Cat 5 etc are used as data cables, but they're really just telecommunications cables (telephone). In telephony they have 1000 pair cables carrying voice without crosstalk between pairs. So it's well proven for carrying analogue signals. The other thing is that the reason the pairs don't interfere with each other is that there is no integer relationship between the twist rates of the pairs in the common jacket. That almost eliminates inter pair crosstalk. So great to use for audio even using unshielded twisted pair.
If a line level return is sent and unbalanced at the termination, and the mixer has poor CMNR, the induction and capacitive coupling could could be high enough to feedback at higher frequencies. Ensure all sources and terminations are balanced to reduce crosstalk between line level sends and mic inputs. It is best to not mix line and mic levels in the same cable when shielded if it can cause feedback such as monitors.
We use Cat cable all the time for analog audio and have for years. I have had tech's question it quite aggressively. I tell them to go tell Dave Rat he is wrong!
Thank you. This test shows crosstalk between on shielded twisted pairs with a 60db or so difference in signal level. 10x10x10x10x10x10 or 1,000,000 x more, and the crosstalk was not audible. With speaker cables we may have HF at 10v and LF at 100 or 500v which is only an 10x to 50x differential so we just don't have those drastic level differencea, so crosstalk is less of a concern. That said, to completely avoid crosstalk, separtely shielded pairs is best.
Thanks. In 4 and 8 conductor speaker cable neither the whole cable or the pairs are shielded, and typically the pairs aren't twisted either, correct? Why would that be? I wonder if there is much crosstalk when using nl8 for passive line arrays, or when using it as snake from amp world to passive monitors, etc.
@Connor the level difference and open circuit impeadance just isnt really an issue. On a mic pre your input impeadance is generally quite high compared to cable and the current flows are very tiny. In a loudspeaker cable, even the smallest driver is usually 16 ohms and still in the order of magnitude of hundreds of watts in todays systems.
@@DaveRat, do you think running unbalanced at that length would be just as likely (if not more so) to pick up more of the typical environmental noise than generate crosstalk? I won't be surprised if you hear all the usual junk from computers, light dimmers, FM radio stations (like with some super long guitar cables) before you'd hear crosstalk. Also, isn't crosstalk directly related to the capacitance of the jacket material on each wire? If data cables work in the GHz range that would explain why anything in the audio range should be practically immune to crosstalk on data cables. Sorry if my thoughts are somewhat scattered on this topic.
Yeah, I will look at doing some more cable tests. I'd like to do a star quad vs high quality vs low quality vs snake cable test. Just need to start gathering the various cables.
ordinary balanced twin-screened vs star-quad was a real eye-opener for me, back in the day, & then seeing tv data travelling 200m over twisted pair data cable.
This looks like it works well for balanced (XLR) microphone level signals with high impedance preamps. But If you try to use this for unbalanced cable runs (1/4” plug into guitar or keyboard) or worse, speaker cable you will definitely get cross talk. I’m sure this is obvious to sound engineers out there but maybe not to the lay person.
No, it will carry dmx, com, audio and aes3 at the same time without crosstalks issues just like any 4 channel audio snake. And just like with an audio snake it is generally not recommended to run extremely low level signals down the same cable as very high level signals. When using supercat sound cable or cat6a or cat 7, the crosstalk is superior to audio snakes as the perfection of twists on cat cables are much better
Yeah, working on that. I do the videos spontaneously. Often without planning and try and do a single take and when things are not quite right with the sound or other bits, if the content is there, I just post rather than retake.
@@DaveRat This is common as most audio engineers use plenty of headroom between 12 to 18db, where most UA-cam audio bloggers compress to less than 3 db of headroom, so unless there is post processing on the audio, it is considerably lower than most content. In digital, 0 is the max before clipping.
Could run a 100w amplifier and a speaker down a pair, or down two pairs using each pair connected in parallel, that might get some crosstalk. Should also be noted that for fast ethernet, ie anything in use today, only four of the eight RJ45 (or more strictly 8P8C) pins are used, so you can use additional pins for anything you want in theory, and you can build all kinds of nifty breakouts/bodge plugs for stuffing things down Cat5 that don't belong down there.
Yeah, I don't see mixing speaker level with mic level as something that would be implemented, but I do like the extremes. Maybe I will do a video on using cat5e for 70 volt systems
@@DaveRat it would be a baaaad idea on many levels, but I do wonder *how* bad the crosstalk leakage would be. Worst case would presumably be mixing unshielded vinyl cartridge phono with... I dunno... an arc welder? PWM power for something? Feed line for a high power radio station antenna? Power supply for a neon light? Torture / stress tests often reveal features which are difficult to observe under normal operating conditions.
@@DaveRat I have seen people using Cat5 as speaker cable, and have considered hijacking cat5 network wall boxes in office buildings / DIY recording spaces for an analogue talkback mic & speaker, so while it's not a good idea, it might have some applications.
Gig Ethernet and faster uses all the pins, and gig ethernet is likely FAR more common that fast (100mbit) or slower today. But back in the day, yes, you could use those unused pairs for any number of things from a whole second run of ethernet to analog signals.
