Article: SOtM sNH-10G Network Switch Review

[Jud]

58 minutes ago, plissken said:

They don't transfer jitter. They can carry noise.

 

Yes, exactly.  What happens if/when the noise gets into the DAC chip and clock?  (I recall at least one ESS white paper with info on this.)  It can cause jitter.  And/or it can get into the analog side of the system through ground.  Now I'm just speaking conceptually, and not about whether the levels of this stuff would be audible.

[Jud]

Just now, The Computer Audiophile said:

I'm not following you about Ethernet not transferring jitter. Can you provide more details?

 

Timing of the bits through an Ethernet or USB cable (to an async input) shouldn't matter because they go into a buffer.  Conceptually, you could be typing ones and zeroes with totally inconsistent timing, and as long as they went into a buffer somewhere and were clocked out accurately, no worries.  There are two aspects of noise that I believe could conceptually make a difference regarding jitter: (1) The DAC chip compares signal to ground in order to determine the timing of changes from 1 to 0 or vice versa.  If there's noise on ground, the timing changes - jitter.  (2) Noise getting to sensitive clock circuitry could make the clock less accurate.

[Jud]

15 minutes ago, The Computer Audiophile said:

Doesn't the receiving end just buffer exactly what has been delivered?

 

Then there'd be little reason to have a buffer in a local file playback situation (as opposed to streaming).  At least my conceptual understanding is that the bits in the buffer are values in a location that stores those values but doesn't store arrival times.  So the only thing controlling the timing (jitter) of the bits as they come out of the buffer in a DAC with async input is the local DAC clock.

 

Edit: Ever fly Southwest Airlines?

[Jud]

Just now, plissken said:

 

Then a dipshit designed the DAC.

 

Noise is always going to be present, the only question is how much (how effectively the design protects against that noise).  Lots of methods of isolation from transmitted noise are noisy themselves, so complete isolation isn't the complete answer either.

[Jud]

1 minute ago, plissken said:

WiFi is the defacto, highest fidelity, connection available to the audiophile.

 

How noisy is the WiFi receiver circuitry?  🙂

[Jud]

1 minute ago, The Computer Audiophile said:

I like optical because one gets the reliability of wired and the isolation of wireless. 

 

Optical is a lovely idea.  But just like any other isolation method, the question is what level of noise the isolation method itself creates (in the case of optical, the receiver circuitry to convert the optical signal back to an electrical one).

[Jud]

2 minutes ago, plissken said:

 

More is better than less Jud. WiFi certainly eliminates AC hum/ground loop.

 

It does, which is great.  The point is that you're not done then.  And it could conceivably be the case that a well designed wired connection might result in less noise getting into the DAC parts than at least some optical or WiFi connections.  It's always implementation (i.e., "the devil is in the details").

[Jud]

Just now, plissken said:

I had a white paper at one point comparing the power supply rail characteristics and optical was something like 10^3 better in consumption. So that's a win.

 

I know I've read that at least some opto-isolators are electrically noisy.

[Jud]

On 2/8/2019 at 1:37 PM, incus said:

Pulling a cable just means you're listening to the buffered signal from whatever is doing the buffering - which is the whole reason you are able to hear music play without the cable attached in the first place. By definition this signal would sound the same as the signal coming over the switch and its attendant ethernet cables - because it's already passed through these things and has therefore already been affected by them prior to reaching the device (computer) where it is buffered. So no change in sound tells me nothing.

 

As @PeterSt has already said (though at some length and introducing other topics, which may have confused things), you're not picturing the experiment correctly because you're not understanding what goes on when a DAC has an asynchronous USB input, as most do these days.  The bits are collected in a buffer at the DAC input, then are clocked out by the DAC's internal clocking.  So nothing in the *timing* of the bits before they go into the buffer affects the timing of the bits out of the buffer by the DAC's clock.  That's the meaning of "asynchronous:" the timing of the bits in the DAC is not synchronized with the upstream timing, so no "jitter" (i.e., timing effects) upstream means anything to what occurs in the DAC.

 

Picture an airplane loading: No matter what happened in terms of timing through the boarding of the plane (bits moving into the buffer), everyone takes off at the same time (bits being clocked out of the buffer).  The two aren't interdependent.

 

So what might conceivably affect the timing accuracy of the clock in the DAC?  Electrical noise, in two ways: (1) Because the switch from a 0 to a 1 or vice versa depends on comparing signal to ground, noise on ground might affect timing of that change; or (2) Electrical noise might disrupt the accuracy of the clock itself.

 

What @plissken's experiment would do is remove any electrical effect of the switch by pulling the Ethernet cable connecting it, thus taking it out of the circuit.  When that happens, there are still bits in the buffer, so as the rest of that buffer plays you can compare the sound when the switch was in the circuit to the sound when the switch is out of the circuit and can't be causing any noise.

 

At least that's the general idea.  There has been some discussion on the forums about upstream clocking effects passing through into the DAC, but I don't have a sophisticated enough understanding to evaluate that, and it remains to be demonstrated that this can actually occur.

