Phasure NOS1 24/768 async USB DAC

[PeterSt]

Hi Joachim,

My reservations about USB appeared justified allright because this really won't go by itself. The "problems" themselves didn't change, but my approach did. It is fair to summarize it with two main aspects :

1. Noise which is like (relatively) crazy;
2. Jitter which is higher.

Both are related, because at these speeds (480mbs) no means exist to surpress noise other than incurring for additional jitter. Notice that this is not about audible noise as such (around 2mV), but it influences the sonics heavily (well, listening to the NOS1).
And so I sat down to achieve the 7-8uV from the original NOS1, which again just has been a reference, or IOW "see, that is possible !".
This whole little "project" is about prooving where actually "noise" comes from, and it is here where Nick has been a great help. Not only about looking for even better sonic qualities, but merely about the proof what did NOT help, hence what it all was NOT about. Funnily enough we both worked with the same sources, but onto different applications : Nick with batteries on the Juli@ department, and me with batteries on the USB stuff. We started with this at the same time, but I didn't want to apply it to the Juli@ interface anymore because it seemed the wrong traject. In the mean time though, Nick's sound improved while mine only got worse. It didn't work out, or at least we couldn't find the consistency in it all. I couldn't and Nick couldn't either, both working on a complete diffrerent interface. But here too only one thing seemed important : eliminating all the regulators to be found. Nick's 1, 2, 3, 4 and 5. Maybe 6.
I worked out a complete battery operated design (so, 100% of the DAC including chargers of different types for the different types of batteries, the connections), but noticing the inconveniency of charging (though the whole thing could kept on playing for 48 hours continuously) made me work in the other direction more : achieve noise levels with normal PSUs.

A real strange coorporation which merely looked like sparring to me, but -and I can't say this enough- which for me was about "seeing" the reference, thus knowing what must be achieveable.
And so the USB interface ended up with 7uV (6uV maybe, depending on the soldering quality), even lower than with the Juli@ interface. Notice though that this isn't about noise at the outputs only, but more importantly about jitter influence which doesn't show as noise at the outputs at all. However, the theories were applied throughout, and while I can't let show jitter of any kind anyway, the sonics told me sufficiently it helped, as it did with Nick's. Measurable or not, the jitter must be way lower now because of the crazy difference in sound. Fun is and remains : no difference with Nick, although working on the complete different interface - and keep in mind, the exact same DAC.

Will you recognize your music ? I guess so. But somehow you won't recognize your listening room ? What you will not recognize is the basses. Not only the upright ones (they are just live now) but also the stiffness of electric ones. Oh, that already was so good ? well, it could be better. This too is just jitter again, no matter how low it is or was.

USB3 ? I wouldn't know why. There's 8x 32/384 (24/384 output) and it really is enough. It would only allow 8x 32/768 as an addition, but more has to change because the on-board processing will have ran out of bandwidth then. Also, keeping in mind this is all meant to be an 8ch XOver, 768 will be too much for everything. Well, that's what I estimate.

September/October ? yes, there should be sufficient stock then. And if not I'll have reordered.

Peter

[Flecko]

Hi Peter,
congratulations to your newes achievement. Didn't expected the thing you are working on was so big
 

[Nick]

Nick, consider your efforts to be beneficial to all, hence the new NOS1. The latter above is just part of it. I think we have been driving eachother to the sort of ultimate. At least I have been tempted to "see" the least noise possible, which is -as often- only the reference of what's achieveable. Next it "only takes" to push yourself to achieve that reference in whatever other situation. This is what happened, and it is exactly this why USB started to work in the first place. Without this ? no way it would have been any good.

Peter,

Thanks for your kind words,

The fun for me has been that as I was trying to sweat a little extra performance from the Juli@ interface, you were working on similar design decisions on  your entirely new interface. So we have had some fun conversations about clock power supplies, regulators, and various low noise supplies.

