Eureka !

[PeterSt]

As in : Too stupid not to realize earlier on ...

Working on our Linear Power Supply, at some stage I noticed strange things; very adverse to earlier findings. Examples :

- Albums which showed the most silky previously now sounded the most snappy;
- Albums which sounded the most snappy previously, now sounded very silky (which quickly is perceived as too silky);
- The Silverstone card previously was needed to make al sound more realistically. Now the Silverstone card is not to be used because it makes the sound a sort of "bad" (say that suddenly you are unable to find a good sounding setting).

Yesterday for the first time all (for the LPS) was officially assembled in a case, meaning that all was screwed and bolted and everything needed to, say, ship it. Call it the final test to see whether all could really fit like I anticipated and of course that nothing gets hot in its final position.

But nothing sounded right to me. Or say that last night I had no real pleasure in the music.
I survived the night with good sleep because I also changed back to a 47000uF capacitor of which I earlier on had the idea it was too large to be good in its particular place (but also to see whether the PSU would still fit in its final place with that on it). So I blamed the capacitor ...

When one is working on a prototype, this often goes in some wooden box, just to be able to carry it back and forth between the test location and the soldering bench.
When one, like me, is not using protectice earth explicitly for some means of grounding, and only uses that for safety and say connect PE to the case only, in such a wooden box setup there is no connection with PE; you wouldn't know what to connect it to (wood really is quite useless now ).

When the setup becomes final, in its intended (kind of) final case, we must think of protection and thus blindly connect PE to the chassis. Blindly means : you really don't even consider it. Well, at least I didn't when I received the question (this really happened yesterday) : to what point in the chassis shall I connect the PE lead.

So now you think where I am going to, right ?

I don't think so.


This morning, in bed with my eyes half open, I recalled that one of the last things I suddenly noticed last night, was the again hammering emphasis of Windows 10 10586.0 "distortion" sound. You know, that sound I could take out with the Intona and after which 10586.0 suddenly was the best of them all.

I am sure that back at the time I described that sound as "too much of the Silverstone". IOW, it is an on/off sound which can do things right in the highs (it makes cymbals more realistic) but when too much of it, it is sheer distortion (which I think it is anyway, but alas). Now :

I don't know the real combination of matters yet, but it is now obvious that the past 10 days or so I had been listening to no PE connection at all - that the Silverstone had to go out, and that the PE connection now back in place destroys much of what has been for the better since my LPS. But it must go very sneaky ...

That Audio PC already was not connected to PE. BUT - and this is with all my audio gear - it connects to the PE lead of each device. So envision I have this power block and while it is not connected to PE, internally it has the PE leads connected among the sockets. So what I'm actually doing is connecting the chassis' of each device, as long as the power cord (and connector) used, carries a PE lead.

The sound with the LPS in use, generally is inmensely more lean. Think super light-footed and this is also where the vast amount of additional detail comes from. This is also how (again) the slightest disturbance in the sound becomes obvious.
So what actually may have happened with my two types of connections - prior to the LPS and when using it of which we know I removed the Silverstone (which btw was because of too black sound) ?

SMPS

a. Chassis of PC connects to (e.g.) Chassis of NOS1a via PE leads and power block.

b. In NOS1a PE is not used further.

c. In PC I never investigated really, but we do know that the ground of USB connects to PE hence connects to the PC chassis.

d. We use an Isolated Silverstone card, meaning that we don't allow at least the shield of the USB cable to connect to the PC chassis. What happens further with that is up to the PC (again I didn't investigate and I don't know).

e. We use an Intona which isolates the USB concection itself.

f. We use an NOS1a which means another level of isolation, this time between the USB receiver and the i2s interface.
The famous "black wire" is disconnected in order to prevent a back door gnd connection to the PC (which again would be the PE lead connection).



