Wiring unbalanced connectors with balanced cable
Gareth Magennis
Hi,
just curious as to what is the best practice: If you have an unbalanced
connection, say RCA to RCA or jack to jack, and use balanced cable, what is
the best way to wire it? Parallel both cores, only connect one core with
the other floating, or ground the unused core?
I'm guessing that for most applications there will be little difference, but
for something like an RIAA preamp the capacitance/impedance/noise
differences between connection methods may be rather more significant.
Cheers,
Gareth.
Scott Dorsey
>
>just curious as to what is the best practice: If you have an unbalanced
>connection, say RCA to RCA or jack to jack, and use balanced cable, what is
>the best way to wire it? Parallel both cores, only connect one core with
>the other floating, or ground the unused core?
One wire goes from tip to tip, the other wire goes from sleeve to sleeve.
Connect the shield to sleeve on one side only, not both.
>I'm guessing that for most applications there will be little difference, but
>for something like an RIAA preamp the capacitance/impedance/noise
>differences between connection methods may be rather more significant.
For a phono preamp, you really want to use special low-capacitance cable
and not generic twisted pair.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."
Arkansan Raider
See how easy that was, Gareth?
<g>
---Jeff
Gareth Magennis
"Arkansan Raider" <yom...@yomama.com> wrote in message
news:hc7b83$uss$3...@news.eternal-september.org...
Yep, a far better experience, a lot less CAPITALS.
Cheers,
Gareth.
William Sommerwerck
> sleeve. Connect the shield to sleeve on one side only, not both.
Agreed. The end where the shield is wired to the sleeve should go to the
input.
Gareth Magennis
"Scott Dorsey" <klu...@panix.com> wrote in message
news:hc7984$4cu$1...@panix2.panix.com...
> Gareth Magennis <sound....@btconnect.com> wrote:
>>
>>just curious as to what is the best practice: If you have an unbalanced
>>connection, say RCA to RCA or jack to jack, and use balanced cable, what
>>is
>>the best way to wire it? Parallel both cores, only connect one core with
>>the other floating, or ground the unused core?
>
> One wire goes from tip to tip, the other wire goes from sleeve to sleeve.
> Connect the shield to sleeve on one side only, not both.
Would that actually be any improvement over unbalanced cable?
Thanks,
Gareth.
William Sommerwerck
>> sleeve. Connect the shield to sleeve on one side only, not both.
> Would that actually be any improvement over unbalanced cable?
In theory, yes, because the shield no longer carries the signal.
Scott Dorsey
>>>
>>>just curious as to what is the best practice: If you have an unbalanced
>>>connection, say RCA to RCA or jack to jack, and use balanced cable, what
>>>is
>>>the best way to wire it? Parallel both cores, only connect one core with
>>>the other floating, or ground the unused core?
>>
>> One wire goes from tip to tip, the other wire goes from sleeve to sleeve.
>> Connect the shield to sleeve on one side only, not both.
>
>Would that actually be any improvement over unbalanced cable?
No. But if you have big rolls of twisted pair and you don't have any
single-conductor lying around, you might as well. Plus maybe one day you'll
want to cut one end off and put an XLR on.
Mark
> Gareth Magennis <sound.serv...@btconnect.com> wrote:
> >"Scott Dorsey" <klu...@panix.com> wrote in message
>
> >>>just curious as to what is the best practice: If you have an unbalanced
> >>>connection, say RCA to RCA or jack to jack, and use balanced cable, what
> >>>is
> >>>the best way to wire it? Parallel both cores, only connect one core with
> >>>the other floating, or ground the unused core?
>
> >> One wire goes from tip to tip, the other wire goes from sleeve to sleeve.
> >> Connect the shield to sleeve on one side only, not both.
>
> >Would that actually be any improvement over unbalanced cable?
>
> No. But if you have big rolls of twisted pair and you don't have any
> single-conductor lying around, you might as well. Plus maybe one day you'll
> want to cut one end off and put an XLR on.
