dedicated mains earth

[Diede]

Hi,

I’m currently setting-up a dedicated listening room, and to get the cleanest current I’m installing a dedicated fuse box (Siemens KE) with silver connections and HiFi-Tuning Ultimate Silver-Ceramic fuses in order to get the least possible interference with other appliances in the house.
I’m also installing a separate mains earth directly into the listening room.

My questions is what to connect to the this dedicated mains earth and what to the general house mains earth?
As I have multiple mono amps per side (Hypex Ncore with switching power supplies) my thought was to connect them to the house earth (in order to get rid of any high frequency stuff), and to connect the DAC and audio PC to the dedicated mains earth.

Anyone any thoughts on this?

Thanks,
Diede

[PeterSt]

Hello Diede, my 2c :

my thought was to connect them to the house earth (in order to get rid of any high frequency stuff)

Yes, that is the logical thought and in itself good thinking.
However, it will be more devistating because of other effects;

When you have two separated earth pins (or other means for separate earthing) you will be able to measure the potential difference between the two rings. What you will see is this :

Side note first :
Each device connected to the mains, gives the earth a slightly higher potential. Thus, connect a 220V light bulb of 100W (and switch on) and the earth potential may rise 1V. Of course I make up this figure, but it gives the idea of what's happening.
In my further view this is related to "voltage rails" as in :
The mains provided to you is 220V (or 240V) which assumes a 0V ground potential. Thus because this ground is 0V, the 220V delivered to your house remains that (remains 220V). If now your ground potential rises with 1V, you've 119V left and besides clipping may occur (on the sine (sinus) of the voltage). Please notice that I did not really test / measure such things explicitly, BUT that I have seen sufficiently strange sines from my own mains to find explanations like this.
In my own house I have measured as much as 107V potential difference between the house ring and a clean ring (nothing connected). 107 Volts ...
(but just oer 50V is normal in my situation)
So far the side note.

Whether this is detrimental to to the shape of the wave form is not so important now. Far more important anyway is the 107V difference which exists between the two rings. And, nothing is the matter with that as long as they stay separated. But now you are going to connect the two rings ...

From experience I know that there's only a very low current flowing, but the point is : where you connect the two by means of you amplifier(s) over the interlinks being connected to the DAC (for one example), you are molesting the signal ground in the interlinks. So there's a 107V wobbling on that, and this is not going to work well.
Notice that the strength of the 107V, in my view, depends on the current vs the current of the signal but point remains that it is there for influence.
Btw, the other day I tested laptops with batteries and for the same reason that would not work out (at all); possibly I posted pictures of that. Thus what happened here is that my Audio Mains Ring received a ground potential because of its own usage (say 10V all together) while the laptop is on a different potential (namely 0V) and that implying noise which was super easy to measure. Several 10s of dB's.

I'll hand you this as Part I of my 2c response.
Peter

[PeterSt]

I only now think about this again. I will come up with Part II tomorrow. Apologies !

Peter

[PeterSt]

Part II finally ...

In my previous post I implied that the power amps should not be on a separate ring. Nothing should be, actually. But true, such a noise spitting amp (btw not only with switching power supplies but also with digital (sawtooth etc.) behavior) should not be connected to the audio ring indeed. But what to do once you have them anyway ...
(I have no answer to that apart from the obvious)

In my earlier PC & Audio days, for a same reason as you suggested, I had the PC on a separate ring. Wrong too, and easy to measure that. But for that component of our chain there's this :

Have that PC as lean as possible, maybe give it a linear power supply, and connect that to the audio ring. I measure nothing wrong from that.
The last part now is the music storage, which best is to be another PC;

That other PC, the Music Server, is connected to another ring. Whether the normal house ring ? I guess that is fine (I have again another separate ring+earth for that, but alas). Now the monitor (one of the worst beasts for audio (noise)) is connected to that PC too, while the audio PC is controlled over RDC/RDP. Nothing different from what we talk about all the time (lately). Mind you, while the two PC's (Music Server PC and Audio PC) are connected by means of Ethernet, this is a galvanically isolated connection. Thus the Music Server PC is not destroying the audio ring, also not with its connected monitor.