LPS without PE connected to the PC chassis

A. We do not use the isolated Silverstone - just the USB3 MoBo connection.

B. There is now no connection to PE at all, because the PE lead just is not connected. There is thus also no PC connection to any of the other devices. This is also because of :

C. We use an Intona which isolates the PC from the NOS1a.

D. We use an NOS1a which means another level of isolation, this time between the USB receiver and the i2s interface.
The famous "black wire" is disconnected in order to prevent a back door gnd connection to the PC (which again would be the PE lead connection).


Important side note :
I once observed something like the case of the NOS1 being able to act as a capacitor; it charges and discharges through air. Do notice though that this is explicitly without PE connection (so to the real protective earth (pin)), or otherwise this would not happen.
Do also notice that the say 1 meter of PE wire length in your mains cord, already is a capacitor which is very tough to discharge (read : it takes 30 minutes to do that). I am now talking about the situation that the power cord is connected to the NOS1(a I did not check back at the time because it did not exist) and that I can read a voltage between (IIRC) the output and the chassis (which is the PE wire - which wire is NOT connected at the other side, but obviously sits in between a hot and neutral mains wire (induction ?).


With the knowledge that "signal" travels through air and is radiated by the one device while captured (antenna) by the other, all sorts of things happen we can not see and most probably don't like either. Measuring it almost goes by coincidentally seeing it - to keep it in mind and hopefully utilize it later.
Could be today.

The crux is somewhere in the above two lists;
Because of what I observed yesterday, which now accidentally is with the reference of that not being present the past 10 days (no PE used towards the PC), I suddenly hear USB backfiring onto itself. This is high stipulation, but something like that which also would suit the connection of USB gnd to PE, which thus via the huge backdoor again connects to the NOS1(a), although outside of the real protective earth (and possibly crucial : with a same beat it has interently, but now with different timing (think 20ns because iof power cord lengths). Not using the PE connection in the PC now makes any USB connection impossible (?).
Mind you please, prior to the Intona we could not think like that. Also not because of the "a" version of the NOS1, because when the USB cable entered the case, at that moment a connection to the PC is still there and easily think now some elements start at least to act as a capacitor (like ground itself, in front of the isolation of the "a"). This should at least influence the working of the USB connection (inherent jitter or something) which could be taken out by means of the Intona.
What remained tough - at least in my situation, is that the USB gnd plane which bounces up and down in the PC itself, fires back via the PE leads on to the NOS1(a) chassis, which again is going to be a capapacitor (but a very small and fast reacting one, which exactly could be the culprit).

Now before you all think that you won't be suffering from this because you do not use such power block and you really don't have the PE leads connected anywhere, think twice. You 100% most certainly will be using that power cord with PE lead in it. This is fine, as long as it does not connect to any chassis, or otherwise it is there (the PE lead) where your capacitor is charging and discharging (I am pretty sure of that). Sadly it now is a bit difficult to solve because the standard connector which goes into our devices, has that lead connected. Easiest would be to break off the PE pin ?
So notice that in-NOS1 you are able to disconnect the PE wire by means of the bolt it is connected to, but with your PC you can not do this while I am not even sure you should attempt this (over here the situation is easy because the LPS is all mine and I know what is happening).

Eureka ?

Peter

[pedal]

Thanks for the update and "Lab diary". Very interesting to learn about your discoveries and the ongoing process. -So many variables makes it difficult to judge if a tweak is one step forward, backward or sideways.

Silverstone: Since Day 1 I have voiced my concern about its edgy and colored treble. (Although I have to admit it works very well in my current system, after the Clairixa arrived). But I won't be surprised if it ends on the scr*pyard in the end...

I have had the privilege of being ringside with the constructor of my amplifiers, Norwegian brand Dynamic Precision. -Very big power amps, DC-coupled and approaching 1MHz bandwidth. He was sitting just like  you now, with his nose into the chasis while measuring ground loops at "0,000000" level. It was his obsession for 30 years. Actually it started even earlier, from his lab-days at Tandberg in the 70s. He worked on their FM-tuners, were high frequency noise and signal transfer is key parameters. I remember he explained me how ground loops at (very) high frequencies bypasses almost anything. Even through air. And certainly an ordinary iso-trafo doesn't stop it.