> --scott
>
> --
> "C'est un Nagra. C'est suisse, et tres, tres precis."
OP
you will get better RF shielding if you connect the shield to the
shell on BOTH ends.
Otherwise I agree, for an RIAA pre-amp the capciantce is critical and
you should use cable /length that provides the correct C that the
cart. is expecting to see.
Mark
Scott Dorsey
>
>OP
>you will get better RF shielding if you connect the shield to the
>shell on BOTH ends.
Not really, Telescoping is actually something of a win. Check out
Henry Ott's book on grounding for a really nice discussion.
Mark
a shield open at one end will do a perfect job as an electoSTATIC
shield protecting against high impedance low frequency sources, like
high voltages at 60 Hz for example, but a shield open at one end will
do nothing against electroMAGNETIC waves like RF signals at 100 MHz
for example.
Mark
Scott Dorsey
>a shield open at one end will do a perfect job as an electoSTATIC
>shield protecting against high impedance low frequency sources, like
>high voltages at 60 Hz for example, but a shield open at one end will
>do nothing against electroMAGNETIC waves like RF signals at 100 MHz
>for example.
Try the other way around.
ALL foil and braid shields deal only with electric fields. They
do not protect against the magnetic field in any way.
Electromagnetic waves consist of both an electric and a magnetic field,
at 90 degrees. That's why they call them electromagnetic waves. To
block RF, you can either block the B field or the E field, and needless
to say it's a lot easier to block the E field with a shield.
All this stuff is nicely discussed in Henry Ott's book on grounding.
Don Pearce
Of course by shorting down the E field, you deal simultaneously with
the associated B field. It is the B field on its own - from
transformers and the like - that can't be handled by a braid or foil
shield. You either need a mu metal magnetic screen, or minimized net
loop area, as you get from twisted pair.
d
Mark
>
> - Show quoted text -
I'm not talking about a low frequency 60 Hz magnetic field, yes you
need mu metal for that.
I'm talking about a plane EM wave at 100 MHz in the far field where
the wave impedance is that of free space 377 Ohms. More commonly
known as RFI.
A shield not connected at one end is NOT effective against RFI at 100
MHz.
There are many things I don't know and some things I do know, and that
is one of the things I DO know.
Mark
Don Pearce
wrote:
You are dead right. A chunk of coax with only one end grounded will
make a pretty decent antenna at 100MHz. It will shield nothing.
d
Scott Dorsey
>
>Of course by shorting down the E field, you deal simultaneously with
>the associated B field.
Right. In the case of RF that's sufficient.
>It is the B field on its own - from
>transformers and the like - that can't be handled by a braid or foil
>shield. You either need a mu metal magnetic screen, or minimized net
>loop area, as you get from twisted pair.
Right, and that's why balanced lines are such a big win. Well, it's one
of the two big reasons anyway.
And steel conduit isn't too bad either. Permittivity isn't as good as
mu metal but it can be a lifesaver in permanent installs.
Scott Dorsey
>I'm not talking about a low frequency 60 Hz magnetic field, yes you
>need mu metal for that.
>
>I'm talking about a plane EM wave at 100 MHz in the far field where
>the wave impedance is that of free space 377 Ohms. More commonly
>known as RFI.
>
>A shield not connected at one end is NOT effective against RFI at 100
>MHz.
>
>There are many things I don't know and some things I do know, and that
>is one of the things I DO know.
Hint: the two grounds are kept equipotential by the return wire.
Try it. Or check Ott's book.
Mark
I "try it" every time I get a defective coax cable in the lab and it
radiates all over the place....
It's like a "farady shield" with one side cut open...
Any current induced on the outside of the shield by an external field
will flow at the gap into the inside of the shield and then be induced
onto the inner conductors.
I know I have read about the open shield being effective in
___audio___ books and it may be effecitve in certain special cases
against a particualr type of interference (Hi Z low frequency), but in
general it is NOT effective at shielding RF.
