FreeLists / si-list / [SI-LIST] Re: Coaxial Cable crosstalk

[Rob Hinz <rob@xxxxxxxxxx>]

Hi Craig,

In my view, there are principally two mechanisms that result in coupling 
between coaxial cables:

1) Direct coupling due to fields penetrating the coax shield
2) End effects

Field penetration of the coax shield can occur in braided type cables 
because there are small gaps in the shield. While significantly attenuated, 
the fields will penetrate these gaps and couple to the interior of the 
cable. Field penetration can occur even in solid shield cable because of 
the finite conductivity of the metal used. We often think of skin depth as 
the maximum depth at which current flows in a conductor at a particular 
frequency. This is not the case, however. Fields penetrating good 
conductors decay at an exponential rate proportional to the conductivity of 
the metal and the frequency of the incident signal. So, for a finite 
thickness of metal, there will be some penetration of the metal by the 
field. This is much more pronounced at low frequencies, like those you are 
using.

If I were to try to characterize this, as a first approximation, I would 
use a good 2D field solver that correctly handles fields within metals, as 
opposed to assuming surface currents only. I would create a model of the 
twinax using actual dimensions and material properties and let the field 
solver compute a coupled line model of the system. I would probably try 
this in Ansoft 2D extractor first since that is what I have but I suppose 
there might be others that would work. To be fair, I am not certain how 
well this will work. My main concern is that the coupling will be very weak 
compared to all the other effects of the system. This may result in a lot 
of numerical noise in the solution. There is also a question of how 
accurately the field penetration in the metal is modeled. The field 
strength will be changing very quickly as it penetrates the metal. This 
would seem to require very fine discretization of the metal for accurate 
resolution.

The end effect coupling may well dominate in any case. At the end of the 
coax, the all encompassing shield is broken. Depending on how the cable is 
terminated, signal currents might even flow back down the exterior of the 
twinax shield. In addition, it sounds as though the cables terminate in the 
same enclosure. It is likely that the termination structures at the end of 
the lines, as they are collected at a common point, will have significant 
coupling. This coupling is probably quite difficult to characterize.

As surprising as it seems, you have a difficult modeling problem. I hope 
this helps with some of the concepts, even if it falls short of a good answer.

Rob Hinz
Senior Electromagnetics Specialist
SiQual Corporation
rob@xxxxxxxxxx
phone (503)885-1231
fax   (503)885-0550
http://www.siqual.com





At 11:48 AM 2/5/2002 -0500, C Deibele wrote:



>"Daniel, Erik S., Ph.D." wrote:
> >
> > Craig-
> >
> > I think you're right -- this is a hard problem.  In our experience with 
> good
> > quality coax cables in the lab, the crosstalk is so low (easily better than
> > -100 dB) that I think you'd find it extremely hard to predict -- much 
> easier
> > to measure with a good spectral quality signal source and a spectrum
> > analyzer with fairly low noise / narrow resolution bandwidth.  Because the
> > crosstalk between very good cables is so low, you may find that the
> > effective crosstalk between signals on the two cables is totally 
> dictated by
> > what is happening at the cable ends (e.g., connection to a PCB where the
> > signal is not totally enclosed by ground).
> >
> > If the cables you are interested in are not as high a quality (e.g., 
> thinner
> > ground shield with a looser braid, etc.), it may be even easier to measure
> > the crosstalk, but it seems it would still be difficult to 
> predict.  Can you
> > give some more information on the type of cables you are interested in, the
> > approximate level of crosstalk you care about, and the frequencies of
> > interest?  Our limited experience here in the lab was with good quality
> > flexible cables with SMA or 3.5mm connectors, looking at frequencies of a
> > few hundred MHz to a few GHz, lookinf for crosstalk at the ~-100 dB level.
> >
> > Some cable manufacturers specify crosstalk per foot or some such thing
> > (assuming some cable separation I suppose).  I think I've seen this for 
> RG-6
> > cable before.
> >
>
>Well, I am not interested in any actual cable.  I am more interested in
>some general theory/theoretical concepts.
>
>We have an application where we have twinax (twinax is like coax, except
>it has two conductors surrounded by an outer shield instead of one
>conductor surrounded by an outer shield like coax).  The frequency is
>DC-1MHz or so.
>
>These cables are long....separate sources, and they will all be
>collected at a common point where they will share a small space, and
>then separate again.
>
>There is some cables with high currents on them...on the order of a few
>thousand amps....which may/may not have switching supplies driving them.
>
>So, if I have a small signal cable nearby...how can I reliably predict
>what the crosstalk/might be?
>------------------------------------------------------------------
>To unsubscribe from si-list:
>si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
>
>or to administer your membership from a web page, go to:
>http://www.freelists.org/webpage/si-list
>
>For help:
>si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
>
>List archives are viewable at:
>                 http://www.freelists.org/archives/si-list
>or at our remote archives:
>                 http://groups.yahoo.com/group/si-list/messages
>Old (prior to June 6, 2001) list archives are viewable at:
>                 http://www.qsl.net/wb6tpu
>

Rob Hinz
Senior Electromagnetics Specialist
SiQual Corporation
rob@xxxxxxxxxx
phone (503)885-1231
fax   (503)885-0550
http://www.siqual.com

------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field

or to administer your membership from a web page, go to:
http://www.freelists.org/webpage/si-list

For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field

List archives are viewable at:     
                http://www.freelists.org/archives/si-list
or at our remote archives:
                http://groups.yahoo.com/group/si-list/messages 
Old (prior to June 6, 2001) list archives are viewable at:
                http://www.qsl.net/wb6tpu