My friend is now retired, spending most of his time practicing his jazz guitar, but fortunately I have another friend in Peter, who lets me continue the joyride into unknown technical territory towards State Of The Art sound 

[AlainGr]

Peter,

From what I understood, the Silverstone USB card may not be the best for the task now that you have added a LPSU ?

(OT)
As you know, I have had many problems with voices recently but with your advice I was finally able to consider the sound "acceptable" even if it still sounds too "scintillating" to my taste at times (not all voices, just some).

I brought back my Sotm PCIe USB 3 card in the portrait to check if something would change. It does, but to be honest, I don't know if it is for the better. The only reason I brought back this card is that I can connect a LPSU on it and have "clean" Vbus power for the Intona, in hope it would help something.
(End OT)

You seem to mention (...) that the ground does not drain correctly depending on the PE lift we apply to some components ? And that can accumulate and discharge at a periodic rate (or when it reaches a certain level), affecting the SQ ?

I read about Entreq ground boxes... Would this potentially help for this situation or am I still in "dreamland" ?

And if it affects treble, I would then not be too surprised for the results I can hear at times... Could that partly explain for what I hear already ? I am using Win 10 build 10586.0 from the os-ram drive...

Regards,

Alain (still talking on behalf of Anonymous)

[manisandher]

Wow... I tried to understand your post Peter, but there are simply too many PEs via power cables/blocks, isolated/non-isolated grounds via interconnects, and isolated/non-isolated grounds via USB for my little brain to cope with 

But getting this all sorted is way more important than the vast majority of people would ever be able to accept, let alone understand.

Mani.

[PeterSt]

You seem to mention (...) that the ground does not drain correctly depending on the PE lift we apply to some components ? And that can accumulate and discharge at a periodic rate (or when it reaches a certain level), affecting the SQ ?

Hi Alain,

You hit quite some nails on the head here, were it for my thinking. So I don't think I explicitly told about all that, but these things do happen in my view and they happen in a sneaky nasty way. An example I know from my own situation :

I have the amps (in my case in the speakers) 100% always on. However, what can happen is that I disconnect the lot from the inputs (this goes by a switch), so I can mess around with software or DAC or other tests without the windows going out.
By now I know that the trick is to disconnet from those inputs for an hour or so, and after that playback is optimal.
This can last a day or 2-3 and then I have to do it again.

To me this means that something gets charged and THAT I saw so nicely happening with tha powercord which wasn't even connected. And now I don't know what this does or how it works. But say that when something can get charges (very the same to a battery !) it can (thus !) also discharge and this latter may (or will ?) go by radiation. And although there may hardly be current involved, the (radio) waves may influence.
Maybe look at this one again : WiFi impeded noise ?

Not that we are talking about WiFi, but that too is radiation from an as random source, and it can be seen at the output of the D/A converter.

So if you think like that, anything can be happening;
If we know the possible sources (at least theoretical ones) we better try to eliminate those, just in case. Btw, once I found out about this WiFi stuff (see link) I could readily hear it when that PS3 was on, never mind nobody had switched on a TV for it - and was only downloading an update etc..
Of course you must learn the sound of such a thing, but this is "doable".

Mind you ... that I have connected my devices' cabinets was not intended at all. As a matter of fact I only "dicovered" when I wrote the first post in this topic ...
(at this moment this is still so, but the PE in the PC is disconnected and sound was great again).

Regards,
Peter

[PeterSt]

Wow... I tried to understand your post Peter, but there are simply too many PEs via power cables/blocks, isolated/non-isolated grounds via interconnects, and isolated/non-isolated grounds via USB for my little brain to cope with 

Mani, I was thinking about some response, but couldn't readily find one.
So I forced myself into this one :

Supposed it would be true that by whatever means we "leak" noise or even signal via an unintended path ...