I am not a professional audio guy and I come here to learn about
audio but I AM a VERY experinced professional RF guy and I can assure
you, a shield disconnected at one end is not effective agaisnt RF.
hint: "slot antenna"
Mark
Mark
Scott Dorsey
>On Oct 29, 2:02=A0pm, klu...@panix.com (Scott Dorsey) wrote:
>>
>> Hint: the two grounds are kept equipotential by the return wire.
>>
>
>I "try it" every time I get a defective coax cable in the lab and it
>radiates all over the place....
>
>It's like a "farady shield" with one side cut open...
No, it's not. The two grounds are kept equipotential. The shield is
complete, it's just joined only at one end so no current flows.
But UNLIKE a coax cable, the two ends ARE joined together, by the
return conductor.
Think about it this way...
If you have a magnetic field, the return wire and the signal wire both
have the same signal induced in them, since they are part of a twisted
pair. Consequently the signal is more or less cancelled assuming the
receiving device is completely floating.
If you have an electric field, it's blocked by the Faraday shield, which
goes the length of the cable and IS an effective shield because it is
bonded to ground at one point.
If you bonded the shield on BOTH ends, you'd be creating a small but
nonzero-area ground loop because the shield and the return conductor
don't occupy precisely the same place in space. If you omitted the
return conductor, you would lose the small but measurable benefit
against magnetic induction mentioned above there.
Remember, we're talking about DC continuity between pin to pin and
between sleeve to sleeve... it's not like the sleeves have become
disconnected like in a bad coax cable. They are bonded together
by the return conductor in the pair. But, and this is a huge but,
no return current flows in the shield. The shield acts _only_ as
a shield.
>Any current induced on the outside of the shield by an external field
>will flow at the gap into the inside of the shield and then be induced
>onto the inner conductors.
What gap? The thickness of the jacket inside the connector? Yes,
that's a break in the shield but it's no worse than the breaks already
resulting from the braid not having 100% coverage.
>I know I have read about the open shield being effective in
>___audio___ books and it may be effecitve in certain special cases
>against a particualr type of interference (Hi Z low frequency), but in
>general it is NOT effective at shielding RF.
It is actually a minor win for any sort of unbalanced signals, and
I first encountered it in the manuals for the Grass EEG amplifiers.
Honest, you _need_ to read Henry Ott on the subject.
The shield is still complete, it's just only bonded at one end.
>I am not a professional audio guy and I come here to learn about
>audio but I AM a VERY experinced professional RF guy and I can assure
>you, a shield disconnected at one end is not effective agaisnt RF.
>
>hint: "slot antenna"
It is indeed a tiny slot antenna, and that might be an issue in the millimeter
wave region. However, the benefit that you get from the mild rejection of
induced magnetic stuff is pretty good and more than makes up for it.
Honestly, you need to check out Henry Ott's book on grounding.
Ficchi's book "Electrical Interference" also talks a bit about the
technique. There's also a really good book on instrumentation
electronics whose author escapes me right now but I'll find it when
I am back in the office.
It's one of those goofy things that instrumentation engineers see a lot
of and which not too many other people do. RF guys don't care about any
of that stuff... they have easy ways to break grounds and they don't
get to use twinax much anyway....
Meindert Sprang
news:e819f8c2-1470-41b4...@a31g2000yqn.googlegroups.com...
> I am not a professional audio guy and I come here to learn about
> audio but I AM a VERY experinced professional RF guy and I can assure
> you, a shield disconnected at one end is not effective agaisnt RF.
You and Scott are talking about two different things here.
Scott is talking about a *balanced* (2 wire) connection with an additional
shield, whose sole function is being shield. You are talking about an
unbalanced (1 wire) connection where the shield is part of the return path.
In that case you *must* connect the shield at both ends.
Meindert
Scott Dorsey
No, go back to the original question.
I am talking about an unbalanced line being run over shielded 2-conductor
cable rather than coax. This gives you, as I mention, one of the benefits
of balancing without any of the others.
