SMPS and grounding

[pl_svn]

 

Alex... the MeanWell SMPS I'm using to power an LPS-1 comes from Regen's very first batch: in doubt if those are already grounded or not   ... I  just finished grounding its negative using John's method and all sounds good so far  

 

too early to say (and maybe totally unrelated) but looks those tiny EMI/RFI "noises" my tubes system was suffering now and then are gone 

(though, actually, they always go whenever I disconnect/reconnect something in the chain then... eventually come back later on  )

[BigGuy]

1 hour ago, BigGuy said:

Definitely appreciate the patience shown by both John and Alex in answering our questions.

 

I too will be looking to purchase a Netgear switch, e.g., Fs105,  to insert between my router and PC but wonder how one knows if the switch is V5 or higher per John's recent post particularly buying onnline?

Duh!  Its not V(er)5. 

[BigGuy]

Re John's comment re crosstalk between STP vs UTP ethernet cables, I had gotten this info from Blue Jeans Cable some time ago and decided against STP...

 

"The crosstalk would be between cables if you are running bundles of them over long distances.  This article (written by the owner of BJC) explains shielding in ethernet cabling in detail:

 

Shielding is, in most cases, not of much use on Ethernet cable.  The noise rejection characteristics of the pairs (common-mode noise rejection) are the most important consideration whether the cable is shielded or not, and there are all sorts of practical issues -- foil isn't a very low-resistance path to ground, and noise currents on the shield can sometimes hurt as much as they help. 

 

Performance-wise, the shield causes problems.  For one thing, the shield-to-pair spacing is hard to keep consistent, and it affects the pair impedance slightly; for another thing, a lot of the energy at these frequencies bounces off of shields, so internal crosstalk suffers for the benefit of alien crosstalk; and for a third thing, the nature of four pairs in a shielded bundle is such that the symmetry of the pairs is effectively disrupted by a shield which each conductor approaches, and then recedes from, with each twist (the other pairs don't affect this symmetry as profoundly because of the effect of different twist rates). 

The result is that if you take two patch cords: our Cat 6, and our Cat 6a, and test both of them both at Cat 6 and at Cat 6a standards, you find that both of them pass both standards, but the Cat 6 passes by larger margins than the 6a due to the effects of the shielding.  One might suppose that this means that the 6a should never be used -- but the difference is that in order to be certified the bulk cable must meet tests as well as the finished assembly, and Alien Crosstalk is tested only on bulk cable -- the Cat 6 would fail AXT, while the 6a passes it.  It's a design tradeoff -- deterioration in cable internal electricals in order to achieve external targets.

 

Now, that has some interesting implications in a home or small network environment; if the patch cables are deployed at work stations, rather than on patch panels, then they typically do not run close to one another in bundles.  Regardless of what the spec says, the Cat 6 patch would be better than the 6a in such a case, because the actual installation conditions make

alien crosstalk no longer a meaningful consideration.

 

But, more importantly, in most American network installations the backbone cable is unshielded.  What this means is that use of shielded cables in these systems can result in partial shield paths -- which can cause ground-loop problems and the like that can be rather hard to diagnose.  The best practice is, if the installed horizontal cabling is unshielded, to use unshielded patch cords, and correspondingly, if the installed horizontal cabling is shielded, to use shielded patch cords.  Either the entire system should be shielded, with shields tied to ground at patch panels, or none of it should be grounded.  This is one of those "how to tell the customers" conundrums -- we need, when we do make a shielded product available, to be sure that people understand that it really should only be used in particular environments, because the assumption tends to be "shielding = good" and this really does depend on the installation."

 

[JohnSwenson]

1 hour ago, BigGuy said:

Re John's comment re crosstalk between STP vs UTP ethernet cables, I had gotten this info from Blue Jeans Cable some time ago and decided against STP...

 

"The crosstalk would be between cables if you are running bundles of them over long distances.  This article (written by the owner of BJC) explains shielding in ethernet cabling in detail:

 

Shielding is, in most cases, not of much use on Ethernet cable.  The noise rejection characteristics of the pairs (common-mode noise rejection) are the most important consideration whether the cable is shielded or not, and there are all sorts of practical issues -- foil isn't a very low-resistance path to ground, and noise currents on the shield can sometimes hurt as much as they help. 

 

Performance-wise, the shield causes problems.  For one thing, the shield-to-pair spacing is hard to keep consistent, and it affects the pair impedance slightly; for another thing, a lot of the energy at these frequencies bounces off of shields, so internal crosstalk suffers for the benefit of alien crosstalk; and for a third thing, the nature of four pairs in a shielded bundle is such that the symmetry of the pairs is effectively disrupted by a shield which each conductor approaches, and then recedes from, with each twist (the other pairs don't affect this symmetry as profoundly because of the effect of different twist rates). 

