USB audio cracked... finally!

[mmerrill99]

28 minutes ago, manisandher said:

 

And yet you claim the following...

 

 

... implying that all non-NOS1 DACs do not use the 5V for sensing the connection of the USB cable.

 

Mani.

Thanks for the correction but a 'proviso' means that it's something which non-NOS1 owners should consider - it doesn't mean what you are suggesting.

 

It's easy enough to determine this whole 5V issue - can your DAC run without the 5V connection? Then it has no internal 5V electrical connection.
 

Look I know you are besotted with the LUSH & NOS1 - you want to tell the world - that's understandable & natural.

But for non-NOS1 owners, do you not think this recent information is useful?

[manisandher]

3 minutes ago, mmerrill99 said:

But for non-NOS1 owners, do you not think this recent information is useful?

 

Here's some useful information for non-NOS1 owners: the Lush has a profound affect on my iFi Nano iDSD.

 

Mani.

[mmerrill99]

15 minutes ago, jabbr said:

Let me propose an electronic explanation -- and again with the disclaimer that I do not have knowledge of the actual cable details --

 

As we know every transmitter, cable and receiver can be described as an equivalent circuit with capacitors, inductors and resistors. Cables are often drawn as straight connecting lines but we know that in reality there are capacitances, inductances and resistances at each point along the cable as well as at the source and end terminations. These can be expressed by equivalent Thevenin networks.

 

Source and/or end termination can itself have many variants including capacitors and resistors connected between D+/- but also between D to VCC as well as GND.

 

As such VCC participates in this network completely aside from any power needs that it is supplying to either the transmitter or receiver electronics. As such for specific topologies there may be a requirement that VCC be actually connected. If the resistance is measured between D+/- and VCC it might not be infinite, or if the capacitance or inductance, not 0. If VCC is disconnected the intended circuit is modified.

Right, I was not going to get into that aspect but yes, any parallel wire running alongside the D+/- USB signal wires can have a current generated in it by capacitive coupling with the signal wires & a certain amount of leakage occurring from the currents running on th esignal wires to the 6V wire.

 

This applies to the ground wire & the shield too!!

 

It doesn't mean that this wire should therefore be left in the cable - it actually means that the 5V wire is best taken out of the cable as some USB cables have done - some run the 5V wire as a separate outside the main cable.

 

Again, I can't see any advantage of working with this capacitive coupling & leakage between the USB signal wires & 5V wire?  

 

Your wider point about this as a network may have value, I'm not sure without some details - maybe that is the problem - lack of details

[mmerrill99]

13 minutes ago, manisandher said:

 

Here's some useful information for non-NOS1 owners: the Lush has a profound affect on my iFi Nano iDSD.

 

Mani.

And does the iFi nano iDSD use the 5V ?

 

Remember, I didn't start this issue about 5V - PeterSt stated it explicitly on his forum & I just asked about it

 

All I was trying to do was clarify what he said.

[manisandher]

5 minutes ago, mmerrill99 said:

And does the iFi nano iDSD use the 5V ?

 

When running off its battery, I'm not sure. But whether it does or doesn't, the fact remains that the Lush has a massively positive affect on this particular non-NOS1 DAC.

 

I would recommend anyone, with any DAC, to give the Lush a go. Not for Peter's sake, but for their own.

 

Mani.

[elcorso]

37 minutes ago, manisandher said:

 

Here's some useful information for non-NOS1 owners: the Lush has a profound affect on my iFi Nano iDSD.

 

Mani.

 

The same with exaSound e32, that with many USB cables (Curious® inclusive) did not show important changes.

 

Today I'm going to try the Lampi B7.

 

After 100 hours of continuous use is still more profound the improvement !

 

Roch

[Boleary3]

I too, with the Yaggdrasil, have found a more pleasing sound with the Lush. Schitt is advertising the new Gen 5 usb board update to explicitly not need the 5v power of the usb cable. Will let you know how that new usb board sounds with the Lush after it arrives. But that might be a while as I'm waiting for them to open sales up to those who want to install the board themselves, like me. 