If it's a balanced line, telescoping shields is also important because it
allows you to configure your grounds properly. (ref. Ott again). But there
it's not a thing you do casually.
GregS
>On Oct 29, 2:02=A0pm, klu...@panix.com (Scott Dorsey) wrote:
>> Mark =A0<makol...@yahoo.com> wrote:
>> >I'm not talking about a low frequency 60 Hz magnetic field, yes you
>> >need mu metal for that.
>>
>> >I'm talking about a plane EM wave at 100 MHz in the far field where
>> >known as RFI.
>>
>> >A shield not connected at one end is NOT effective against RFI at 100
>> >MHz.
>>
>> >There are many things I don't know and some things I do know, and that
>> >is one of the things I DO know.
>>
>> Hint: the two grounds are kept equipotential by the return wire.
>>
>> --scott
>>
>> --
>> "C'est un Nagra. =A0C'est suisse, et tres, tres precis."
>
>
>
>
>I "try it" every time I get a defective coax cable in the lab and it
>radiates all over the place....
>
>It's like a "farady shield" with one side cut open...
>
>Any current induced on the outside of the shield by an external field
>will flow at the gap into the inside of the shield and then be induced
>onto the inner conductors.
>
>I know I have read about the open shield being effective in
>___audio___ books and it may be effecitve in certain special cases
>against a particualr type of interference (Hi Z low frequency), but in
>general it is NOT effective at shielding RF.
>
>I am not a professional audio guy and I come here to learn about
>audio but I AM a VERY experinced professional RF guy and I can assure
>you, a shield disconnected at one end is not effective agaisnt RF.
>
>hint: "slot antenna"
>
> Mark
The thing with grounding both ends creates a ground 60 Hz loop.
Thats the main consideration, AT 60 Hz.
If the signal is not on that then it doesn't matter.
I would agree that RF shielding must be continous. Even
when designing larger shielded rooms.
Ground is only ground when its phase is in phase
or out of phase, and not 75 degrees or something like that.
Quarter wave stubs and half wave stubs, nothing in between
multiples.
greg
Arny Krueger
> On Oct 27, 2:41 pm, klu...@panix.com (Scott Dorsey) wrote:
>> Gareth Magennis <sound.serv...@btconnect.com> wrote:
>>> "Scott Dorsey" <klu...@panix.com> wrote in message
>>>> Gareth Magennis <sound.serv...@btconnect.com> wrote:
>>>>> just curious as to what is the best practice: If you
>>>>> have an unbalanced connection, say RCA to RCA or jack
>>>>> to jack, and use balanced cable, what is
>>>>> the best way to wire it? Parallel both cores, only
>>>>> connect one core with the other floating, or ground
>>>>> the unused core?
>>>> One wire goes from tip to tip, the other wire goes
>>>> from sleeve to sleeve. Connect the shield to sleeve on
>>>> one side only, not both.
>>> Would that actually be any improvement over unbalanced
>>> cable?
>> No. But if you have big rolls of twisted pair and you
>> don't have any single-conductor lying around, you might
>> as well. Plus maybe one day you'll want to cut one end
>> off and put an XLR on. --scott
Looks to me like a pretty pragmatic answer, and one that will work well in
just about every case.
> OP
> you will get better RF shielding if you connect the
> shield to the shell on BOTH ends.
But this is audio, so ground loops are strong potential problem. If you
rely on the shield as your ground connection at both ends, then it should be
the only signal ground connection between the two ends.
I look at it this way.
No way are chassis grounds identical at all frequencies at both ends of a
cable.
I have a choice - put the ground around the signal conductor at ground level
at the source (an output), or put it at ground level at the load (an input).
Now right up front, grounding the cable's shield at *either* end is far
better than no grounding it at all. And practically speaking, they could be
the same.
OTOH, in many more cases, the ground that I trust the most of the two is
the one at the load which is an input.
There must be continuity for both + and - signal connections between the
ends, so we relegate that to the otherwise unused other inner conductor.