The result is that if you take two patch cords: our Cat 6, and our Cat 6a, and test both of them both at Cat 6 and at Cat 6a standards, you find that both of them pass both standards, but the Cat 6 passes by larger margins than the 6a due to the effects of the shielding.  One might suppose that this means that the 6a should never be used -- but the difference is that in order to be certified the bulk cable must meet tests as well as the finished assembly, and Alien Crosstalk is tested only on bulk cable -- the Cat 6 would fail AXT, while the 6a passes it.  It's a design tradeoff -- deterioration in cable internal electricals in order to achieve external targets.

 

Now, that has some interesting implications in a home or small network environment; if the patch cables are deployed at work stations, rather than on patch panels, then they typically do not run close to one another in bundles.  Regardless of what the spec says, the Cat 6 patch would be better than the 6a in such a case, because the actual installation conditions make

alien crosstalk no longer a meaningful consideration.

 

But, more importantly, in most American network installations the backbone cable is unshielded.  What this means is that use of shielded cables in these systems can result in partial shield paths -- which can cause ground-loop problems and the like that can be rather hard to diagnose.  The best practice is, if the installed horizontal cabling is unshielded, to use unshielded patch cords, and correspondingly, if the installed horizontal cabling is shielded, to use shielded patch cords.  Either the entire system should be shielded, with shields tied to ground at patch panels, or none of it should be grounded.  This is one of those "how to tell the customers" conundrums -- we need, when we do make a shielded product available, to be sure that people understand that it really should only be used in particular environments, because the assumption tends to be "shielding = good" and this really does depend on the installation."

 

This and every other article on Ethernet shielding is talking about the differential signal between the two wires in the pair. Leakage is common mode, it is the same on both wires (NOT differential) thus the twistedness does not matter. Because it is common mode noise the twisted pair just looks like a weirdly shaped single conductor and makes a nice antenna.

 

I do NOT recommend shielded Ethernet in most cases, leakage will go through the shield, bypassing all things I have written about blocking it, none of that works if you have connected shields.

 

That is why I recommend using UTP and putting one of the grounded named swithces some distance away from the audio system, there will be nothing on the cable to radiate.

 

John S.

[JohnSwenson]

2 hours ago, BigGuy said:

Duh!  Its not V(er)5. 

FS105 goes up to V3, GS105 goes up to V5. The highest versions are the ones that almost everybody is selling. (FS105V3, GS105V5).

 

If you buy one used on Ebay you might be getting an earlier one, but getting a new one from a normal online channel (Amazon, Newegg etc) will have a very high probability of getting the latest version.

 

John S.

[BigGuy]

13 minutes ago, JohnSwenson said:

This and every other article on Ethernet shielding is talking about the differential signal between the two wires in the pair. Leakage is common mode, it is the same on both wires (NOT differential) thus the twistedness does not matter. Because it is common mode noise the twisted pair just looks like a weirdly shaped single conductor and makes a nice antenna.

 

I do NOT recommend shielded Ethernet in most cases, leakage will go through the shield, bypassing all things I have written about blocking it, none of that works if you have connected shields.

 

That is why I recommend using UTP and putting one of the grounded named swithces some distance away from the audio system, there will be nothing on the cable to radiate.

 

John S.

Sorry for my misunderstanding,  FS105 arriving tomorrow hopefully V3.

[Monge]

I googled Netgear GS108 v5

Ihttp://www.downloads.netgear.com/files/GDC/datasheet/en/GS105v5-GS108v4.pdf

Mark the words In the tittle of the pdf file.

i have a Netgear GS108GE v4 here. The 230V/12V/1A EU version.

Hopefully I Can use this as well.

[rickca]

3 hours ago, JohnSwenson said:

I do NOT recommend shielded Ethernet in most cases, leakage will go through the shield, bypassing all things I have written about blocking it, none of that works if you have connected shields.

Is BJC Cat 6A OK to use?  It's supposed to have a floating shield.

[Theobetley]

Just additional parts (2.5 mm) to ground my other docking station and it sounds wonderful. Thanks again John. 

[lmitche]

33 minutes ago, rickca said:

Is BJC Cat 6A OK to use?  It's supposed to have a floating shield.

I have lots of belden based bjc cat 6a cables and having re-terminated several, know that the cable is shielded but floating on both ends.

 

The internal structure of this cable is really something. I have no idea how they can cheaply manufacture something with this complexity.

[BigGuy]

17 minutes ago, lmitche said:

I have lots of belden based bjc cat 6a cables and having re-terminated several, know that the cable is shielded but floating on both ends.

