# Four wires transmission line for data

1. Jul 19, 2011

### SredniVashtar

I read on Kraus' "Electromagnetics", 4th edition (p 583), that four-wire transmission lines (in a quadrupole configuration - the conductor with the same polarity are in diagonal) improve shielding.

+ -
- +

I tried to find some more information about this sort of transmission lines but I couldn't find anything sensible. Are these T.L. used for data communications, how can one compute the characteristic impedance (considering the the lossy case too), are there tables with properties based on geometrical specs? Are there experimental studies? (And so on...)

I would like to know if it's possible to adapt a standard Ethernet cable to mimick this sort of transmission line to improve its shielding from external interference.

Any pointers?

2. Jul 20, 2011

### SredniVashtar

No takers?
Bummer.

Let's see if we can compensate the lack of theory with some sound practice.

Anybody with a vector analyzer and the skills to use it (I guess one of those rules me out as a candidate :-) ) would care to measure the characteristic impedance of a cat5 or cat5/e Ethernet cable by using different 'quadrupole' configurations?

Also, what happens using all solid colors for plus, and all striped colors for minus?
Yep, that would be an octupole. Murphy's law is categorical on that: someone must have tried it!

How I wish there were T&M Points, where one could access instrumentation for half an hour, sipping a good coffee in the meantime...

Last edited: Jul 20, 2011
3. Jul 20, 2011

### Staff: Mentor

I haven't ever seen this configuration of TL, but perhaps it has some uses. It does look like it has the potential for better B & E shielding in some situations, but since it would be about twice as expensive as simple twisted pair, maybe that's why it is not commonly used.

4. Jul 20, 2011

### Staff: Mentor

Also keep in mind that for data cables that have multiple twisted pairs within the same sheath, the lay length of each twisted pair is different, to minimize crosstalk. I'm not sure how you would handle multiples of the 4-conductor straight TL in the same sheath --- the crosstalk would likely be worse than with regular twisted pairs.

5. Jul 21, 2011

### SredniVashtar

Your remarks on price and cross-talk are correct, yet I was thinking about this solution as a 'EMI suppression first' trade-off. In order to minimize EMI (from soldering stations operating nearby, for example, or from atmospheric discharges) I would be willing to pay not only twice, but four times the price of a twisted pair. Let aside the fact that Ethernet cable comes cheap in the four pairs flavor. Using a single cable for what would have been a two-wire communication should make crosstalk a non-problem, or am I missing something?

Anyone can spare a hint on how to simulate a four or eight-wires transmission line?
Should I consider capacitive links and mutual inductances between every couple of wires?
Is there some general 'impedance composition' formula for paralleled _completely uncoupled_ transmission lines? Like "two 50-ohms TL in parallel give an impedance of 25 ohm"? :-)

EDIT: Now I have found something sensible: a 1964 article about four wires transmission lines.
Sandia Corp. Monograph
C.W. Harrison Jr.
"http://www.ece.unm.edu/summa/notes/In/0065.pdf" [Broken]"
June 1964

It'll take some time to read it and digest it.
But it confirms the enhancement in EMI rejection and gives a formula for the characteristic impedance of a 4-wires line.
I wonder how different that would be from that of a line made with the four pairs in a UTP cat5 cable in this way

++ --

-- ++

Also, I cannot avoid asking myself what would happen if I connected the wires in different configurations, like

+- -+

+- -+
,
+- +-

-+ -+

I understand that the symmetry is broken by the twisting of the pairs inside the sheath. But perhaps the different geometries average out, due to the twisting, giving a completely different characteristic impedance.

I guess it's about time I bought a serious scope.

Last edited by a moderator: May 5, 2017