Unbalanced shielding questions

[Colin Bigam]

Hey all;

What started as a thread over on rec.audio.tech got me wondering about
how shielding works in different configurations.

Basically, in an unbalanced connection (remember, the original discussion
came from consumer audio), you typically see the following:

1) single shielded conductor (not usually coax, but could be)
2) unshielded twisted pair
or...
3) shielded twisted pair, with the shield connected at (only) one end
of the ground return

Now the question is, how do all of thse work, and how WELL do they
work?

For example, if shielded pairs work by inducing the same (stray) voltage
equally in both of the (opposite) conductors and thus cancelling it out,
then can shielding them accomplish anything at all?

Another ponderable: Why tie the shield to the conductor at all?

Hope someone can answer these for me. My curiosity is burning. BURNING,
I tell you!

Thanks,
Colin

 

[N. Thornton]

1) single shielded conductor (not usually coax, but could be)
2) unshielded twisted pair
or...
3) shielded twisted pair, with the shield connected at (only) one end
of the ground return

Now the question is, how do all of thse work, and how WELL do they
work?

For example, if shielded pairs work by inducing the same (stray) voltage
equally in both of the (opposite) conductors and thus cancelling it out,
then can shielding them accomplish anything at all?

twisted pair is fairly effective using a differential input, but it
isnt perfect, since each conductor is at a different distance to any
noise source, so both L and C coupling arent perfectly matched. Also
CMRR isnt perfect at the following input.

Screening the twisted pair resolves these 2 matters.

Hope someone can answer these for me. My curiosity is burning. BURNING,
I tell you!

I've noticed. For some reason I'm just not with you on that.

Regards, NT

 

[Steve]

Now the question is, how do all of thse work, and how WELL do they
work?

They all work the same way, google for more info on shielding. The
effectiveness of a shield depends more on the quality of the braiding
than how it is connected.

For example, if shielded pairs work by inducing the same (stray) voltage
equally in both of the (opposite) conductors and thus cancelling it out,
then can shielding them accomplish anything at all?

I thought you were talking UNbalanced. You are describing a balanced
system.

Another ponderable: Why tie the shield to the conductor at all?

Huh? You don't.

 

[Walter Harley]

Hey all;

What started as a thread over on rec.audio.tech got me wondering about
how shielding works in different configurations.

Basically, in an unbalanced connection (remember, the original discussion
came from consumer audio), you typically see the following:

1) single shielded conductor (not usually coax, but could be)
2) unshielded twisted pair
or...
3) shielded twisted pair, with the shield connected at (only) one end
of the ground return

Now the question is, how do all of thse work, and how WELL do they
work?

Google 'Morrison' in this newsgroup to see a recent thread with a discussion
of some references on how EMI reduction techniques work.

As to how well they work, you'll probably not get any better info here than
on r.a.t - the problem is that there are so many variables in measuring EMI.
Picking the signal out of the noise in opinions about EMI is a hard problem

For example, if shielded pairs work by inducing the same (stray) voltage
equally in both of the (opposite) conductors and thus cancelling it out,
then can shielding them accomplish anything at all?

No. But that's not quite how pairs work (even in balanced signal
transmission, which you're not talking about...). Even if it were,
"equally" and "cancelling" imply some kind of ideal process, and we're
talking about the real world.

For instance, if the common mode signal to a differential amp makes the
input stages clip, then it can't effectively reject it any more. In that
case, shielding the signal could help, by allowing the differential amp to
do its thing.

Another ponderable: Why tie the shield to the conductor at all?

Intuitively: so that the induced energy has somewhere to go. It's going to
go SOMEWHERE; you want it to go somewhere safe, like "ground".

 

[Ben Bradley]

Hey all;

What started as a thread over on rec.audio.tech got me wondering about
how shielding works in different configurations.

Basically, in an unbalanced connection (remember, the original discussion
came from consumer audio), you typically see the following:

1) single shielded conductor (not usually coax, but could be)
2) unshielded twisted pair
or...
3) shielded twisted pair, with the shield connected at (only) one end
of the ground return

Now the question is, how do all of thse work, and how WELL do they
work?