 

The internal structure of this cable is really something. I have no idea how they can cheaply manufacture something with this complexity.

Interesting re the bjc cat 6a cable since when I just inquired about STP ethernet, I was told they do not offer.

Is there a difference between STP and their 6a?

[rickca]

18 minutes ago, lmitche said:

I have lots of belden based bjc cat 6a cables and having re-terminated several, know that the cable is shielded but floating on both ends.

 

The internal structure of this cable is really something. I have no idea how they can cheaply manufacture something with this complexity.

So BJC Cat 6A is OK meaning it does not defeat John's ground shunt because it does not have connected shields, right?  I was concerned when I read his comment

I do NOT recommend shielded Ethernet in most cases, leakage will go through the shield, bypassing all things I have written about blocking it, none of that works if you have connected shields.

[lmitche]

Yes, unconnected shield on bjc cat 6a cables.

[JohnSwenson]

3 hours ago, Monge said:

I googled Netgear GS108 v5

Ihttp://www.downloads.netgear.com/files/GDC/datasheet/en/GS105v5-GS108v4.pdf

Mark the words In the tittle of the pdf file.

i have a Netgear GS108GE v4 here. The 230V/12V/1A EU version.

Hopefully I Can use this as well.

OK so it seems the GS105V5 and the GS108V4 are in the same series, which PROBABLY means they will have the same behavior.

 

John S.

[R1200CL]

8 minutes ago, JohnSwenson said:

OK so it seems the GS105V5 and the GS108V4 are in the same series, which PROBABLY means they will have the same behavior.

 

John S.

 

And can you make a guess about the E (managed) version ?

https://www.netgear.com/support/product/GS105Ev2.aspx#docs

 

The possibility to switch of ports and even lower the speed should be a good extra positive option. 

[thyname]

How do we know whether the GS version is v4 or v5? Can’t tell from Amazon page

[JohnSwenson]

The GS105V2 is old enough that I can't make a guess as to whether it will block external leakage when grounded. I have tested current and previous FS and current GS, these have the proper behavior.

 

That is not enough of a sample to guess well on significantly older versions. I have seen radically different behavior between two versions from other companies, this makes me leery about common behavior between a V5 and a V2.

 

John S.

[JohnSwenson]

4 minutes ago, thyname said:

How do we know whether the GS version is v4 or v5? Can’t tell from Amazon page

You can ask the question, you might get a reply from the seller.

 

Amazon has a fairly high volume so there is a fairly high probability you will be getting a "latest" version .

 

John S.

[thyname]

And once we get it at hand, how do we check which version?

[thyname]

A couple of (somewhat naive) questions:

 

1 - I know Jason’s ground plug and the SMPS need to be plugged in the same power strip, but can we have other (unrelated) devices plugged on the same power strip?

 

2 - Can we have other devices hooked up on the same Ethernet Switch when grounding the Switch? Or just the Streamer in question?

 

it is really great news that only the switch need to be grounded! Saves me from having my setup look like a high school science experiment if I had to ground all the various devices.

 

thanks!

[lmitche]

12 minutes ago, thyname said:

And once we get it at hand, how do we check which version?

There is a label with the version number on the bottom of the device.

[R1200CL]

If you buy the PoE version of that GS105, you don’t even need a PS ?

 

All problems solved ?

 

(no one has said the PoE is a no go YET)

 

And then a MicroRendu or UltraRendu driven by PoE......

 

Alex once said it will be fairly easy to make a 12 V LPS-1. Only reprogram the FPGA ?

 

Could we hope for a 12VDC version to all the people now buying 12 V switches ?

[sandyk]

2 hours ago, R1200CL said:

Could we hope for a 12VDC version to all the people now buying 12 V switches ?

 

From a different Alex.

The answer is likely to be no, due mainly to the low voltage ratings of Ultracaps.

You would need more of them in series in  each parallel group meaning that the PCB would need to be considerably larger as well.

This would also mean that a larger case would be required.

 

SandyK

 

[kboung]

Regarding combination of iso-regen powered by LPS-1

 

(1) For "tick/dropout" issue with iso-regen, it was recommended to ground downstream of iso-regen

(2) For high impedance leakage with LPS-1, it was recommended to ground Meanwell output negative

 

In this case, both groundings required or (1) will actually cover for (2)?

 

Maybe Alex or John can clarify. Thanks.

[BigGuy]

12 hours ago, lmitche said:

Yes, unconnected shield on bjc cat 6a cables.

IIRC, there was a much earlier post about the need to have the two ends of the shield connected to each other by an insulated wire otherwise the unconnected shield is just metallic sheathing.  Or does it serve a purpose even unconnected?