I want everyone to be in no doubt though that our conversations really touched only on aspects of tuning your new interface. From what I know the new interface you have come up with is truly inspired in its conception and daring in it’s execution. I just don’t know how you do it, but I am very glad you do. I cannot wait to give it a try.

Kind regards,
Nick

[manisandher]

With the knowledge of XXHighEnd and SQ control, together with the development experience of the original NOS1 - at last it was achieved that the software influence has been eliminated. SQ Controls in XXHighEnd won't do a thing to the sonic performance of the new NOS1, as won't different software sound different anymore...

But do different PCs running XXHighEnd sound different? Is it now acceptable to use a laptop? I'm assuming that if the influence of the SW has been eliminated then the influence of the HW has been eliminated also... no?

I ask this because a friend of mine is very interested in buying a NOS1 but would really like to use a laptop, not a desktop, in his main room.

Mani.

[PeterSt]

I better had not talked about eliminating those differences eh ? haha

Let me put it like this :
You (Mani) have the experience with an Atom based system, right ? or maybe even a latop, I don't know;
Now tell me, would you like to use it ?

The story -by now- is too complicated to explain, but maybe it is better to look at 0.9z-6 and how that works. Then you will understand that you still need "the software" first to next be able to eliminate the "controls".
Maybe it isn't even right to say that the software stopped influencing "at last", but it does after first running the software in "that mode" (in 0.9z-5 nothing of that is in there).

Think of this stupid (completely made up) example :
Suppose your only option would be to use your system in a not bit perfect fashion. Like XP without special drivers always keeps on dithering. Now, XX could have been made so that it can change the dither means. So, you could influence sound by that. Next comes a version along that allows for bit perfect playback forever. And now the dither control obviously can go ...

It is also (for me) a psychological thing (hey Mani, for you too, I know); At first with USB I *needed* all the controls I could find to improve on the bad sound. It helped, but if one thing was right, the other was not. Not much different from what we all experience, no matter what DAC we use. This inherently prooves that something is wrong in the first place, but stil the controls could (would) help. This is where my ever espression came from : when your DAC is not 100% right, you can always influence the sound of it. Well, now it had to be the other way around ... how to get the DAC 100% (??) right, so the controls can go.
But the trick is ... both are the same properties. -> Find that major "control" and next fix that. This still needs that control hence the software. In the mean time though the DAC is "100%" also, but now the environment can still destroy (noise). And besides, which those "??" because it won't be 100%. This is where the influence remains.

Riddles ? voodoo again ?
Let's see ourselves.

Peter

PS: Also look at my sig and current settings. Quite the opposite of KS Special Mode Ultra Low Latency, right ? That's because I now turned it into a least demanding system. I mean, why not, if it doesn't influence SQ anymore anyway.

[manisandher]

You (Mani) have the experience with an Atom based system, right ? or maybe even a latop, I don't know; Now tell me, would you like to use it ?

Yes I have experience with both. I'd happily use either, especially if "less demanding" system settings help. However, it would have to be silent (unnoticeable from listening position) and not take forever to do things in XXHighEnd (such as process flac files). But if the latter hasn't changed, then an Atom would be out, but a powerful laptop might be OK... BUT only if the SQ was not compromised in any way, shape or form.

The story -by now- is too complicated to explain, but maybe it is better to look at 0.9z-6 and how that works. Then you will understand that you still need "the software" first to next be able to eliminate the "controls".

OK, I think I understand this.

Riddles ? voodoo again ?

Well of course. This is the XXHighEnd forum after all!

Anyway, it'll be easy for me to try an Atom/laptop for myself when I have 0.9z-6 and the NOS1 USB upgrades done. But just to make things clear, I'm perfectly happy with my current PCs - I'm just trying to find a way of making the NOS1 more compelling for my friend, who is less willing to 'ruin' his beautiful living space for an obsession with sound.

Mani.

[PeterSt]

Ok, I will try the due 0.9z-6 on a laptop here. It might even sufficiently work, especially just because of those changes to XX.

