar?

 

[earnmyturns]

I wonder if the new USB implementation would be better than using the lynx soundcard/aes that I'm using now, or if it would be similar?

My experience:

Pi 2 Design Pi2AES AES just noticeably better than Unison USB from Allo USBridge+LPS on dynamics, detail.
Unison USB+Allo USBridge+LPS noticeably better than Auralic Aries Femto AES in every dimension.

TL;DR Unison USB is up there with AES, but won't beat every AES source. USB source may also matter in the comparison.