Maybe these pages will answer some of your questions (and other
questions you didn't know to ask):

http://www.rane.com/note110.html

http://www.jensen-transformers.com/apps_wp.html (especially AN-003 and
AN-004)

 

[Mike Page]

Hey all;

What started as a thread over on rec.audio.tech got me wondering about
how shielding works in different configurations.

Hope someone can answer these for me. My curiosity is burning. BURNING,
I tell you!

Thanks,
Colin

Jensen Transformers have some great app notes about cabling.

 

[[email protected]]

Google 'Morrison' in this newsgroup to see a recent thread with a discussion
of some references on how EMI reduction techniques work.

Ah hah! Don't know how I missed this before. Thanks.

As to how well they work, you'll probably not get any better info here than
on r.a.t - the problem is that there are so many variables in measuring EMI.
Picking the signal out of the noise in opinions about EMI is a hard problem

Well the thin is that asking questions that aren't applicable to audio
in r.a.t. often gets a reply of "it's not applicable!"

No. But that's not quite how pairs work (even in balanced signal
transmission, which you're not talking about...). Even if it were,
"equally" and "cancelling" imply some kind of ideal process, and we're
talking about the real world.

Ah! Then if I understand, this can also explain why it's important to
dump any induced voltage in a shield, rather than letting it float around
in there. If it's in the shield and not drained, then it gets induced
in the conductors, and once again you can overload the input stage. Is
this right?

Anyways, thanks for the info. Might have to hunt down a copy of Morrison
now...

COlin

 

[[email protected]]

I've noticed. For some reason I'm just not with you on that.

Oh, that's easy: Because I don't know!

Colin

 

[Walter Harley]

Ah! Then if I understand, this can also explain why it's important to
dump any induced voltage in a shield, rather than letting it float around
in there. If it's in the shield and not drained, then it gets induced
in the conductors, and once again you can overload the input stage. Is
this right?

Well, that's the basic idea.

The shield itself isn't perfect either; it's got finite inductance, which
means that for high-frequency signals, there can still be substantial
voltage across it even if one end is grounded. That's the main argument for
grounding shields at both ends, by the way. For that matter, "ground"
itself is not perfect; when the noise on the shield is dumped into the
ground wire, it will cause current in the ground wire, and the ground wire
has nonzero resistance and inductance, so there will be a voltage across it.
Which means that you have to decide which end you want to call "ground".

Dealing with EMI is an interesting problem, because it's real-world
engineering at its most real. The laws of physics all apply; but the
convenient generalizations and simplifications that we sometimes make - like
thinking of wires as ideal conductors - break down.

Shielding against high-frequency noise is a different problem than shielding
against low-frequency noise. What defines "high frequency" is the
wavelength of the signal, compared to the length of the wires you're using,
and eventually compared to things like the size of the little holes in the
shield braid or the gap in the shield foil.

However, even a seemingly "low frequency" source like a light dimmer
actually has very high frequency components. The wave is getting switched
on and off at only 120Hz or so, but when it gets switched, the voltage
changes very abruptly; so this isn't a 120Hz signal, it's a high-frequency
burst that just happens to occur 120 times a second.

Also, shielding against voltage fields (like that from a flourescent light)
is different than shielding against magnetic fields (like that from a nearby
motor or power transformer).

 

[R.Legg]

Hey all;

What started as a thread over on rec.audio.tech got me wondering about
how shielding works in different configurations.

Basically, in an unbalanced connection (remember, the original discussion
came from consumer audio), you typically see the following:

1) single shielded conductor (not usually coax, but could be)
2) unshielded twisted pair
or...
3) shielded twisted pair, with the shield connected at (only) one end
of the ground return

Now the question is, how do all of thse work, and how WELL do they
work?

http://emcesd.com/
http://emcesd.com/pdf/roma94.pdf

RL