 

Edit: Actually, I just thought of a way the switch could still have an electrical effect after the Ethernet cable was unplugged: through the power cord to the wall, and from there to the rest of the system.  So perhaps unplugging the switch from the wall would be a better test?

[Jud]

52 minutes ago, Jud said:

Edit: Actually, I just thought of a way the switch could still have an electrical effect after the Ethernet cable was unplugged: through the power cord to the wall, and from there to the rest of the system.

 

By the way, this is one reason a Wi-Fi or optical unit can still allow noise to get into the system - through the power connections.  (Another way, as already discussed, is through the electrical activity of the receiver of the Wi-Fi or optical signal.)

[Jud]

1 minute ago, EdmontonCanuck said:

 

I can attest to that....

 

 

 

I remember....

 

Flew into Edmonton one evening when it was 40 F below (actually 40 below is the same temperature, F or C), and 50 mph winds.  When I (thankfully) made it to the hotel from the airport, I turned on the news to see a pie chart saying "Exposed flesh freezes in...."  The part of the pie chart that was highlighted said "30 seconds or less."

[Jud]

7 hours ago, PeterSt said:

 

Um, my signature over at Phasure tells since the beginning of Windows 10 (2015 IIRC) "Switching power supplies removed everywhere", which is true for the sub net of my mains of concern. So hehe, sure. But what is even more easily to overlook is the fact that with that you would also switch of WiFi. And if anything is measurable at the outputs of a DAC it is that.

So ... WiFi not allowed.

Did I introduce an other topic again ? I don't think so. You just don't want all this sh*t, which coincidentally is router/switch related. And over at Phasure this is not even a topic (at any times) because "we" simply avoid them. But we also created the infrastructure for it ...

 

Yes, I remember well.   

[Jud]

4 hours ago, Summit said:

 

That’s incorrect! The buffering and re-clocking doesn’t make the DAC immune to the gear you use upstream. If it would be true the phase noise in gear upstream, like the ultraRendu for example, wouldn’t matter.

 

Yes, though there we are talking about something with a different factor to consider, at least in theory.  (There are people who don’t think such devices ought to work, and it must be said no one has done an end to end measurement of an analog effect. Nevertheless, I own a microRendu and IsoRegen, and subjectively prefer to have them in the chain. One very practical reason is the computer I own that is capable of easily running the software I like is a desktop that lives in the office away from the audio system.)  The theory of operation of these devices is that the receiver electronics in the DAC should be less noisy with them in the chain.  So it isn’t a matter of timing upstream directly relating to timing downstream.

[Jud]

4 minutes ago, Summit said:

 

Why is it that every damn time someone report that they hear a difference between gears upstream of the DAC, the same old explanation is used (buffering and re-clocking in the DAC)?

 

If people really believed in that old dogma, they wouldn’t use any so called “audio grade” devices at all upstream of their DACs, it would be irrational.

 

It is well known that jitter, noise etc upstream affect the final sound and its why “all” manufacturer of upstream gear try to keep them low.       

 

I was actually agreeing, but attempting to give a more precise explanation.  Please do look at what I wrote rather than assuming I’m trying to argue.

[Jud]

7 hours ago, rickca said:

Peter it sounds like you don't believe the stuff that Alex and John Swenson describe ... upstream phase noise fingerprint affecting downstream components and getting past many kinds of isolation techniques?  Am I misinterpreting what you're saying?

 

My impression (I could be wrong) is that though this is what they believe to be happening, they have not yet had a chance to measure and see whether or not it is definitely the case.

[Jud]

7 hours ago, plissken said:

 

What I'm saying is you can't tell the difference and we can easily present different load for the OS to run.

 

Let me know when you are up for some blind testing. We can setup an Intel box with Prime 95. I can control it from remote, run it headless, and you tell me during playback when the system is under increased load.

 

If you are talking about my screenshot it was Win10Pro X64. I forget the build #.

 

Way back near the dawn of digital, people like Ed Meitner were already discussing “logic induced modulation,” and right up to the present day folks like ESS talk in white papers about lowering noise getting into the chip.  It’s a situation where the level of noise may not be quite so important as even small changes in it (changes in noise on ground = jitter at the clock), so at least theoretically you don’t want things in the playback chain to be changing noise characteristics while engaging in playback.

 

To what extent this or anything to do with jitter may be audible has always been a subject of some discussion and controversy.

[Jud]

23 minutes ago, Ralf11 said:

well, same for output resistance & Temperature variation

 

but the key is "how much"...

 

Not only.

 

 The key is how rapid the variation is, as much as or more than absolute level.  Something that varies over minutes or hours may as well take a glacial age with respect to causing jitter.

[Jud]

25 minutes ago, incus said:

Thank you.

 

But notice: The measurement hasn’t happened yet.  And even after that, we’re left to ponder the question of how significant/audible it might be.

[Jud]

22 minutes ago, plissken said:

My Emotiva DC-1 seems impervious to the systems I have had it connected to when it comes to testing the upstream noise possibility.

 

I am sure this is true, but I was so sure it was coming that I had to smile, my friend.