If we connect the shield at both ends, then there is no need for any other
connection between the two ground, so the otherwise unused inner conductor
should be connected to either ground or the signal + at both ends.
> Otherwise I agree, for an RIAA pre-amp the capciantce is
> critical and you should use cable /length that provides
> the correct C that the cart. is expecting to see.
Very important with MM cartridges, especially those made by Shure. IME
Grados are less picky, but still have some sensitivity.
GregS
Ground must not not in phase to any degree.
A microwave oven has a 1/4 wave stub around the seal,
about 1.25 inches. A 60 hz stub would be 4100000 feet,
or 776 miles. An electrostatic shield will become
inneffective over a few feet to a AM broadcast station,
thats ended with a stub.
greg
geoff
"GregS" <zekf...@zekfrivolous.com> wrote in message
news:hcesvf$399$1...@usenet01.srv.cis.pitt.edu...
> The thing with grounding both ends creates a ground 60 Hz loop.
Not in all of the world. Much (most ?) of the world is 50Hz, but of copurse
the earth-loop effects are not only bad at the local mains frequency.
geoff
fredbloggstwo
"GregS" <zekf...@zekfrivolous.com> wrote in message
news:hcesvf$399$1...@usenet01.srv.cis.pitt.edu...
Can I suggest that everyone has a look at Tony Waldron's excellent set of
articles on this subject - they can be found at
http://www.tonywaldron.co.uk/
This old myth about not connecting the cable screen at both ends is clearly
dealt with
Cheers
Mike
Scott Dorsey
>
>Can I suggest that everyone has a look at Tony Waldron's excellent set of
>articles on this subject - they can be found at
>http://www.tonywaldron.co.uk/
>
>This old myth about not connecting the cable screen at both ends is clearly
>dealt with
I agree that his work is excellent, but it does not apply at all in this
case. We aren't talking about lifting shields on balanced lines here.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."
GregS
>fredbloggstwo <fredblo...@hastings1066.plus.com> wrote:
>>
>>Can I suggest that everyone has a look at Tony Waldron's excellent set of
>>articles on this subject - they can be found at
>>http://www.tonywaldron.co.uk/
>>
>>This old myth about not connecting the cable screen at both ends is clearly
>>dealt with
>
>I agree that his work is excellent, but it does not apply at all in this
>case. We aren't talking about lifting shields on balanced lines here.
>--scott
Another thing, I was thinking of turntable hookups and never ground
both shield ends. You actually have to look into the phono cables to see how they are wired.
greg
fredbloggstwo
I thought we were: GregS' post, to which I replied, was clearly talking
about shielding and RF interference, which is why I thought Tony's article
would be useful reading.
Regards for a wet and windy UK
Mike
GregS
I was just trying to say shielding can be different at RF. The majority
of noise is at 50 or 60 Hz.
greg
Scott Dorsey
>"Scott Dorsey" <klu...@panix.com> wrote in message
>news:heh8jg$qt$1...@panix2.panix.com...
>> fredbloggstwo <fredblo...@hastings1066.plus.com> wrote:
>>>
>>>Can I suggest that everyone has a look at Tony Waldron's excellent set of
>>>articles on this subject - they can be found at
>>>http://www.tonywaldron.co.uk/
>>>
>>>This old myth about not connecting the cable screen at both ends is
>>>clearly
>>>dealt with
>>
>> I agree that his work is excellent, but it does not apply at all in this
>> case. We aren't talking about lifting shields on balanced lines here.
>
>about shielding and RF interference, which is why I thought Tony's article
>would be useful reading.
The WHOLE subject of this thread is using a 2-conductor shielded cable to
connect two unbalanced sources. With this method, the tip and sleeve are
connected to the two leads of the shielded pair. The question is whether
the shield should be bonded only on one end, or on both ends.
The subject is NOT about balanced connections, about lifting grounds, or
about telescoping grounds in ANY other context. It's not about unbalanced
50 ohm connections with broken ground connections either.
The subject in fact is NOT about grounding topology at all, but about shielding
topology when it's separated from grounding.