[manisandher]

OK, so off the top of my head, we know the Lush has a positive affect on the following DACs:

 

- Phasure NOS1

- Aries Cerat Kassandra

- Pacific Microsonics Model Two? (not sure if @cmarin uses a USB-to-AES converter)

- Schiit Yggdrasil (pre Gen 5 board, for now)

- exaSound e32

- iFi Nano iDSD

 

I'm sure there are loads more now.

 

Mani.

[lmitche]

1 minute ago, manisandher said:

OK, so off the top of my head, we know the Lush has a positive affect on the following DACs:

 

- Phasure NOS1

- Aries Cerat Kassandra

- Pacific Microsonics Model Two? (not sure if @cmarin uses a USB-to-AES converter)

- Schiit Yggdrasil (pre Gen 5 board, for now)

- exaSound e32

- iFi Nano iDSD

 

I'm sure there are loads more now.

 

Mani.

Add the IFI microIDSD to your list.

[89reksal]

1 hour ago, mmerrill99 said:

I agree & it shouldn't be an issue for him - it wasn't a threat - it simply represented what I had to do, withdraw my recommendation to two people to buy this cable as it was stated here by Peter & others that it will be no better than existing USB cables in DACs which don't have 5V connected internally.

I've been following this thread with interest since the beginning. I have to say I believe you've misunderstood what Peter was trying to say. I believe he's simply saying that in the situations where the DAC being used does not have 5v Vbus connected internally, then any USB cable, including the Lush, will be working sub-optimally. It seems, based on what I've read in this thread so far, that the Lush, even though it may be working sub-optimally (as any other USB cable would be with the same DAC), it is still able to achieve "better" results than those other USB cables due to Peter's design. Which should mean that the Lush will still be an improvement over those other cables.

 

I'd hate to see a misunderstanding derail one of the better threads (imo) on CA. I think we're incredibly fortunate to have Peter contributing here.

[cmarin]

29 minutes ago, manisandher said:

OK, so off the top of my head, we know the Lush has a positive affect on the following DACs:

 

- Phasure NOS1

- Aries Cerat Kassandra

- Pacific Microsonics Model Two? (not sure if @cmarin uses a USB-to-AES converter)

- Schiit Yggdrasil (pre Gen 5 board, for now)

- exaSound e32

- iFi Nano iDSD

 

I'm sure there are loads more now.

 

Mani.

Hi Mani,

 

I'm using a Playback Designs USB-XIII USB-AES Converter (one wire) to feed the Pacific Microsonics Model 2 from a Sound Galleries Monaco server.

[mmerrill99]

35 minutes ago, lasker98 said:

I've been following this thread with interest since the beginning. I have to say I believe you've misunderstood what Peter was trying to say. I believe he's simply saying that in the situations where the DAC being used does not have 5v Vbus connected internally, then any USB cable, including the Lush, will be working sub-optimally. It seems, based on what I've read in this thread so far, that the Lush, even though it may be working sub-optimally (as any other USB cable would be with the same DAC), it is still able to achieve "better" results than those other USB cables due to Peter's design. Which should mean that the Lush will still be an improvement over those other cables.

 

I'd hate to see a misunderstanding derail one of the better threads (imo) on CA. I think we're incredibly fortunate to have Peter contributing here.

Yes but my experience & those of others reported in this forum & elsewhere, is that taking the 5V out of action in a USB cable (even removing the wire altogether) or running it as a separate external wire to the USB cable is beneficial to the sound.

 

So, I don't concur with the notion that no 5V wire or no 5V connection is sub-optimal for any USB cable & the evidence seems to suggest otherwise.

 

And therefore I can't relate to the idea that the LUSH will be working sub-optimally but better than other USB cables.

 

Anyway, keep the reports coming in - that's useful - the world doesn't always behave according to our models or predictions.

[jabbr]

1 hour ago, mmerrill99 said:

It doesn't mean that this wire should therefore be left in the cable - it actually means that the 5V wire is best taken out of the cable as some USB cables have done - some run the 5V wire as a separate outside the main cable.

 

Again, I can't see any advantage of working with this capacitive coupling & leakage between the USB signal wires & 5V wire?

 

That entirely depends on the cable. In an ideal world we could model the transmitter, cable and receiver network and derive ideal resistance, capacitance and inductance parameters a.k.a. termination. When the intrinsic capacitive coupling between D+/- and VCC/GND contributes to this circuit then this coupling is intended and desirable. Similarly for inductance.

