Bits is bits?

[Miska]

High frequency DC is why I use ferrite beads on my DC cables...  

 

In fact the medical grade ~25 EUR Mean Well PSUs I use (5V DC, 6A) have both shielded cable and ferrite bead. Unlike the cheaper 10 EUR options...

[Miska]

2 hours ago, DonaldT2109 said:

This is where you are going completely wrong. This is where a lot of people have their huge, huge misunderstanding.

The analogue signal is not carrying any audio. The analogue signal is transporting the bitstream of digitally encoded audio

You cannot possibly affect the quality of the source audio by reducing distortion on the analogue signal. Can you really not understand this ?

 

There is some amount of misunderstanding here. We are talking about two separate things.

 

1) Correctness of data transfer

2) Noise injection to analog parts through galvanic connection of the source and DAC

 

(2) completely and independently bypasses the data path. It doesn't require any data to be transmitted and unless excessive, doesn't affect data transfer itself. But it tends to affect clocks, D/A conversion part and other analog parts of the DAC - this is where the crossing from digital to analog happens and this is where the electronics are sensitive to noises at levels down to -140 dB or so.

 

I've found audiophile USB cables to be usually non-spec compliant and cause more trouble than good. I personally use cables that are officially USB HiSpeed (USB2) or SuperSpeed (USB3) certified. But depending on DAC and many other factors, analog implementation of a digital USB source matters... This is one reason why I prefer Ethernet instead of USB, either in copper (transformer isolated) or optical form.

 

[Miska]

1 hour ago, DonaldT2109 said:

How does a USB cable change the sound of the audio that has been digitised like some people believe is possible.

 

Well, with my Chord Mojo I had so many challenges finding USB cable that doesn't eventually end up with DAC blasting full volume white noise on my ears...  If you want to test USB cables and sources, that's a perfect product for the purpose. I guess that's why also HiFi-News is measuring USB source differences with it...

 

But various USB powered DACs are good candidates too. Like Meridian Explorer (2), and probably Audioquest DragonFly as well.

 

If there are differences, they are likely more due to point (2) in my earlier post, apart from the cases where there's a problem with (1), however symptoms are quite different.

 

 

My analog cables are primarily Supra, because the price is decent and I agree with their design philosophy. Plus they also have specs.

 

On digital cable front, I've been just trying to tell people NOT to use those damn shielded STP Ethernet cables in order to avoid totally screwing up the galvanic isolation Ethernet otherwise provides...

 

[Miska]

42 minutes ago, DonaldT2109 said:

The reason not to use anything but cheap generic ethernet cable is very simple to explain

 

it is impossible for an ethernet cable or switch to change the audio. One of the advantages of using Ethernet is that the digitised audio is enclosed with  protocols which utilise crc ( cyclic redundancy check ). That means if you flip just one bit, the packet will be rejected for retransmission. This is how you manage to see a perfect page in your browser after the digitised data has traveled across countless ethernet cables and converters to get to you. The bits MUST arrive as sent. Take a look inside a Google or Microsoft server farm. Generic ethernet cables connecting everything.   No Ethernet cable can make a difference

 

You are again forgetting the important electrical path that works outside of the data transmission itself. Of course having been working on networking protocols and hardware design for ~30 years I have some kind of vague understanding how these things work.

 

Of course data transmission is error-free. But this is not the issue. It hugely matters if you use UTP or STP cable. Because STP cable connects grounds of two devices, potentially allowing dirty ground currents to flow, while UTP cable doesn't and the signals are galvanically isolated using transformers (if you don't use PoE). Transformer isolation exists there for a reason! STP cable can be only used if you have a very carefully designed grounding layout on your entire system (and not for long runs for various reasons).

 

44 minutes ago, DonaldT2109 said:

I have forgotten Miska,  Are you in the 'USB cables can make a difference' camp or the ' thats impossible' camp ?

 

I'm not in any black-and-white camp, I'm in the grey "devil is in the details" camp.

 

[Miska]

2 hours ago, DonaldT2109 said:

Can you not understand that the cables do not matter.   The protocols involved give perfect delivery

 

Can you not understand that cables can have perfect data delivery, and at the same time deliver electrical noise that bypasses the data path. You know, this is also why many cables have ferrite beads...

[Miska]

12 minutes ago, pkane2001 said:

I have a number of XMOS and Amanero USB receiver-based devices. None of them suffer from USB noise, except for the old Emotiva XDA-2, which appears to be using a Cmedia CM6631 receiver.

 

I have too, and I have many that suffer from USB noise. Although for many that is baked-in feature of the USB interface instead of delivery from the cable...

 

For example is is quite common to see 8 kHz USB packet ticking (125 µs packet interval of UAC2) in the DAC output...

[Miska]

6 minutes ago, DonaldT2109 said:

Once more   Did the CRC check ?   No ?    Send me the data again

 

By the way, UAC2 doesn't do resends on errors...  (and many DACs don't even check the correctness)

 

6 minutes ago, DonaldT2109 said:

Can you not understand that using the protocols that are used what is sent is received

 

Yes, but what we are discussing here is completely outside of scope of protocols or digital transmission.

 

Just like if you use powerline Ethernet. You have electricity on the wire, and you have data and the high level protocols on the wire. Electricity is delivered completely independently on the same wire regardless of your data transmission. And your bits don't have a faintest idea if the electricity is 110V 60Hz or 230V 50Hz, or that the electricity exists there at all.

 

[Miska]

2 minutes ago, pkane2001 said:

Right. That’s why it’s important to check measurements before buying a DAC. Proper receiver implementation will make specialty USB cables and regen devices unnecessary.

 

You cannot fix that part with such devices either.

 

But Chord Mojo is one good example of a DAC that doesn't have that particular problem, but is extremely sensitive to noises coming from the USB cable...

 

[Miska]

6 hours ago, asdf1000 said:

" isolation of the ground plane using ferrite beads to form a definite moat, this avoids any capacitive coupling between the input and the output. "

 

 

@Miska this is exactly what Hugo2 USB data input has and  Mojo's USB data input doesn't have - ground ferrite bead isolation.

 

Mojo came first. Safe to assume Mojo2 will have it.

 

That is not enough to fix Mojo's sensitivity to USB sources and cables. I've tried it with a USB cable that has ferrite bead.

 

It is not isolation, it is just a filter for high frequency noise. In isolated case you would have galvanic isolation (for example optical) between USB controller and rest of the DAC.

 

[Miska]

26 minutes ago, asdf1000 said:

But is a USB cable with ferrite bead the same  as the ground plane ferrite bead to filter RF, that @marce mentions in that link?

 

Practically yes...

 

26 minutes ago, asdf1000 said:

Hugo2 doesn't have galvanic isolation on it's USB input and doesn't have anywhere near Mojo's sensitivity.

 

I'm not going to buy it to find out, Mojo was already such a massive disappointment... I tried to find some use for it, but got sick and tired of eventually getting blasted by white noise every now and then, even when using something like iPhone as USB source and only playing simple 44.1k stream from Tidal.

 

26 minutes ago, asdf1000 said:

isolation of the ground plane using ferrite beads to form a definite moat, this avoids any capacitive coupling between the input and the output.

 

That is not isolation by any stretch in my opinion. If ground plane has any connection, it is not isolated.

 

[Miska]

35 minutes ago, asdf1000 said:

 I have a Mojo and use it’s Toslink input.... zero incoming RF issues. And jitter is a non-issue...

 

And bypasses the flaky USB... But useless for me because it limits available rates and sources...