Assumed the law of conservation of energy still holds true ()
( https://en.wikipedia.org/wiki/Conservation_of_energy )
...

then the other way of approaching this is stating that part of the signal travels another route, but that *thus* the original must have been degraded (say got less in energy) because some was taken out of it and goes elsewhere.

That this part of the signal, next, may arrive at the intended inputs again with a time lag (like I told about) is yet that other reason to be not happy. But for now thus :
If we allow to let leak signal away (where the ground (gnd) participates in) then we just degrade the signal by that alone.

This could be new, fresh thinking (aiming for better solutions).

Peter

[juanpmar]

I don't know if it has something to do with it or not. When I am a few days out of home I disconnect the system and when I return the first day I'm listening music the sound seems to be much better. I always thought that it had do with my ears that forgot how good it was sounding but also suspected that unplugging it for some time the system was cleaned.

Juan
 

[AlainGr]

Peter,

The reason why I expressed things like this is that like Mani, I was unsure of my understanding from your first post. Each of us has his way to describe something and the more we are unsure of what is happening, the more vague the attempt to explain.

I tend to make a parallel to what I experienced here but I finally concluded that my ears were responsible for the say "excess" of granularity I hear through many voices. And I only perceive this with voices (not the rest, at least not in my perceptions).

I even cut all the power in my house except for 2 wall outlets to determine if something was still happening (and it was still there). The subs were completely disconnected and all that was left was my audio chain on one circuit and my server PC on another. No lights (it was during the day) ON, just the system.

Applying your suggestion to blow my ears, though not working first, seemed then to help a little. I say "seemed" at lot, because I can't pretend I nailed anything. I even thought at one point that the Intona could be affected because the signal passing through it was weak (since I apply the XXHE digital volume between -18 and -21DB). Putting back the digital volume at -3DB (since I use a preamp with an analog volume), it did not really prove what I was thinking. Even removing the Intona did not really change a thing. Sor far for my hypothesis...

So your post brought me back to my situation. When we lift the ground on a component (say with a "cheater"), how and where does the energy that should normally flow toward the return path go and if not, how does it behave ?

I suppose that, like Juan reports, something may drain itself when totally disconnected.

So your post, in my twisted way of understanding, "lifted a flag".

In my layman understanding, the third branch (PE or ground pin) adds to the complexity with the neutral. I may be completely wrong...

I would like so much to do meaasurements with a standard meter to see what are the different potentials but... I have a problem. How do I do this... I touch the computer casing for example, but where do I put the other wire to measure anything ? Do I do this between two different components ?

As you can see, my ignorance is showing and I feel... Ignorant.

That is why I thought about a "ground drain" and I was wondering if something like en Entreq box could be part of the solution...

Sorry for that long post (by the way)...

Alain

 PS: Not necessarily an Entreq box, but a Phasure Ground box maybe ?

[PeterSt]

So your post brought me back to my situation. When we lift the ground on a component (say with a "cheater"), how and where does the energy that should normally flow toward the return path go and if not, how does it behave ?

Alain - this could be mistake #1;
The Protective Earth lead is not a return path as such (jn our situation). That would be the Neutral (say the 2nd of the three wires). (Another) Point is though :
It is quite easy to let PE act as "massive" return path by means of, well, just using it. For example (at least in my thinking) : Cut the gnd of an interlink at both sides, lead each side to PE instead and done. You now will have ground to your audio signal and it is only that it doesn't go as direct as it could have (namely via your interlink). And OK, that ground path (thus via PE) is noisy as hell, but it will work (uhm, assumed that both devices connect nicely to PE of course).

This is not soooo strange thinking, if we see that often a cabinet's chassis is used as central ground point, while the chassis is also connected to PE for safety reasons (this latter should always be the (normal) situation, while the former isn't good thinking - alas, in my personal view).

That PE is actually bonded to the mains neutral is something else which is vague and although logic when you dive into that, it is "strange" when you did not sort that all out (but let's not mind for now).

That is why I thought about a "ground drain" and I was wondering if something like en Entreq box could be part of the solution...

