PC electrical isolation from DAC

[PeterSt]

Alain,

Before I implied a misunderstanding ...
My no means any BIOS update will change that driver behavior. Any driver update won't either. Ok, but one : that from Adnaco (but forget about that, because they won't even know it's there like this, and otherwise why would they care (there's nothing wrong about it. -> Look at a Weiss driver haha)).

Drivers tend to sit in the last core, but with Hyperthreading active (not your case) it will be the one but last. All logic (to me), and no way to change it (I could for the NOS1, but there's no point and it will only make XXHighEnd less compatible with e.g. your device).

No need to worry about not being a NOS1 owner. Look at what cars we drive. If they start at all, they go too fast, stop their engine in the middle of nowhere, have no rooftop in the UK (that's a fault to begin with) or otherwise you will be a shrink and if *that* is not the case you will be ending up at one. There are just many ways to spend money I guess.
Okay okay ...

[AlainGr]

Hi Peter,

No problem at all. There is always something to discover each and every day and I am happy about this. Life would be boring if we knew everything

I have flashed the BIOS to the latest and the BSODs have stopped. There was also an update for USB3 (I guess this is in case I add a USB3 controler card ?).

But the behavior is still there, no matter what I do.

Will I accelerate the build of a new PC - not sure for the moment, it depends if that behaviour changes, or at least the driver puts itself in another core or hyperthread than XXHE engine ?

Anyway, now I will write to Adnaco and check with them if there is something to do - maybe like having a second CPU

Life is wonderful don't you think ?

Alain

[PeterSt]

Life is wonderful ? for sure when your posts are not as long as your emails - and they aren't. So, good. Haha.

I don't remember anymore, but I think indeed it is so that when two cores are available, XXHighEnd (or XXEngine3.exe actually) will put itself in the second core (depending on the Appointment Scheme - but never in the first alone). This was made when I didn't know about the driver behavior. So, without me changing that (which is a dangerous thing, already for all those people who found their best SQ setting regarding this), the solution is a 4 or more core processor. So, that will let stay Engine3 in the second, while the driver wil be further down the line (3rd for 4-core, 5th for 6 core, etc.).
But the solution looks a bit heavy to me.
Anyway, forget about Adnaco changing this because it is a deep down (OS) thing, and most probably they didn't make the driver themselves anyway plus that it will be long ago if they already did something about it. This means : hireing a programmer with the knowledge plus a few months to read into the stuff. So, no.

Life stays wonderful though.
Peter

[AlainGr]

Peter,

Coming from you, I find your comment about the length of my emails quite amusing

I am stuck with the Adnaco, may be it Maybe when I will decide to get a more powerful PC it will do the trick, but I will not bet my life on this :-)

Kind regards,

Alain

[Scroobius]

Out of interest I just checked my processor usage whilst running the ADNACO USB3 extender and there is no noticeable difference compared with normal USB connection.

P

[AlainGr]

Hi Scroobius,

I looks as if I would need more than 2 cores to get better results, since Peter indicated that the drivers tend to go to the last core (or hyperthread).

Maybe this expense is not totally wasted for me...

Thanks.

Alain

[PeterSt]

Nah, I don't think that will show a different result in something which happens in one core anyway. The point is that we don't know what Paul thinks is normal. It could show the same as you showed Alain. But it can't be as easy as it looks (from Paul's post) anyway, because there is no way that any two settings show the same for what Paul calls "normal USB". What he could be saying though is that no matter what setting, they differ amongst eachother all right, but he sees no difference with normal USB and the Adnaco.

When the latter is true, it should be so that you must be able to see differences between settings, which you say you don't (and which is hard to believe to begin with). BUT :

What I forgot about is the "physical" buffer length setting in the driver. Which driver ? well, like the one from the NOS1. So, *that* one encourages for the most visible behavior changes, while you, Alain, probably are not able to set anything at all regarding this part.

So I think I have nailed it down; The AudioPhilleo will obviously carry "one" buffer setting and that interacts with the Adnaco in one of these ways of which many exist (with some effort I could see a Picasso in the CPU graph at trying all the combinations). So what you lack is one of these parameters to tune this for the better (FWIW of course !).

Remember, that's what I think and nothing more than that.
Peter

[praphan]

Hi Paul and Nick,

There were report suggesting that fiber optic is more sensitive to vibration than copper. You may like to test this mechanical characteristic too.

