"Hotwires" EMI, mains hum and data transfer....

[Nick]

“Hotwire” Drains..

A very exciting post to be making and after some significant time working on this .....

An idea about clock and music data signal integrity has grown from working on my system / Chronos and through pondering on a number of posts that people have been made here.

In part this is about grounding but much better to see it as improving signal integrity as data / clock signals are moved through the PC, over the USB link and within the NOS1a.
A lot of stuff has fed into this:

• The some principles applied to develop my Chronos server. It actually took 6 weeks to package the server whilst maintaining sound quality, well SQ improving in the end but it was far from easy.A lot of time was spent measuring environmental noise to assist the packaging and this and the stuff below was the real kick off point,
  
• This post on EMI and clocks,
  http://www.phasure.com/index.php?topic=2888.msg30456#msg30456
  
• This post (make what you will of this haha),
  http://www.phasure.com/index.php?topic=3069.msg33264#msg33264
  
• Peters posts on Silverstone Card earth set up and,
  
• Rather too many experiments on USB cards and Hard drives linear ATx and SMPS ATx supplies.
  

The current position here with my systems is that I have put in key positions 3 tuneable grounds within my PC, and 4 of these grounds within my NOS1a. Its satisfying that although they are applied to different systems, (eg 4 clocks, USB Card, USB interface and DAC PCB), they all behave in exactly the same way and very predictably. Each has a very similar sweet spot that can easily be found !

As I said before I am sure this is more digital about signal integrity than lowering analogue noise floor, in fact I would not be surprised if a few DB of analogue noise is sacrificed here in order to improve digital signal transfer !

So sound quality,

These comments have to be seen as relative to the starting position which is with a Chronos and NOS1a, from this position the changes are not subtle, the sound has;

+Silky smooth with a bell like clarity to the presentation,

+Dissociates sound from the speakers

+Very deep and involving sound stage (occupy’s up to 50% of my room on some tracks)

+VERY nice bass (dance music is like being in a club, always a key objective for me)

+“Snap”, “speed” and “rhythm” all superb, real seat of the pants listening with the right music.

A few possible caveats

+I have no measurements to show this but as mentioned its possible that analogue noise might be increased to improve data transfer conditions,

+It could be this it only works in my system. There are some strange things happening at times like the -144db i2s breakthrough sound get here. But I would guess this is very generally applicable and my guess right now is it could remove a lot of the variability you hear between “similar” systems.

+“Mileage may vary” as they say in other systems, the impact may be relate to PE conditions in other systems. But again I would be optimistic that this would work.




Its been a very nice afternoon listening and tuning these “hotwires”, next step is to try to measure their effect  quantitatively and then they need to be tested in another system.

I really am VERY happy with sound here right now. 

Nick.

[PeterSt]

So Nick ...

I had a longer response at first (soon after you posted) but in the end I could not see the relevance and merely thought that this (your post) should lead to something for the even better and now what. So today, with a hopefully fresh mind, this is my re-attempt :

In my first attempt I started out with "you like to talk in secrets these days". Today I can't get around that again, so I just said it (again). However, clearly thinking in a constructive fashion ... what's next ?

And so I must assume that "we" need to work this out and where my contribution could be the measurements. But mind you, this won't be more than noise measurements, assumed that what we hear for SQ improvement can't be measured in qualitative fashion anyway. Still the noise will tell a lot, especially if it can be measured in the digital domain (but more "analogue" noise can be measured for sure of course, and let's state that the more of that will not be for the better).

But now I need to know what to do. I mean, where to connect what etc.

Thanks,
Peter

[Nick]

Peter hi,

No problems regards the arrangements within the NOS1. Inside the PC the arrangements are specific to the Chronos and useless to anyone that does not have one, so I will keep those to myself.

I will pm you with the details.

The caveats may hold, mileage may vary but here the sound is great. I think that there may be less of a need to do stuff to the usb card ground 

Best regards,

Nick.

[PeterSt]

Hey Nick,

So I gave your PM some thorough analysis and will answer that PM directly as well, but say for public notice I will respond here first - again with a fresh and open mind.

I will first try to define what you are doing, without unveiling secrets might they be there anyway (I think not) so you too will know what we are dealing with :

The control of ground paths or ground loops of you will. So where (return) current flows and where it should not.

Key thing to emphasize is that you say yourself clearly "YMMV" etc. etc. and I think this is justified. But what I like to add to it is that this is needed (the YMMV) because it is nothing official. And the only part official in there is related to PE (protective earth) while PE is a pile of sh*t. Well, YMMV because this is the case in my house accidentally, so nothing to build on (so to speak).

You also suggested that possibly this is nothing to measure hence measuring possibly is not worth while. I actually agree with that in advance, but in aftermath I am not sure any more. So this could be part of my PM response :
If you measure noise to go down by this largely, then I wonder what you measure in the first place because I measure all under 160dB anyway (maybe a peak here and there up to -145dB) and I don't use such tweaks. If I combine this with your low resolution analyser then ...
Then you measure wrongly or you have it very wrong. But for example, I know how things change for the worse when I use PE ... (and with the isolated NOS1a this now counts for everything (previously the PC was allowed to be in PE)).

Which brings be to the last part : the isolation.
I think you say it yourself that what you apply somewhat defeats the isolation, but in my view it is completely illegal what you are dowing. Ok, it is "half" illegal because of the means / the way you apply it - I hope you understand what I mean.

Something quite else - and even interesting - is that you could be focusing on EMI. And this *is* in the title of this topic, you know. This doesn't get highlighted all the way, but actually it is me who tells that isolation itself is not allowed. I could be the only one in the world - I ignorantly did it anyway, but what one should get from that is EMI.
And is that bad ?
If you now a. observe that and b. tune on that, then it is my advice to break the isolation all together (ehm, back to the NOS1 that is). But but but ... of course not.
So the real advice is :
1. Don't pay attention to it in the first place because what tells you what it does for SQ ? (hence you are guessing in that department);
2. Solve it how you can (yes please) but not by shortcutting the isolation (shall we call that "daft" ?)

That you net receive a better SQ because of the appliances - why not. But as usual I would look elsewhere for the real cause hence the real solution.

What could be nice for this topic - and hopefully helpful - is that you try to explain (or show when you like / can) what you mean with mains hum etc. So I can tell you that that too won't be worse than -120dB measured anywhere (supply or output) while I think that your analyser won't even allow measurement that deep. So what actually is it that you see and measured from where ? or ... if you still talk about the hum etc. audible from the speakers ... (are you ?)

Best regards,
Peter

[Nick]

Peter hi,

I am on the way up to London now so a more detailed response to come.

A couple of points for now, my PE is a dedicated array of ground spikes so not as dependant on the house mains wiring.

I will explain the measurements but remember I have only the most basic fft. Measurement is based on trace amplitude of 50hz hum pickup on an otherwise floating ground plain.

There is scope for some interesting conversations about the theory but try it out and listen, the caveat about milage may vary applies the first possible difference relating to PE is called out above.

Have a think about the way that the ground refferance and the circuit components on the input of the use interface to the nos1 usb codec.
In not saying this it is but if it's about better data transfer then could be something there.

Kind regards,

Nick.