[marce]

5 minutes ago, jabbr said:

 

That entirely depends on the cable. In an ideal world we could model the transmitter, cable and receiver network and derive ideal resistance, capacitance and inductance parameters a.k.a. termination. When the intrinsic capacitive coupling between D+/- and VCC/GND contributes to this circuit then this coupling is intended and desirable. Similarly for inductance.

We can and do, that's what signal integrity verification software does. The parasitic coupling is minimal for data transfer as the signal is LVDS and each signal uses the other as its return. The wires will have some coupling but not a lot. A cable is a cable and their are only limited things (very) that a cable can do to a signal, without the addition of components. I have not seen any hints or information that gives any indication of what this cable is doing (and yes I have read the whole thread .).

Crosstalk checking is also readily simulated, some EMC advisers will give some basic indication, but the best results come from the 3D field solvers. Of course then there is the USB spec...

[lmitche]

33 minutes ago, mmerrill99 said:

Yes but my experience & those of others reported in this forum & elsewhere, is that taking the 5V out of action in a USB cable (even removing the wire altogether) or running it as a separate external wire to the USB cable is beneficial to the sound.

 

So, I don't concur with the notion that no 5V wire or no 5V connection is sub-optimal for any USB cable & the evidence seems to suggest otherwise.

 

And therefore I can't relate to the idea that the LUSH will be working sub-optimally but better than other USB cables.

 

Anyway, keep the reports coming in - that's useful - the world doesn't always behave according to our models or predictions.

Many of us have experienced improved SQ by eliminating or modifying vbus power on our USB connections. Nevertheless I think it important to say that the impact has diminished as lower noise DC power sources have been available like the lps-1.

[mmerrill99]

52 minutes ago, jabbr said:

 

That entirely depends on the cable. In an ideal world we could model the transmitter, cable and receiver network and derive ideal resistance, capacitance and inductance parameters a.k.a. termination. When the intrinsic capacitive coupling between D+/- and VCC/GND contributes to this circuit then this coupling is intended and desirable. Similarly for inductance.

Yea, but I don't know of any USB cables that are designed in this way by intent - it's certainly not part of USB conformance testing. If an audiophile USB cable was done with this modelling, I'm sur eit would be used as a selling point & I haven't seen such a USP, have you?

 

I contend that it would be far easier to get the characteristic impedance stable along the length of the cable & closer to the USB spec if the 5V wire is excluded from the cable - thereby dealing with shield, ground, D+ & D- wires & the twist conformity of just 3 wires instead of 4 wires.

 

Maybe I'm wrong?

[hifial]

1 hour ago, mmerrill99 said:

Yes but my experience & those of others reported in this forum & elsewhere, is that taking the 5V out of action in a USB cable (even removing the wire altogether) or running it as a separate external wire to the USB cable is beneficial to the sound.

 

So, I don't concur with the notion that no 5V wire or no 5V connection is sub-optimal for any USB cable & the evidence seems to suggest otherwise.

 

And therefore I can't relate to the idea that the LUSH will be working sub-optimally but better than other USB cables.

 

Anyway, keep the reports coming in - that's useful - the world doesn't always behave according to our models or predictions.

 

 

I concur, and that in my experience USB cables that had no 5V wire or separated it from the body of the cable gave a better sound to my T+A DAC 8 DSD, which does not need the 5V. Now that is with a limited number of USB cables but ones that are highly rated.

But progress does not stand still and Peter may have found a better way.

 

So I plan on ordering a Lush cable. When it arrives I will compare it to those type of USB cables and report my impressions. I will also try and have some NJ Audiophile Society members try it in their systems (both with and without 5V) and report back.

 

For the record, I am not a fanboy of Phasure (though I understand why some are and think very highly of Peter's equipment) but I highly respect Peter, even when I disagree with him. And that also goes for Mani, lol.

[89reksal]

34 minutes ago, marce said:

I have not seen any hints or information that gives any indication of what this cable is doing (and yes I have read the whole thread .).