Best regards,
Praphan

[AlainGr]

Hi Peter,

Thanks for the precisions. I forgot to say that with 15ms resolution, the graphs does not show the same. Instead of "peaks" or "bursts", if shows a lower "noise", but with not so steady peaks. If I can give an image of this: it would show continuous variations within 2-12% - same as .5ms... Nothing steady as those peaks at 1, 5, 10ms resolution...

My concern is actually the BSODs that keep on happening... I wrote to Adnaco about this and they sent me the previous driver (1.12.18) to compare results with the version I got with the S3B (1.12.24)... Same behavior. I wonder if it is not a problem with my S3B alone... But as you also mention, it could be related to a buffer management from the Audiophilleo - at this point, I don't feel knowledgeable at all...

I will wait and see what Adnaco will say about this (BSODs), but I would also like to see some pics from people who can post here with "Without Adnaco" and "With Adnaco".

I know it will sound futile, but the pics I posted were taken while music was playing...

Regards,

Alain

[Scroobius]

Hi Guys:

Nick was a most welcome guest here on Saturday and we had a really great day playing with our toys (actually we got diverted listening to music but anyway that is another story we did get back to business eventually). I had made up my mind during the week before but it was good to get Nicks view. On Saturday the result was clear and both Nick and I were in agreement  :  Fibre USB3 does not sound as good as normal copper USB3 in my system. With Fibre the sound has more edge to it and loses bass definition and warmth. Just the sort of sound I might associate with jitter (on asynch?) but who knows. Maybe jitter is introduced in the conversion but a total guess of course.

To start with when I first installed the fibre extender Peter's random setting generator set Q1=1 without me knowing. Some very odd things happened which did strange things when the very first comparison was made but as Q1=1 is all wrong I forgot all about it and made the comparison with my normal settings shown in the signature below.

The test was with 10m fibre - some say that if the length of fibre is short things improve - it will  be interesting to try that just to find out if light transmission does have an effect (not sure how it can though!). It does not seem likely but it will be interesting to try. It looks as though I will not be using fibre any time in the near future.

Actually this result is a big disappointment as I would have liked to get my PC out of the living room.

So the question is does electrically isolating the PC make an improvement? well if conversion to/from light was perfect maybe I could answer that but I cannot. I just got the feeling that even if we could make significant improvements to the fibre link that its performance might approach that of the copper USB link but it is difficult to imagine that it could significantly improve it. So could it be that provided our overall grounding arrangements are good that the electrical link to the PC might not be a problem.

The electronic ground in my amps is not connected to PE it is just connected to the electronic ground of NOS1 which in turn is not connected to PE but is connected to the ground of the USB cable which in turn is connected (possibly) to PE in the PC. Maybe as it turns out that is not such a bad thing after all. But if your electronic ground is connected to PE anywhere else AND via the USB link to the PC then that could of course change the results. Just a thought.

By the way I need to say that the switched mode power supply that comes with the ADNACO extender is very very bad the sound quality using it is truly bad. I used a beefy linear regulated  supply for the fibre receiver card which draws a hefty current. Sound quality with the regulated supply was much better and overall not that much worse than copper - but certainly not good enough for me.

Paul

[PeterSt]

Paul, Nick too, this doesn't leave anything else than complimenting you with this very thorough description of your findings. Not that I know everything about everything, but at least I can't shoot holes in your consistent elaboration of what is and what could be. Hats off.

There are so many trade offs ... too many to think of. The whole business of audio is a trade off they say. Well, I don't agree. As long as we keep judging seriously and keep both feet on the ground. You both just did.

Drove any nice cards lately ?


PS: I still want to try it myself, knowing that the only reference I can really have is me myself. Next, if I can't get it to work satisfactory I wish to split costs with Nick. I really do. After that we'll have a few drinks on it.

[AlainGr]

Hi Paul, Hi Nick,

Thanks for this review. It helped me to clear my mind. I will return my own S3B (provided that I can).

I did a silly test yesterday - I just used the S3B between my external drive and the PC. I wasn't sure at first, but when I connected my drive back with Esata, the sound regained its "plentiness". I did not like what I was hearing with the Adnaco...

Thanks again,

Alain

[Nick]

Paul Peter hi,

My visit to Pauls on Saturday was a great day, we had a fun time listening and trying things out and I was made to feel exceptionally welcome. This and the quality of the music being played resulting in my completely losing track of time  

I agree with Paul's description of the Anadco. It puts in a good performance but such is the resolving power of Paul's system (more in a moment on this....) that the results were very clear. So much so that in the end it only took a short time to compare and understand the differences. Our observations about the sound were very consistent so I think we were hearing the same characteristics of the copper and fibre usb systems.

