Author Topic: Measuring crosstalk in a cable  (Read 4666 times)

0 Members and 1 Guest are viewing this topic.

Offline david77

  • Frequent Contributor
  • **
  • Posts: 806
  • Country: de
Measuring crosstalk in a cable
« on: November 07, 2012, 12:39:09 pm »
I want to send multiple stereo audio channels over a short piece of cable, max. 2m. Naturally I'd like to keep crosstalk between L+R and between individual channels to a minimum. My first bright idea was to use a multicore audio cable, that works well but is just to thick. I can't fit it to any connectors I'd like to use and multiple connectors increases cost and looks unprofessional.

I've done some quick and dirty measurments yesterday using the multicore and using a shielded cable, type LIYCY.
The multicore excells at channel crosstalk as one would expect as each channel is individually shielded. L+R crosstalk attenuation is about 67dB. I have to say I've done the tests using unbalanced signal lines, so L and R signal are together in one core.

Now in a piece of LIYCY cable I measured crosstalk of 89dB, even better if I ground some of the wires in the cable.
The LIYCY I'm using is not a twisted pair, so am I right in thinking I'd get even lower crosstalk using a TP cable connecting one of the wires in each pair to ground?

My measuring setup looked ike this:

Audio sinewave generator @ 1kHz/1,55V/600Ohms impedance
feeding into "sending" wire
          |
         \/
cable under test, shield connected to ground
          |
         \/
termination into 560R resistors, both send and receive wires
          |
         \/
HP400F AC millivoltmeter
connected to "receiving" wire

Anything I'm missing here please let me know.
The crosstalk increases rapidy when I go over about 6kHz, that was expected and doesn't bother me too much...or should it?
« Last Edit: November 07, 2012, 12:43:14 pm by david77 »
 

Offline DavidDLC

  • Frequent Contributor
  • **
  • Posts: 740
  • Country: us
Re: Measuring crosstalk in a cable
« Reply #1 on: November 07, 2012, 05:50:32 pm »
My first bright idea was to use a multicore audio cable, that works well but is just to thick.

If it does not totally work, then is not a "bright" idea !

 ;D ;D ;D ;D ;D ;D ;D ;D
 

Offline david77

  • Frequent Contributor
  • **
  • Posts: 806
  • Country: de
Re: Measuring crosstalk in a cable
« Reply #2 on: November 07, 2012, 06:03:21 pm »
Well, I ordered it out of the catalogue and was a bit surprised when I received a length of hose pipe a few days later. I read the datasheet and knew it was 17mm in diameter, but without seeing it that seemed reasonable to me  ::).
 

Offline beaker353

  • Supporter
  • ****
  • Posts: 102
  • Country: us
Measuring crosstalk in a cable
« Reply #3 on: November 08, 2012, 04:07:56 am »
I don't see why you would need nearly that low value of termination resistors. What exactly are you sending from and to?

-EM
 

Offline vk6zgo

  • Super Contributor
  • ***
  • Posts: 5336
  • Country: au
Re: Measuring crosstalk in a cable
« Reply #4 on: November 08, 2012, 07:38:02 am »
His generator has a 600 Ohm output impedance,so he has terminated the other end in 600 Ohms.
 

Offline david77

  • Frequent Contributor
  • **
  • Posts: 806
  • Country: de
Re: Measuring crosstalk in a cable
« Reply #5 on: November 08, 2012, 06:49:05 pm »
The 600R input impedance is a standard value in studio/broadcast equipment and I want to stay compatible.
Also isn't it easier to maintain low crosstalk, good s/n ratio using a low impedance line?
The audio signals come out of a rack of summing amps and need to go into another rack to be distributed to their destinations.
In the end the whole arrangement will consist of three separate enclosures all of them together making up a rather complicated HiFi mixer/preamp/distribution box.
 

Offline SeanB

  • Super Contributor
  • ***
  • Posts: 15381
  • Country: za
Re: Measuring crosstalk in a cable
« Reply #6 on: November 08, 2012, 07:59:09 pm »
Have you considered using twisted pairs? Works well enough for telephones and they are 600R loads.
 

Offline david77

  • Frequent Contributor
  • **
  • Posts: 806
  • Country: de
Re: Measuring crosstalk in a cable
« Reply #7 on: November 09, 2012, 12:30:10 am »
Yes, I've done some measurments using twisted pair.
Using the same setup as before I measured 83dB for a 1m piece of TP cable. This could probably be improved somewhat using balanced lines - as POTS telephony does I believe - but that means considerable extra effort.
 

Offline beaker353

  • Supporter
  • ****
  • Posts: 102
  • Country: us
Measuring crosstalk in a cable
« Reply #8 on: November 09, 2012, 02:45:15 am »
I highly doubt you will interface with anything near 600ohms. Other than a small handful if rather odd special mic preamps, all pro level gear produced in the last two decades is going to have line level input impedances of 5kohms and up. You will find gear with OUTPUT impedances of 600ohms, but this is something complete different than INPUT impedances.

-EM (professional audio systems engineer for over 15 years)
 

Offline david77

  • Frequent Contributor
  • **
  • Posts: 806
  • Country: de
Re: Measuring crosstalk in a cable
« Reply #9 on: November 09, 2012, 07:22:31 am »
Yes Beaker you are right, of course. I was half asleep when I wrote that.
At this point in my system I don't have to interface to anything else anyway. This is just an interconnection point between two separate 19" cases.

I can't really use a 12-pair multicore so want to use regular shielded cable. My question is how do I use regular LIYCY and still keep crosstalk to a minimum?
 

Offline SeanB

  • Super Contributor
  • ***
  • Posts: 15381
  • Country: za
Re: Measuring crosstalk in a cable
« Reply #10 on: November 09, 2012, 02:45:27 pm »
Shielded Cat5 multidrop cable?
 

Offline beaker353

  • Supporter
  • ****
  • Posts: 102
  • Country: us
Measuring crosstalk in a cable
« Reply #11 on: November 10, 2012, 04:07:01 am »
You have many options as far as cable and connectors. Lets start with how many total transmission lines you need between the two racks and I can help you narrow down your options.

-EM
 

Offline david77

  • Frequent Contributor
  • **
  • Posts: 806
  • Country: de
Re: Measuring crosstalk in a cable
« Reply #12 on: November 10, 2012, 04:18:18 am »
Ok, let's see what you come up with  ;).
In this particular case I need 7 individual stereo lines.
 

Offline beaker353

  • Supporter
  • ****
  • Posts: 102
  • Country: us
Measuring crosstalk in a cable
« Reply #13 on: November 10, 2012, 10:12:46 pm »
DB25's are often used for 8 channels of balanced lines in and out of multitrack recorders. You can use off the shelf cables like below to link the two racks. Panel mount DB25's are easy enough to mount and wire. Each of the 8 lines are individually shielded. Ground the shield and one of the conductors in each twisted pair for maximum noise immunity. If you want to spend a whole series lot of money you can use their MASS connectors for up to 56 balanced channels in a single connector.

http://whirlwindusa.com/catalog/snakes-splitters-and-multiwiring-systems/snakes/connect-series-dbmd-010

-EM

 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf