2
\$\begingroup\$

I need to send an audio signal over 3 meters and back again on the same line (the plan is to put an on/off switch and a volume regulator at the end of the line).

Additionally I would like to send 12 volts power in the same cable, with small amperage.

Is it possible to use an Cat5 Ethernet cable for that purpose? I have those in abundance, and want to avoid buying a dedicated cable for that purpose.

I would use following signals:

  • Audio Right In
  • Audio Left In
  • Audio Ground In
  • Audio Right Out
  • Audio Left Out
  • Audio Ground Out
  • 12 V +
  • 12 V -
\$\endgroup\$
7
  • 1
    \$\begingroup\$ It might introduce too much noise/interference - what are your in and out signal levels. Also, why not just try it and see - it doesn't seem like it will be much of a hardship to perform a small experiment. \$\endgroup\$
    – Andy aka
    Commented Jul 18, 2018 at 10:07
  • \$\begingroup\$ If you keep the impedance (to ground) of the signal lines low you can minimize the influence of noise and hum. In general line outputs might be too high in impedance (a couple of kilo ohms) but a headphone output could do the job (a couple of hundred ohms or less). Then also use a low-resistance potmeter, like 100 ohms to adjust the volume. \$\endgroup\$ Commented Jul 18, 2018 at 10:11
  • \$\begingroup\$ Use the cable in pairs as ethernet does.. That will help reduce interference and crosstalk. It is also easier to wire correctly. \$\endgroup\$
    – JRE
    Commented Jul 18, 2018 at 10:37
  • \$\begingroup\$ Hang on, why are you sending the signal all the way out and then back again? Or are they different signals? \$\endgroup\$
    – pjc50
    Commented Jul 18, 2018 at 11:07
  • \$\begingroup\$ It will be a line signal going to a amplifier after that. I am sending the signal back and forth because it will be modified (volume and on/off). \$\endgroup\$
    – Paul Weber
    Commented Jul 18, 2018 at 17:49

4 Answers 4

2
\$\begingroup\$

I think a good question to ask here is: what kind of audio signal?

If you can spare e.g. an Opamp per channel and end, plus the power for that, and a terminating relatively low resistor (e.g. 120 Ω, 47Ω, or even less), then you'd waste power, but you'd get a very stable transmission.

  • Audio Right In
  • Audio Left In
  • Audio Ground In
  • Audio Right Out
  • Audio Left Out
  • Audio Ground Out

Don't do that – your twisted pair cable is meant for differential pairs. So, instead of transmitting ground (which is just a recipe for all kinds of avoidable hum and DC currents), convert your single-ended audio on the transmitting side into a differential one, and transfer that on a twisted pair (blue / blue-white, and so on):

  • Blue: left -> +
  • BlueWhite: left -> -
  • Brown: right -> +
  • BrownWhite: right -> -
  • Green: left <- +
  • GreenWhite: left <- -
  • Orange: right <- +
  • OrangeWhite: right <- -

Converting from single-ended to differential can be done using inverting op-amp circuitry, or opamps with differential output, or dedicated differential line driver ICs, or signal transformers.

  • 12 V +
  • 12 V -

Nah. Ethernet cable's not a good conductor of current, so you'd avoid this.

Also, if you use a signal transformer on the end, you can just use the center tap to impose DC over your differential audio signal, and can avoid having dedicated power pairs. That's how Power-over-Ethernet works!

\$\endgroup\$
2
\$\begingroup\$

Cat5 cable intended for ethernet has four twisted pairs. There will be significant capacitive coupling between pairs, but each pair is twisted at a different pitch so that the coupling is largely common mode over reasonable lengths.

To minimize crosstalk, keep the signals low impedance and definitely make them differential on each pair.

The main problem is that you're one pair short. If this is a one-off then, just run another two wires for power and ground. If you really have to cram all this analog information plus power onto a 4-pair cable, then transmit power as a DC common mode offset between two of the pairs.

A totally different approach is to send the audio data digitally. That eliminates crosstalk and other noise pickup, and doesn't require separate physical wires for separate channels. At only 3 meters, you should be able to use a bit rate of a few MHz with relatively simple drivers and receivers at each end.

Let's do the math. Let's say the audio is "HiFi" quality. That requires 16 bits/sample at over 40 kHz sample rate. Let's use 50 kHz. That comes out to 100 kBytes/s per channel. For two channels, that's 200 kBytes/s, which is 1.6 Mbit/s raw data rate. Add some overhead for protocol at low and high levels, and you need 3 to 5 Mbit/s. That really should be doable over 3 meters rather easily.

Of course if you only need to transfer "voice" quality, then the data rate requirements are significantly relaxed. 8 bit samples at 8 kHz, especially with a little companding, can do quite a nice job of transferring voice.

\$\endgroup\$
0
\$\begingroup\$

To send audio across Cat-5, follow the practice of the original purpose of those sockets - NOT Ethernet but audio (telephone systems).

That means, use balanced audio to eliminate interference.

And THAT means, you need 4 pairs (L/R out, L/R return) leaving no room for DC power.

Fortunately (if your power needs are relatively low) you can use a trick called "phantom circuits" to communicate power : dedicate at least one (possibly 2) pairs to 0V, and another to 12V, carrying power as common mode voltages on each pair.

Ethernet has started adopting this (as POE) but it's much older than that.

Marcus's answer just popped up as I was editing my comment into an answer : follow the original colour codes to keep the pairs together, as he suggests.

\$\endgroup\$
0
\$\begingroup\$

Audio level matters. How much signal is already there affects how noticeable noise will be.

If this is a "speaker" level signal of more than a few watts, it'll probably be too powerful for the cable and you might suffer attenuation and heating in the cable.

If this is a "line" level or headphone extension, it'll work fine unless it's in an electrically noisy environment.

If this is a microphone level, or going into another amplifier, then you'll amplify the noise and it will probably be noticeable. Especially from high-power radio sources like cellphones.

\$\endgroup\$
1
  • \$\begingroup\$ The signal will be "line" level and going to an ampflifier after that (final stage before the speakers) \$\endgroup\$
    – Paul Weber
    Commented Jul 18, 2018 at 17:47

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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