1
\$\begingroup\$

I'm currently using a cable that has multiple twisted pair conductors. I'm attempting to send a several single ended signals down some of these conductors at a frequency of 12.5MHz. Essentially, I'm only using one conductor of each twisted pair.

Will this pose any issue? The fact that there is an unused/unterminated conductor twisted around one of my conductors, will this have any transformer like effects? I have been unable to find any resources online.

Thanks,

\$\endgroup\$
  • \$\begingroup\$ Sure, using twisted pair conductors in sigle ended fashion makes no sense. Why don't you use a flat cable? \$\endgroup\$ – Marko Buršič Jul 15 '16 at 14:19
  • \$\begingroup\$ Also thought about this. I think if you use wire-pairs & treat the signal at the fair end as balanced pair, you might get some of the benefits of twisted-pair-noise-cancellation. \$\endgroup\$ – st2000 Jul 15 '16 at 14:34
  • \$\begingroup\$ @MarkoBuršič The manufacturer created a device that can operate in both single ended and differential mode. The cable/connector type features twisted pairs. However, my receiving device is single ended thus I'm required to use the twisted pairs. \$\endgroup\$ – Izzo Jul 15 '16 at 14:51
4
\$\begingroup\$

Essentially, I'm only using one conductor of each twisted pair.

Actually, you're using 2 conductors. One conductor of the twisted pair, and another that you haven't specified, which returns the current, or is the reference voltage. It will be there, even if you don't know where it is at the moment.

Is a ground conductor in the cable, or a ground loop from your transmitting end to your receiving end? Or is it the other wire in the twisted pair.

If it's a ground conductor in the cable, and it's working OK to share that between several circuits without too much crosstalk, then that's OK. The other redundant conductor twisted with it sould not be too much of a problem either left open or paralleled with the signal conductor, as long as the frequency is fairly low.

If it's a random ground loop, then that's almost always bad. You need to know where your return is.

If it is the other conductor in the pair, then you will be using it at the far end as the reference for the signal, by taking both conductors into a differential receiver. This will isolate the signals from externals signals, and the pairs from each other. That's the best way to do it.

| improve this answer | |
\$\endgroup\$
  • \$\begingroup\$ I'm using two twisted pairs. For each pair, I'm only implementing a single conductor to send a signal. The other two conductor have not been terminated. I then have a separate GND conductor that each signal references. I'm curious if I should GND terminate the unused conductor for each twisted pair. \$\endgroup\$ – Izzo Jul 15 '16 at 16:39
  • \$\begingroup\$ If you are running a ground conductor per signal in the same cable, then it would make most sense for that conductor to be the one that's twisted with the signal. There will be slight differences depending on whether you ground, parallel, or leave floating the other twisted conductor. If you can cope with the ground returns being seperate from the signal lines, then it sounds like your signals are not taxing enough to be able to tell the difference between them. \$\endgroup\$ – Neil_UK Jul 15 '16 at 16:50
  • \$\begingroup\$ I want to carry multiple signals via some distance and thiught about using twisted pair cable because off the noise rejection. I'd use a signal and a ground per pair, but couldn't multiple ground paths cause problems? \$\endgroup\$ – Julian F. Weinert Sep 16 '19 at 17:37
  • \$\begingroup\$ Multiple wires connecting a ground at one end to a ground at the other will not cause a problem. \$\endgroup\$ – Neil_UK Sep 16 '19 at 20:03
2
\$\begingroup\$

I'm attempting to send a several single ended signals down some of these conductors at a frequency of 12.5MHz. Essentially, I'm only using one conductor of each twisted pair.

Firstly, crosstalk will be an issue if you are not balancing the unused conductor in each twisted pair. This crosstalk is due to inter-pair capacitance.

To go into more detail, you need to think of a single twisted pair like a transformer (ignoring the complication of reflections). Look at this circuit: -

enter image description here

This shows how a single ended driver can drive a twisted pair. Note the controlled impedances Rs at the sending end. This provides a balanced impedance at the sending end. If Rs were not used in the grounded wire ("cold") there would be an impedance imbalance and noise impregnating the cable would produce a differential signal that would be detectable at the receiver.

That's the simplest proper way to drive twisted pair from a single-ended driver. However, the driver output may start to get higher impedance at higher frequencies and what people do is connect the ground end of Rs to a driver identical to the one marked "line output". It can still just drive a constant ground signal but usually it will drive an antiphase signal to improve SNR.

With the single ended driver above, transformer action in the twisted pair ensures that the signal at A is the inverse of the signal at B.

So, it's imperative that the "unused" wire in each twisted pair is grounded via Rs. This has another benefit - because the AC voltages appearing at A and B are antiphase, this will be partially true along most of the cable. This means that electric field crosstalk between pairs is also significantly reduced. This is why a cable carry several twisted pairs has the twist spacing different for each pair. It's the same reason why telephone poles swapped their wire pair positions quite often: -

enter image description here

Also, because of the signal inversion due to terminating the sending end correctly, the currents down each wire in the pair are "opposite" thus, magnetic balance is largely achieved and this also cuts down on crosstalk problems between pairs.

The next problem is how you terminate the receiving end when you only have a single-ended receiver. You must terminate both ends to local ground/screen with the same impedance for all the benefits to accrue. Given that you haven't detailed your circuit I'm not going to make a guess.

| improve this answer | |
\$\endgroup\$
  • \$\begingroup\$ Hi Andy. My position is unique, the device I'm working with can operate as either a single ended or differential driver. However, I'm using the single ended mode because the receiver is single ended. However, due to constraints in the mechanical design, I have no ability to re terminate the unused twisted pair at the sender side. Assuming that the output impedance of the unused twisted pair conductor is 30 ohms, should I be terminating that conductor with 30 ohms on the receiver end? \$\endgroup\$ – Izzo Jul 15 '16 at 16:36
  • \$\begingroup\$ All bets are off if you can't terminate at the send end, sorry \$\endgroup\$ – Andy aka Jul 15 '16 at 16:39

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.