It seems peter has been saying that he designed this cable from an analogue perspective instead of digital theory. Here's some hopefully relevant quotes from Peter:

Quote

OK, a far too long post about all what is not related at all. But a bit of background.

The post could be an introduction of something which *is* important : I see the Lush compared with jitter bugs and regenarators and that kind of things - no. It is just a cable. No active devices in there. Also no resistors. A cable. But nothing like a USB cable. But USB is transferred over it. 

 

That makes it pretty clear there's no secret components built into the cable. "it is just a cable".

 

Quote

To stay on topic, not so with the Lush and I actually just talked about it;

I aimed for something explicitly with a most good idea ... bought many pieces of test cable and braid / sleeved etc. before I knew I was going to right direction and once the direction was set I already created the first "commercial" about it for the Phasure forum.

 

That quote hints that it was his experimenting with the topology (different types of physical cable and sleeving, etc.) that led in the right direction. Much like the process someone building their own analogue cables would follow.

 

Quote

Posted July 16 · Report post

Before I sit down for the (non Oled) TV for some F1 watching, this is what I wrote in another topic yesterday ... food for thought perhaps (and call me crazy if you like) :

 

What thus Lush cable does (and I described it on my own forum a couple of weeks back) is "connecting" individual wave cycles. I know, this is too far out to ever come up with, but this *is* what I attempted. You can say it smears as hell, but in a "too narrow pipe" fashion. So there is no ringing in order and no digital filter or analogue filter could do it (as they all ring). Maybe it is best to call it "shaping" (of the tops of wave cycles). And now this cable suddenly sounds analogue as LP. I am serious.

 

Before someone asks me how I'd think this may ever happen in a USB cable which only passes on digital data, even in packeted form ... no answer. But my approach has been in a fashion as if it was an analogue cable (like an interlink) *because* I could hear the very same properties for the conductor material as we can perceive from loudspeaker cable. The stupid thing is, I was already hearing that myself, and then someone came for audition and he described the very same without knowing a thing about any USB cable (test, which was active at the time). And *then* I started working with it.

 

Till later.

 That again makes it clear this cable design was based on analogue cable design considerations, nothing about "normal" USB or other digital design priorities.

 

I think there's been lots of clues offered by Peter but some want to look at the Lush from a digital design stand point, while Peter appears to have made it clear that the digital design properties were almost a non-factor. It seems he's looking at the digital data being transferred as analogue waveforms as opposed to typical digital data. He's designed the Lush to modify those "waveforms". How or why this is done is another story but I think focusing on the Lush from the digital aspect is taking us down the wrong road.

 

This is only my own understanding of what Peter has tried to share. 

[marce]

USB cable that are correctly manufactured are 90R differential, I would not worry about the 5V line or the GND line. Only the two data wires are twisted, its a twisted pair. Beldon etc. may have some info on how they design their cables as may others, what I can say is that they are designed correctly form the main wire manufacturers, its mainly down to geometry. USB 2 is pretty basic and even slight impedance mismatches are not going to make much difference, its a pretty rugged interface (it has to be as the new universal serial bus, replacing good old RS 232 and the 9-way d-type). I do wonder though wether there ever will be any digital interface that will make audiophiles happy

[mmerrill99]

 

1 hour ago, marce said:

We can and do, that's what signal integrity verification software does. The parasitic coupling is minimal for data transfer as the signal is LVDS and each signal uses the other as its return.

In a perfectly balanced LVDS, there should be no return path as the signal on each wire of the pair is the antiphase of the other - so there should be no return currents as they will cancel each other out, no?

 

In a slightly imbalanced pair when signals are not the exact opposite of each other what do you understand will happen to the return currents? Does each slight return current return in the opposite wire of the pair or does it return in the lower impedance (usually) ground wire?

 

If there were no ground wire (& this can be achieved by disconnecting USB ground after the single ended USB stage of handshaking is complete - without this ground wire then any return currents must return via the opposite wire in the pair. Does this give rise to common mode noise on the USB signal wires?

 

None of which effects digital comms but again, we are beyond that in this discussion - it's a given

[mmerrill99]

51 minutes ago, lmitche said:

Many of us have experienced improved SQ by eliminating or modifying vbus power on our USB connections. Nevertheless I think it important to say that the impact has diminished as lower noise DC power sources have been available like the lps-1.