I can't do that right now though, because it first needs to finish whatever it is on this "procedure". Somehow I'm more busy with emails and stuff at this moment ...

[pedal]

Dear Peter,

Re. Upgrade of my NOS1: When should I pack it down and ship it to you?

Best regards
Petter

[PeterSt]

Hi there Petter,

I answered to your email yesterday. Maybe you didn't get it ?
(and as how it looked I received your email only after several days after you sent it ...)

Regards,
Peter

[earflappin]

Anybody heard the latest and greatest all that NOS1 USB version in their own system who can comment on its performance versus the previous version which appeared to be a bit sterile sounding for many?  Thanks.

[PeterSt]

David,

Unless you receive private emails from people with contents I don't know about, where do you get this "sterile" from ??
I don't recall this was written in here somewhere. Maybe elsewhere though (link ?).

Of course "sterile" can be the or your interpretation of super clean or something, which is what I myself maybe dedicate to the original NOS1, and what sure could be people's perception. But all is relative ... The NOS1-USB is outrageously "clean", but in a way which can't be made clear "anymore". I say it like this because it is beyond understanding. For example about the old recordings which may sound dirty, the original NOS1 really making something of them, but the USB version shows what's really happening. And suddenly no dirt or anything exists anymore. It must be all about the brain; old recordings carry old technique of doing it, and once the brain understands it doesn't hurt. At all !
It really is very strange, and as I said elsewhere, it needs other phenomena than we currently know of to describe this.

With the perceived "sterile" (which really sounds as a negative to me), there was this superb accuracy. It seems to go along, right ? Well, it does not. Not if you listen to the NOS1-USB. There's accuracy beyond accuracy, because now, no matter what, you will be unable to detect what's actually wrong with any music, unless ... -watch out- unless it is about a wrongly digitized transfer. You can now just hear what is wrong with it, and where it is wrong. Strangely enough (but actually not) this only happens (doesn't sound good) to those poorly done remasters. This wasn't before at all.

Because this accuracy has impact on everything, the whole perceivement is radically changed. This already was so for the original NOS1, but it is the more so for the NOS1-USB. So, basses change again more to real basses, and for example, now you really can hear the exposure an electric bass has (from its full body wood cabinet). I think this might be *the* example of how again more accuracy can dig up details which seemed impossible before. And let me add to it that all the added detail I perceive myself lately is hardly in the highs - something which seems to be prone for perceiving accuracy in the first place. But to my findings, at some stage this is behind you and further inherent improvements won't exhibit there anymore (except for the impossibility to hear distortions). Instead it sneaks into everything, and in the end "accuracy" exhibits as more natural representation. So, it changed from a technical merit into a functional one.

David, I know you didn't ask for all this, but a couple of weeks further I learn myself as well. It just needs a post like yours to tell about it a little ...

Thanks and regards,
Peter

[PeterSt]

To those waiting ...

Except for one, nobody really knows - but during the stage of getting the first NOS1 shipped a few things happened ...

I'm talking about the day I created this very thread, being confident that all was good to go. Ok, not enough stock to really satisfy all soon, but I knew that and I told about it.
The first NOS1-USB really was completely ready; it was measured and everything, and we were about to book it for shipment. But suddenly there came a message from someone (haha) owning two NOS1's, and both exhibited the same offset change, going from day one to day two. This combined with other experiences, and I said to myself, till here and no further. The story is much longer, and in the end is related to our mains systems and house grounding. Never mind this for now. But I wanted to definitely solve it. And not to forget : NOT by means of something in the signal path ... (how difficult can one make it for himself ).

