audiophile ethernet and USB cables

Let me see if I can throw a wrench into all this, and perhaps get folks to worry even more...

So far we've discussed Ethernet, USB and streaming from the view that there is a magic port coming from the wall which delivers bits sprinkled with magic dust.  The real path of the music to each home is different:

Spotify HDD -> 1 meelion network devices and cables including coax -> your home -> your network gear -> streamer / player -> amp -> speakers

If there is such degradation by different cables and such, it's amazing we can even hear anything except white noise once the bits have passed through so much before getting to us.  If we are to believe in such difference then just moving to your neighbors house, or to a new town or even country would yield different results.

Maybe I'm crazy, but putting a super expensive cable on the tail end of all that is whistling in the dark. 

It was just a very open, lively sounding cable which i found suited my ears. This is in the system context as i can't bring myself to describe a digital cable having a sound

James

Thanks James, your description is similar to my experience with the Chord ethernet cable i tried over my stock cat 6 i use. It suited my NDX but not the Auralic i was demoing at the time.

Graeme 

Isnt this the same as mains cables though?

:-D Oooo, now you've done it.  This is where the difference between USB and ethernet really comes in for me. In a TCP/IP situation where packets can be buffered, validated, and indeed resent if garbage, and where the audio data (and yes, it is merely data at that point) has come all the way from France (to the UK) in my case, really shows up the why one metre of audiophile cable on the end of that particular part of the process should make no difference at all.  However, the situation is different once it is converted to a time critical audio stream supplied via USB, or SPDIF to DAC; the resend option disappears for one thing.

To Graeme's point about mains, yes, it is the same in a sense, but in my mind a least, the USB theory, is just slightly closer to the mains theory.  Yes, the mains supply will have traveled many miles and through very non-audiophile means, but it is what it is.  It's not going to be buffered, validated, and decoded (except if using a Hugo, or other battery powered device), so the best you can do is to preserve it as best you can.  And, however crazy it might seem, I find more audible benefit in this area than most others, Hi-Res, WAV vs FLAC, USB cables, WASAPI / ASIO vs DSOUND.  On a par with Spotify vs Qobuz I would say in terms of how much difference i appreciate from the changes I've made.  

PS no I do not have a dedicated mains spur....yet. :-)

I think most people with 'audiophile' Ethernet cables are hearing differences with red book or Hi-Def files streamed from local media servers in their home, rather than with compressed internet streaming services.

Am I missing something? Why 2 USB plugs? 

Any views on the single USB version? I still not sure what they are trying to do but it looks good. 

We should all have dedicated mains spur, and a really thick optical cable directly patched to the servers of our favorite streaming services.  Then, and ONLY then will we truly know what its supposed to sound like. 

SomgStream, to be pedantic, Asynchronous audio USB is rate critical rather than time critical.

With Aync Audio USB every thousandth of a second a frame of samples are sent. The receiver then tells the sender it increase or decrease the number of samples in the frame so as to keep the receive buffer optimal.

SPDIF has no feedback at all, so the send and receive clocks need to be very accurately aligned if occasional gaps are to be avoided whete the receive buffer exhausts or over flows. In the early dayes the receiver clock was derived directly from the SPDIF transport clock to avoid this, but then transport jitter had far more impact than it does today.

TCP/IP when using TCP effectively can control the data rate between sender and receiver and also provide transport integrity and resend of the data if required. Here  the important clock is at the receiver end, and is the clock that spools the collated sample data from the application buffer that sits above the TCP/IP stack.

I think I am correct at the physical analogue level, all three use Manchester encoded voltages. All three effectively use analogue transmission lines and therefore the cable properties become part of the transmission and receive circuitry. Whether it's RFI or other coupling, in my opinion it really would be an incredibly special design that completely decoupled the physical analogue transmission line and connector properties from the rest of the directly connected system.

Simon

Qute 2 has standard type A to B USB cable, so would the Mercury be worth considering do you think?

This thread got me thinking about my own setup - i can hear differences with different audio players on my Mac and also different USB cables so my DAC is sensitive to its source quality. I've ordered an Uptone Audio USB Regen - a lot like the iFi device but with the ability to regenerate the USB data. If it does what it says on the tin, then it should be an interesting addition to my system. 

I would see whether you can try before you buy - OK, not overly expensive but I find that the sizing and weight of the converter mean that it hangs in the Hugo socket and breaks connection at the slightest hint of movement. If in a fixed position probably not a problem, but not much cop if portable is your goal. The same applies to the Audioquest Cinammon micro USB cable, the attachment point just doesn't fit. A cheap as chips converter, ca. €1 from my local electronics store works much better as it is less bulky.

I have worked in a team designing hybrid mode (combined digital and high precision analogue) electronics, and I can attest to all that Simon has said.

These are precisely the problems we encountered, and had to deal with (they cannot be completely solved, only reduced in significance by good design).  I can claim a degree of expertise in this as Hamamatsu specifically sent a team of their engineers from Japan to see how we had achieved such good results using their components.  There was nothing radical in our designs, just careful application of established good engineering practice.