If the criterion is the best possible sound then the Anadco USB 3 does not quite make it. The other virtues like ability to remove the pc from the listening room are possible but there would be a sonic compromise for this. We decided to leave the Anadco system at Paul's for testing of short fibre cables just in case the Anadco has as last trick to reveal whilst on tour in the uk, of course a visit to Holland may uncover more  

This brings me on to what for me was frankly a revaluation, listening to Paul's system has redefined my view of what is possible in music replay in a domestic environment. I was rocked on my heals and open jawed at what I was hearing throughout the day. With the NOS1, gainclones and Audio Note AN-Es coupled with meticulous setup, home made interconnects, speaker wiring and using Bybees, Paul has created something REALY exceptional.

With a decent recording there is not a hint of edge or hardness at all in the sound, there is detail that it is difficult to rationalise coming from a mere mechanical transducer, highs (cymbals etc) are at the same time both very prominent, lifelike and natural.  The system presents dynamics an unbelievable sound stage and presence and remains rock solid regardless of the complexity of music. All music types we listened to we're given this same virtuoso performance. The resolving power and musical ability of this ostensivly quite simple system left me stunned at what I was hearing. Very quickly we just moved on to exploring Paul's great collection of music during the afternoon.

This sounds like a rave but I think Paul has managed to create a very special system indead, god only knows what z09-7 is going to sound like if as Peter says we are listening to distortion right now with pre z09-7 software.

I cannot thank Paul's enough for showing me his system, I now understand where I want to get to with my system and it has confirmed for me that it can be done !


Regards,

Nick.

[PeterSt]

Thank you too Nick, for an explanantion from the heart. And yes, how important it is to be able to set yourself a reference "knowing that it can be done". On that matter I am still into providing a complete system (including amps and speakers), although I guess that I must hold my own reference against others first. I'm, quite a loner too you know (but a never much complaining one - or deaf enough ).

The other virtues like ability to remove the pc from the listening room are possible but there would be a sonic compromise for this.

Well, of course this is of great importance to some. In the end though, it could be an obsessive thing because it's really not needed with some careful thinking. Main thing could be that the keyboard/mouse/monitor can be operated from a distance (I use 12 meters of cable to that) and otherwise that more smart remote control does wonders. All obvious of course, but a different strategy from "not wanting to see the PC".

Of course the objective was that isolation, and I guess the real merits of that will never be unveiled as long as other forces play a role, and apples to oranges comparisons can be in order only. Input jitter can't play a role with this asynch operation, but what's needed to get the whole job done, can. Just think about what the power supply again does at the DAC's end, and you will understand that providing any random beefy PSU there may give as random results, since it matters anyway. So this is what I would do - arrange for that in at least the same fashion as how the native USB  connection was arranged for, which already was difficult enough. So, this needs measuring and what not, and that is what I will do of course. If *that*'s arranged for in well fashion (assumed it can be done) we will see further.

So, this is not to be given up upon yet, but how it can be better I don't see. But then I also don't see how USB3 can be better, while it is. And where some use the MoBo's USB3, others use a PCIe card and that is better again. So see ? this is not even related to any physical connection at the cable's end, nor will it be related to ground noise/loops/whatever. This is the PC alone, and now think how the Adnaco stuff will be different at least to that regard. And, as long as "I" am able to dial in some Q settings for different sound (PLUS my knowledge of it all and what to do to prevent that influence on which I FAIL), it is easy to see THAT ALL MATTERS.

Nice hobby ...
Peter

[manisandher]

Paul, Nick, thanks for sharing. But just in case it's of use to anyone, I've compared a standard blue 10m copper USB3 cable to a standard blue 1m copper USB3 cable and I can't hear any difference between the two. The 10m cable allows me to place Le Monster in the cellar well away from my listening room. No electrical isolation here, but the PC and rest of system run off the same mains circuit anyway (with a 1KV isolation trannie for the PC) to avoid ground loops.

... listening to Paul's system has redefined my view of what is possible in music replay in a domestic environment.

It sounds like Paul's system has come on somewhat since I listened to it, what a year or so ago? It was already pretty spectacular back then. So I may just have to invite myself across in the next few weeks... and put a few more miles on the NSX

Mani.