Summarized :

Audio Ring
- Audio PC as lean as possible and without monitor
- D/A Converter (connected to Audio PC)
- Power Amps (preferably no pre-amp)
- (connection to Music Server PC via direct LAN cable)

House Ring
- Music Server PC
- Monitor
- When wanted, external disks with switching supplies
- (Connection to Audio PC via direct LAN cable)
- Router, Modem, etc.

Peter

[acg]

Mind you, while the two PC's (Music Server PC and Audio PC) are connected by means of Ethernet, this is a galvanically isolated connection.

Hi Peter,

I do wonder just how effective the standard ethernet connection is at separating potentials.  In medicine "Network Isolators" are used to guarantee that a voltage potential does not build up between medical and non-medical components.   From Wikipedia:

A specific hazard stems from possible differences in ground potentials between network components, which when not properly isolated can result in a leakage current that can flow through the patient, which is dangerous, and potentially lethal. Such voltage differences can also arise through incorrect installation and wiring of network systems, electrical shorting within damaged cables and cabling, or shorting between damaged network cables and other voltage sources.

Network isolators work to remove this hazard, by electrically disconnecting medical devices from a network. Isolators may be used as network accessories, built into medical devices, or installed within a medical network system. Networks isolators should be installed as close as possible to the medical device in question.

An example of such a connector is here.  Can you see any application for these in our specific situation?  Perhaps the the next step of isolating the AudioPC entirely from the outside world?

Regards,

Anthony

[PeterSt]

Hi Anthony,

Doesn't seem wrong to me at all.

Maybe we must "investigate" what the leakage atually is. I mean, I only copy words from others when I "claim" that Ethernet is isolated (knowing that transformers exist at the end points). But for how much current ? really no idea (easy to find out I suppose).
I also don't think that Ethernet was invented for its isolation behavior. It just is so (or ?). And it thus will also be "to some degree" I'd say.

Perhaps the the next step of isolating the AudioPC entirely from the outside world?

Isn't it so that a linear PSU just does that ? (the remainder being isolated already, lately )

Regards,
Peter

[manisandher]

Yep, I've been using the Baaske ethernet isolators for years now. Not sure how one would measure their effectiveness though, other than trust the manufacturer.

In the Orelo MkII thread Peter asked:

What was going on that you had that audible noise going on (while it shouldn't) ? I mean, apparently you maybe thought I was kidding or whatever else that made you not believe me, and therefore you didn't attack the problem in the first place (never mentioned it either) ? I now mean : it is quite crucial for me to learn what it was that created this noise. For a next one and helping out ...

Well, I was using a PurePower P2000 'AC regenerator' to power the whole system. This is actually just a glorified UPS that some (e.g. 'Romy the Cat') say imparts an interesting affect on the SQ. I've never been particularly impressed with my unit, which I've had for, I suppose, 5 years or so now. And certainly the Orelo MkIIs are simply too transparent for such a unit - you hear all the cr*p that it emits from its AC outputs. And god only knows what it kicks back out into the AC mains that feeds it.

I now have the following AC setup:

PC System

house AC mains -> PP2000 UPS -> music server -> Baaske isolator -> audio PC

Audio System

dedicated low impedance AC mains line -> 15KVA isolation transformer (with dedicated earth bonded to Neutral output) -> PS Audio P10 AC regenerator -> NOS1a (with Intona USB isolator to audio PC) & Orelo speakers

This setup is totally overkill. It really is. And the ONLY reason I'm doing this is because I spent a fortune on my 15KVA isolation transformer and dedicated earth. If I hadn't, I would not be doing this.

But Peter, I have a question for you: why have the audio PC on the audio ring? Why not have it on the music server ring? The Intona gives galvanic isolation, and now no need to worry about using a linear PSU for the audio PC.

Mani.

[acg]

But Peter, I have a question for you: why have the audio PC on the audio ring? Why not have it on the music server ring? The Intona gives galvanic isolation, and now no need to worry about using a linear PSU for the audio PC.

