2
\$\begingroup\$

I understand how a directional coupler can be used to measure signals travelling in one direction on a wire. What I don't understand is exactly how this would work on a differential signal, E.G. LVDS (which stands for Low Voltage Differential Signal).

My ultimate goal is to test the quality (impedance really) of hand made cables. Ideally, they need to be 100R impedance. One way to test the quality is to measure the amount of reflected signal from the terminator.

Measuring reflection using directional couplers is often done for single ended RF signals, but I have never seen it done on differential signals.

LVDS Differential Coupler

Do I do it like this, using two couplers (these ones), one for each wire? If so, which two output ports should I watch if I want to see reflections from the terminator? A&C, B&C, A&D or B&C? Should I add the signals from the ports, or subtract them?

Alternatively, is there such a thing as a differential directional coupler? I tried googling for it, but nothing much came up.

\$\endgroup\$
  • \$\begingroup\$ It looks like you would take A-C as a differential signal. \$\endgroup\$ – Ben Voigt Apr 10 '12 at 17:27
  • \$\begingroup\$ I'm having trouble figuring out what you are trying to do, and what you mean by "directional coupler". The only directional couplers I could find are for RF and completely unsuitable for differential signals. Maybe you should provide more of a big-picture view of what you are trying to do, and a link to a directional coupler datasheet that you are considering. \$\endgroup\$ – user3624 Apr 10 '12 at 17:30
  • \$\begingroup\$ @David: Digital signals in the RF or microwave frequency bands typically are sent differentially and/or current-mode, for better noise rejection. \$\endgroup\$ – Ben Voigt Apr 10 '12 at 17:38
  • \$\begingroup\$ @BenVoigt I know that. I do 2-5 GHz LVDS all the time. But it doesn't help me understand Rocketmagnet's question any better. \$\endgroup\$ – user3624 Apr 10 '12 at 17:45
  • \$\begingroup\$ @DavidKessner I've added more explanation to my question. I hope that's clearer. If not, let me know. \$\endgroup\$ – Rocketmagnet Apr 10 '12 at 18:17
1
\$\begingroup\$

NORTEL already patented this concept in 1993 but exactly as shown.. http://www.google.com/patents?id=eR8hAAAAEBAJ&printsec=abstract&zoom=4#v=onepage&q&f=false

\$\endgroup\$
  • \$\begingroup\$ @TonyStwart, a link on its own is not considered an answer, can you expand some on this site so that google rot does not destroy this answer? \$\endgroup\$ – Kortuk Apr 11 '12 at 17:21
0
\$\begingroup\$

Adding two different couplers as described in the drawing will create a part of the circuit that is not correctly impedance matched, and it will create reflected waves in both directions and disrupt the measurement. The cable has a certain characteristic impedance regardless of whether the signal it should conduct is a differential or single ended. If you are trying to verify the characteristic impedance of the cable, the measurement method could be the same as with a coaxial cable (i.e, a single directional coupler matched to 100 Ohm, with the ground terminal connected to one of the differential wires)

\$\endgroup\$
  • \$\begingroup\$ Thanks Ilya. I know I'm going to sound really thick and annoying, but can you draw me a schematic of this? \$\endgroup\$ – Rocketmagnet Apr 10 '12 at 20:19
  • \$\begingroup\$ Also, why is it that adding two directional couplers will affect the impedance, but adding only one will not? \$\endgroup\$ – Rocketmagnet Apr 10 '12 at 20:20
  • \$\begingroup\$ I don't know if it makes a huge difference, but I have modified my schematic to better represent the set up I had in mind. \$\endgroup\$ – Rocketmagnet Apr 10 '12 at 20:32
  • \$\begingroup\$ @Rocketmagnet, I agree, adding two or one, the primary cable has to be modified to ensure your impedance is carefully matched, if it is a large issue. If a 1% mismatch is not an issue you might get away with a very weak coupler. \$\endgroup\$ – Kortuk Apr 11 '12 at 17:22
  • \$\begingroup\$ @Kortuk I don't understand. I don't want to modify the primary cable to match the couplers? Are you saying that if I use weak couplers, then the set up I show in my question mill work. \$\endgroup\$ – Rocketmagnet Apr 11 '12 at 19:10

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.