This could be useful when one first thinks differently;
I too, at first, was of the opinion that noise "can be guided away". This thinking springs from the fact that you can see it happening (often). So you see some noise (on whatever measurement device) and when you connect the cabinet (or something else) to ground, the noise disappears. Well, let's say that this can be true for those situations that the ground is easily influencable because it being "weak"; say it is too small and too light and therefore it can be moved (this is not easy to explain, if I make sense at all, and/but think that our closed box is a close ecosystem with its own ground - and it is small). Not so with PE which actually is "the earth" (as in the globe). This is rock heavy. So give that some vibes (noise) and nothing moves. OK, it is polluted a little, but not really visible.
Meanwhile you indeed give your noise to PE and the next device has to digest it (as pollution).
Again, this *was* my thinking ...

This can't be my thinking any more if I first eliminate PE all together - btw for the good reason of not wanting to see another device's noise. But now it is all up to the individual devices ...
Our great (not) example is of course the spitting power supply, of which we know that the chassis is connected to PE and of which we also know that it's signal ground is connected to that. And so the PE is polluted heavily, and thus also the neutral because bonded.
Nice !

So skipping to the end of the other way of thinking : When I am now capable of not connecting the PC chassis to PE (just because I use my own PSU and it is nothing much different than a D/A converter or amplifier for that matter), then I must have achieved something (say new).

a. Nothing is spitting towards the mains for SMPS reasons (it just does not exist);
b. No USB noise (whatever) can additionally spit towards the mains (or as the only source of spitting if you like) because nothing in the case connects to the mains.


Side note :

Although it has been said more often (by me), not everybody will have read this :

Many countries will not understand how it can be safe to live without PE (which in the end mainly is there to let trip short cuts with Hot/Neutral so a main circuit breaker can be triggered to cut off power). In Holland though, most of the houses do not have PE in the living room or bedrooms and such; it will be there in the kitchen and bathroom, but there the danger of shortcutting with water exists (read : your body makes contact with 115/230V and the current flows through you to the earth which is the ground you are standing on). And true, when some idiot let float his D/A converter on water in the living room, he better have PE around in that area as well.
Over here, new houses (since 15-20 years ?) will have PE in every room and every wall outlet, just because it can't harm. But necessary ?
So please notice that I am not claiming anything regarding the safety of this, but it can be determined by what is and what is not allowed by government regulation (are governments to be trusted ?).

https://en.wikipedia.org/wiki/Earthing_system
(which is only an easy document on this - I recall an in-depth research of how a sky scr*per caught fire because of some mis-thinking in this PE area, which was really a bit beyond me (as I recall it)).

Regards,
Peter

[AlainGr]

Ah... It does explain the USB card isolation and the disconnection of the black wire that we did some years ago, before the "a" version can do it automatically at the NOS1 end...

Does the Intona make this application useless now ?

Alain

[PeterSt]

Alain,

Does the Intona make this application useless now ?

This has been asked before. I don't know because I can't reason it out and it is too easy to make mistakes (because of so many unknown things) plus I don't try it (too many other things to test and listen for (besides some normal music listening once in a while)).

We "almost" can say that if each device is able to run independently of the other, such answers can be given with a fair chance of being correct on it (whatever). This requires isolation everywhere (also in the signal paths (Interlinks)) which theoretically can be done.

Next, radiation will be all over the place (because of the isolation and again "in my view") which is that other subject and which makes me say "almost".

Fact seems to be is that our sound is getting better and better and better with a regular interval.

When time permits, I see us ending up with poles in various corners, them acting as the noise capturing devices/antennas and which poles guide that to some earth-ground (or drive a small power plant).

Peter

PS: Remember that LPS (GPS) project I talked about and which I worked on prior to Phasure ?
I did that together with a guy I met on the Internet and we both went smart and did some nice things in this area. But it died out and I started Phasure. I only this minute realize what he was doing next (on his own) ...
... talking about poles driving a power plant ... he was going to grasp energy from mid-air and utilize that. Since this is now 10 years ago I should contact him and catch up a little.