Hi Mani,

You are the guy that can test this by reinstalling the standard ATX SMPS into your AudioPC.  I can't test it just yet but wheels are in motion for my ATX LPS...I have a custom case being cut this week in which to stuff it all and once that is done I will be able to test the ATX Controller PCB and then all manner of LPS scenarios.

But in the meantime, if you had some spare time, perhaps see if your AudioPC less LPS still performs at the same level.

Cheers,

Anthony

[PeterSt]

But Peter, I have a question for you: why have the audio PC on the audio ring? Why not have it on the music server ring? The Intona gives galvanic isolation, and now no need to worry about using a linear PSU for the audio PC.

Did you ever witness that bloke watching some gag, holding his breath for one tad too long, after that squeezing out the air with a blast ?



that was me a minute back.

Mani, I guess you are right. I just didn't think about that ...

Peter

[PeterSt]

Pfff ... I'm still thinking in a flabbergasted fashion what we could be overlooking. But I don't think there is anything ...

That plastic box could be more golden than we imagined.

[acg]

Maybe we must "investigate" what the leakage atually is. I mean, I only copy words from others when I "claim" that Ethernet is isolated (knowing that transformers exist at the end points). But for how much current ? really no idea (easy to find out I suppose).
I also don't think that Ethernet was invented for its isolation behavior. It just is so (or ?). And it thus will also be "to some degree" I'd say.

If we are in the scenario that you suggest (and I am in at the moment) where I have a dedicated low impedance audio only mains line that is separate to the rest of the house, and we are running network cable from those two/three or more reasonably unique lines there could quite easily I imagine become a significant difference in ground potential across the length of an ethernet cable from the Music Server to the AudioPC.

So if we have USB properly isolated...ethernet properly isolated...power properly isolated (LPS) then there is not really too much more isolation work to do other than obsessive compulsive amongst us to attack RFI.

At least that is how I am thinking.  I do not know enough to have any definitive answers.

[PeterSt]

Ok, when we are in the mood anyway ...

I forgot the exact reason, but many years ago I bought some line level isolators (there's a transformer in there). They go in between DAC and Amplifier(s). Coincidentally it was 2 days ago or so that I thought about those again.
Back then I rejected them because they measured not linearly, but God knows what was going on back then, including my measurement skills.

Thus, these things exist;
Apparently I am not good at this (soo obvious) stuff, so maybe others can think of a usecase for them ?

Peter

[PeterSt]

So if we have USB properly isolated...ethernet properly isolated...power properly isolated (LPS) then there is not really too much more isolation work to do other than obsessive compulsive amongst us to attack RFI.

Anthony, the more time it is to now attack EMI !
So I am sure it is high time to get myself a decent meter for that. We can suspect EMI to be all over the place with all this isolation stuff and it should be relatively easy to now attack *that*. But first we must see it ...

Regards,
Peter

[manisandher]

Did you ever witness that bloke watching some gag, holding his breath for one tad too long, after that squeezing out the air with a blast ?



that was me a minute back.

I tried to imagine this, but it got too scary!

Mani.

[manisandher]

You are the guy that can test this by reinstalling the standard ATX SMPS into your AudioPC.  I can't test it just yet but wheels are in motion for my ATX LPS...I have a custom case being cut this week in which to stuff it all and once that is done I will be able to test the ATX Controller PCB and then all manner of LPS scenarios.

But in the meantime, if you had some spare time, perhaps see if your AudioPC less LPS still performs at the same level.

Hi Anthony, yeah I could do this... and I will when I have time. Certainly with my current setup (both music and audio PCs on same 'dirty' line) I simply don't see the point in having an expensive linear PSU for the audio PC.

But what I will try first is putting the audio PC (with linear PSU) on the 'clean' dedicated audio line. I can't imagine how this could improve the SQ... but I can see how it might make it worse - no matter how good the linear PSU (and I think the Teradak is pretty good), it's still a PC that you've introduced onto the 'clean' line.

Mani.