After a search of two full weeks - soldering in and out and back again, SMD stuff and chips and everything - I was as far as two sections on the gain stage possibly influcencing eachother in an illegal way (not that electronics or simulations could proove that), that impeded by whatever groundloops coming from external.
In the mean time, we already found that we were able to create the offset just by hovering the NOS1 like reaching for switching it on ... However, one person could incur for that, and the other could not. The difference was in the shoes (isolating rubber vs. not isolating). Ok, it's not a strange story by itself that "we" guide current to earth, but part of the long story is that once a ground loop is setup like this, you can't get rid of it anymore. So, it is you who incur for the loop, but the loop goes through the DAC and stays. This is all in very brief ...

After tearing down really every part of that new NOS1-USB we expected a possible culprit to be in two "fighting" parts of the gain stage, and so I set myself to eliminate those two parts into one only; no fight possible with one part.
Well, in the spirit of Nick's activities again, this part which could be eliminated is a part which creates 2.5V out of 5V. This is a shunt regulated section consisting of some chips and transistors, and if you think of how a "shunt" operates, yes, this is prone to misery (you can't even see that things go wrong because of the ever the same voltage and actually same current flowing).
Of course the transfer to 2.5V was there for a reason, so that reason had to be changed / eliminated.
In brief again, I managed. Took days of reading datasheets at finding the suitable parts, preferrably all fitting the existing PCB or otherwise hell would break loose on redesigning that - reproducing that (preferrably in test first) - waiting for all that. But this turned out not to be necessary.

As Nick will know, I already was hunting for letting all run on exactly one voltage only, and this is exactly what is happening at this moment. And the big fun : the 7-8uV output noise has decreased to the 6uV range. Haha.
I know, people declare me crazy, but since I'm saying that noise rides ON the signal, and is not under it, it must be a good thing. And in the end you could say that there's 15% less output noise now, which seems a large figure to me.
I'm 100% sure it is audible as well, and at some strange "level" new details pop up. This is so audible that my partner in crime here told me that distortion was exhibiting from a certain familiar track, while really it is in the music itself (and no distortion, but a very fast on/off synthesized sound). So, things look to be faster once again. Btw, an example would be the Mini Moog from Rick Wakeman on Journey to the Center of The Earth and the higher pitched tones. This now is pure on/off sound, while at first it merely was a fluent synth sound.

The mechanism to avoid the ground loop -which is that 140 euro upgrade- actually can go now. Also for the 110V systems where it was created for. However, it will stay in, because in the mean time it's a provision to keep the PSU "hot" while the DAC section is (can be) shut off.
Also nice to know may be the fact that the rails for the output stage now indeed is +/- 5V while 2.1V (+/-) only is needed, and although before there was 0.4V headroom on the 2.5V rails (for overshoots) now there's headroom which never will be utilized.

In the mean time we are out of other stock, this time on ready gain stage PCBs. Well, that happens when you prepare a bunch while afterwards they need to be changed. At least today all the new parts for it will arrive, but it remains a little set back at changing the existing ones, which is a necessity because we're out of fresh PCBs. Wat a laugh.

All 'n all it has been taken longer than planned to get this really rolling, but at least I am glad that no single NOS1-USB has left with "older specs".
In the mean time a possible DC Offset problem (wearing around in two ghost hunted English houses with a leakage to one in Norway) now is history. I sure am glad about that.

Regards,
Peter

[Eric]

Hello Peter,
just curious,
would the USB DAC also need to be shut down when the computer is shut down?

Cheers, Eric.

[PeterSt]

That's all history.

You now can shut off either whenever you like it.

[manisandher]

In the mean time a possible DC Offset problem (wearing around in two ghost hunted English houses with a leakage to one in Norway) now is history.

Yeah, my Victorian ghosts told me they like travelling to more Northern climbs when the weather gets too hot here (not very often, of course). But I'm glad it's all sorted out now as they were getting fed up of all my accusations of what they did to my NOS1s when I was away.

Anyway, really nice story Peter. And as I've said in a PM, there's absolutely no rush on my part - I'm actually quite enjoying listening to LPs right now - but I'm very much looking forward to hearing the 'new' NOS1. (BTW, at what point will it become the NOS2?)

Mani.