So yes, in hybrid mode systems, digital signals can and do interfere with delicate analogue signals.  This cannot be avoided, just ameliorated.  Careful selection of cables (analogue and digital) is just a part of the overall amelioration strategy.

I do wonder how many of these USB 'filters' use something across the USB 2.0 balanced data pair like

http://www.st.com/web/en/resou...sheet/CD00282307.pdf

which no doubt cost a few tens of pence and providing noise filtering and DC galvanic isolation.

If I was designing a USB front end for a DAC I would be looking at something similar to this to decouple noise to some extent from the USB source - and all fits within the size of my little finger nail.

Interesting, if built in as part of the design an L-C filter can be quite effective.

As the cable used isn't under the designers' control, resistors are probably needed between the inductors and the socket to reduce the effect of the attached cable (however as there's only limited variation for 'normal' cables this should be possible to incorporate into the design).

However if this is put into the output end of a USB cable or a 'filter' at that location, then the potential for resonance effects of coupling between this and the input circuitry of the component could be significant and quite unpredictable.  This could cause considerable variation in the effectiveness of different cables with different audio components.

Looks like Audioquest have been thinking along the same lines. 

On the Audioquest Coffe USB cable there is a module, but not sure what it does, something to do with balanced pairs

Oh deary dear - what a lot of misinformation and half truths [in my opinion]. Wat where did you find it?

The USB 1.X specification defines a max cable length of 3m. USB 2.0 defines a max cable length of 5m. USB3.X does not explicitly say - but 3 m is recommended. At 12 m you are well out of spec - your cable may work with your equipment - It may not.

Asynchronous USB Audio in most modern USB DACs uses isochronous USB data transfer. There is no data resend or asking to resend as part of the USB protocol. If there is an error, the data is discarded by the USB app stack. All you can do in Asynchronous USB Audio is increase or decreases the number of samples in each 1mS frame.

Ferrite beads are usually a key part of USB filtering internally. Ferrite Chokes are sometimes used to ensure the cable meets EMC compliance - dependent on design. The chokes applied around the cable must be placed close to the ends to be effective. The chokes impede common mode interference. This does not affect the balanced D- / D+ pair - which carries the payload. However it may slew the frame signalling which is carried in a non balanced mode across D- / D+  (poor design IMO) between the data frames - but the impact will most likely be minimal and unlikely to affect operation integrity unless perhaps you are working outside specification.. like cables of a non conformant length.

The USB specifications require D- / D+ to have a transmission line  characteristic impedance of 90 ohms. You can only measure this with specialist equipment. Why state this here? and then say length doesn't need t comply - confusing.

Gas filled conductors, damping values, gold plated contacts, double shielding for 'low indctivity' is not part of the spec and I suspect is mumbo jumbo and pseudo science

Here is a useful guide by Intel of designing devices and cables for low EMI - not much gold or gas filled conductors in sight

http://www.ti.com/sc/docs/apps...face/usb/emitest.pdf

Simon

Thanks, oh well it must be right then 

Simon, I had to read this about five times, but I think I am with you up to paragraph four.  Paragraphs 1 to 3 above confirm in technical and very precise terminology, what I understood, albeit in layman terms....I think.  Paragraph 4, I'm now lost.  I am very keen to understand though, and really would appreciate your time to explain further, if you can be bothered.  Glove puppets may not be required, but a digital transport explanation for dummies perhaps.  

SongStream in paragraph four I am attempting to say the USB, SPDIF or ethernet line are essentially analogue connections. They use what is known as transmission line theory - that is the connection allows high frequency analogue voltages to be sent along wires with minimal loss or disturbance from a source to a receiver. However  in the real world any deviation from the ideal transmission line will interact with the sender or receiver. This interaction can be electrical loading related, modulated ground currents, reflections, EMI etc. The cable of the transmission line (USB cable, SPDIF cable, Ethernet cable etc) is likely to affect the properties of these interactions. 

These interactions can -and appear to - affect the directly connected audio equipment. This is no doubt because of shared/coupled resources such as powerlines, ground planes, local EMI etc. Our hearing is very sensitive to changes in audio - and this is perhaps why we hear these effects.

The point about Manchester encodings underpins the fact the analogue voltages sent over the lines only represent digital binary values when processed and decoded. The digital data is sent as an analogue signal.

Hopefully that makes some sense?

Pww, did I mention glove puppets?  Thanks, Simon.  I need to read this many more times and make extensive use of Google (dangerous, I know) to get my head around it.  Give me 24 hours.  Thanks :-) 

I just executed a network cable upgrade. I moved from high quality standard connections to an audio quest Vodka cable. I knew what it would bring due to a A-B comparison a couple of weeks ago. But was not prepared to the level of change. Detail, soundstage and tonality have further improved with a margin. Funny thing is that this has been the cheapest upgrade in my system ever. 

I will soon get another cheap upgrade by exchange the standard power cable of my NAS with a special power supply. This will than become my cheapest upgrade.

A question for somebody like Simon. Do this cables also display burn in development or is this irrelevant giving the digital character of the transfer ?

Hi Bert which NAS do you have, and whats the special PSU?

good to hear the ethernet cable change has been positive. Are you using just the one, is it from switch to streamer?

Graeme