I'm not talking about eliminating as a source of power, the usually low quality USB 5V - I'm talking about removing the 5V wire from the USB cable

[mmerrill99]

8 minutes ago, hifial said:

 

 

I concur, and that in my experience USB cables that had no 5V wire or separated it from the body of the cable gave a better sound to my T+A DAC 8 DSD, which does not need the 5V. Now that is with a limited number of USB cables but ones that are highly rated.

But progress does not stand still and Peter may have found a better way.

 

So I plan on ordering a Lush cable. When it arrives I will compare it to those type of USB cables and report my impressions. I will also try and have some NJ Audiophile Society members try it in their systems (both with and without 5V) and report back.

 

For the record, I am not a fanboy of Phasure (though I understand why some are and think very highly of Peter's equipment) but I highly respect Peter, even when I disagree with him. And that also goes for Mani, lol.

Yes, that will be a worthwhile comparison, thank you!

[marce]

3 minutes ago, lasker98 said:

It seems peter has been saying that he designed this cable from an analogue perspective instead of digital theory. Here's some hopefully relevant quotes from Peter:

That makes it pretty clear there's no secret components built into the cable. "it is just a cable".

 

 

That quote hints that it was his experimenting with the topology (different types of physical cable and sleeving, etc.) that led in the right direction. Much like the process someone building their own analogue cables would follow.

 

 That again makes it clear this cable design was based on analogue cable design considerations, nothing about "normal" USB or other digital design priorities.

 

I think there's been lots of clues offered by Peter but some want to look at the Lush from a digital design stand point, while Peter appears to have made it clear that the digital design properties were almost a non-factor. It seems he's looking at the digital data being transferred as analogue waveforms as opposed to typical digital data. He's designed the Lush to modify those "waveforms". How or why this is done is another story but I think focusing on the Lush from the digital aspect is taking us down the wrong road.

 

This is only my own understanding of what Peter has tried to share. 

Its a digital cable though, well in fact there is little difference between a digital cable and an analogue cable... So if it is just a cable then how can it do what is claimed, as I said there are only certain things that can be done with a cable, down to the basics, resistance, capacitance, inductance and conductance. Geometry plays apart for characteristic impedance, shielding etc and the dialectric will have an influence on the propagation speed of the signal. But there is very little you can do with a cable to alter the DIGITAL waveform, very little, especially when we are talking small distances a metre or less. But there has been no substance to any of Peters replies, I was hoping there would be more info but the thread seemed to get derailed over some petty little details.

[mmerrill99]

10 minutes ago, marce said:

USB cable that are correctly manufactured are 90R differential, I would not worry about the 5V line or the GND line. Only the two data wires are twisted, its a twisted pair. Beldon etc. may have some info on how they design their cables as may others, what I can say is that they are designed correctly form the main wire manufacturers, its mainly down to geometry. USB 2 is pretty basic and even slight impedance mismatches are not going to make much difference, its a pretty rugged interface (it has to be as the new universal serial bus, replacing good old RS 232 and the 9-way d-type). I do wonder though wether there ever will be any digital interface that will make audiophiles happy

As I  said before mixed signal domain analysis is required to properly analyse this situation & asked you had you done much of this sort of mixed domain work?

[marce]

7 minutes ago, mmerrill99 said:

 

In a perfectly balanced LVDS, there should be no return path as the signal on each wire of the pair is the antiphase of the other - so there should be no return currents as they will cancel each other out, no?

 

In a slightly imbalanced pair when signals are not the exact opposite of each other what do you understand will happen to the return currents? Does each slight return current return in the opposite wire of the pair or does it return in the lower impedance (usually) ground wire?

 

If there were no ground wire (& this can be achieved by disconnecting USB ground after the single ended USB stage of handshaking is complete - without this ground wire then any return currents must return via the opposite wire in the pair. Does this give rise to common mode noise on the USB signal wires?

 

None of which effects digital comms but again, we are beyond that in this discussion - it's a given

There is always a return current, for LVDS as I said they use the other wire as a return. A signal will use the nest available path for a return. Signals always follow the path of least impedance, for low frequency this is the path of least resistance for higher speed signals the path is always the path of least inductance.