@@everettchris1 Indeed, bear in mind I started on 10base2 coax. Where segregated networks are used, processing audio, video and internet via different switches, 100baseT is not a bottleneck for email, so there can in theory be available bandwidth for some extra analogue lines. This is what I have found myself thinking while eyeballing fixed sockets in walls, and having to use mobile phones to debug media networks from room to room, or floor to floor - could I make an analogue talkback bodge box? All the switches found everywhere can do 10base or 100base as well as gigabit, and 10/100 switches are still in use, and still on sale. Walkie talkies and mobile phones work fine most places, though, so it's remained a beard-stroke rather than a habit.
Cat 5 etc are used as data cables, but they're really just telecommunications cables (telephone). In telephony they have 1000 pair cables carrying voice without crosstalk between pairs. So it's well proven for carrying analogue signals. The other thing is that the reason the pairs don't interfere with each other is that there is no integer relationship between the twist rates of the pairs in the common jacket. That almost eliminates inter pair crosstalk. So great to use for audio even using unshielded twisted pair.
agreed
If a line level return is sent and unbalanced at the termination, and the mixer has poor CMNR, the induction and capacitive coupling could could be high enough to feedback at higher frequencies. Ensure all sources and terminations are balanced to reduce crosstalk between line level sends and mic inputs. It is best to not mix line and mic levels in the same cable when shielded if it can cause feedback such as monitors.
Would be interested to know what mixing desk was it (what the CMR preamp rating is) and whether the music source was fully balanced one.
Now this was a cool practical demonstration. I wouldn't guess this would be possible at all.
👍
We use Cat cable all the time for analog audio and have for years. I have had tech's question it quite aggressively. I tell them to go tell Dave Rat he is wrong!
I've been seeking a good way to demonstrate it in a simple way for years!
You rock, would love to see you test crosstalk between pairs inside nl4/nl8! Thanks for the great videos.
Thank you. This test shows crosstalk between on shielded twisted pairs with a 60db or so difference in signal level. 10x10x10x10x10x10 or 1,000,000 x more, and the crosstalk was not audible. With speaker cables we may have HF at 10v and LF at 100 or 500v which is only an 10x to 50x differential so we just don't have those drastic level differencea, so crosstalk is less of a concern. That said, to completely avoid crosstalk, separtely shielded pairs is best.
Thanks. In 4 and 8 conductor speaker cable neither the whole cable or the pairs are shielded, and typically the pairs aren't twisted either, correct? Why would that be? I wonder if there is much crosstalk when using nl8 for passive line arrays, or when using it as snake from amp world to passive monitors, etc.
@Connor the level difference and open circuit impeadance just isnt really an issue. On a mic pre your input impeadance is generally quite high compared to cable and the current flows are very tiny. In a loudspeaker cable, even the smallest driver is usually 16 ohms and still in the order of magnitude of hundreds of watts in todays systems.
Learn so much from this guy.
Thanks Dave. I was wondering about that
I think the twists in the cable pairs also do a good job of cancelling out any crosstalk.
Agreed
Nice one less test I'll need to do for myself
Yeah buddy I was looking for something to learn. 🤘✌️
Could you lift one side of the twisted pair (on the music, mic or both) to show how much the twist is doing for us?
Hmmm, interesting question. I will ponder using/testing the cable unbalanced. Which will show what I think you seek to hear
@@DaveRat, do you think running unbalanced at that length would be just as likely (if not more so) to pick up more of the typical environmental noise than generate crosstalk? I won't be surprised if you hear all the usual junk from computers, light dimmers, FM radio stations (like with some super long guitar cables) before you'd hear crosstalk. Also, isn't crosstalk directly related to the capacitance of the jacket material on each wire? If data cables work in the GHz range that would explain why anything in the audio range should be practically immune to crosstalk on data cables. Sorry if my thoughts are somewhat scattered on this topic.
@@paulthomas5901 I think unbalanced would have more Crosstalk and more issues with outside noise influence. I will do a test for ya!
10 pair cable is in common use in radio broadcast studios to distribute balanced line level audio.
👍
that's awesome
Agreed!
THANKS... :)
👍👍👍
Hi Dave is it possible that you do the same test with a normal mic cable and or a multicore cable?
Yeah, I will look at doing some more cable tests. I'd like to do a star quad vs high quality vs low quality vs snake cable test. Just need to start gathering the various cables.
Dave Rat thank you Dave big fan!!!!
CMRR is like magic. isn't there a spec for twists-per-inch in this stuff too, & minimum installation bend radius, that sort of thing?
ordinary balanced twin-screened vs star-quad was a real eye-opener for me, back in the day, & then seeing tv data travelling 200m over twisted pair data cable.
There are specs for Cat cables. Audio cable is kind of a free for all. That's why cat5 does so well for audio
This looks like it works well for balanced (XLR) microphone level signals with high impedance preamps. But If you try to use this for unbalanced cable runs (1/4” plug into guitar or keyboard) or worse, speaker cable you will definitely get cross talk. I’m sure this is obvious to sound engineers out there but maybe not to the lay person.
Agreed it works for balanced, as far as how it works for unbalanced, rather than making assumptions, how about I do a test!
Great idea
If you add a dmx signal on one of the channels, will you hear that?
No, it will carry dmx, com, audio and aes3 at the same time without crosstalks issues just like any 4 channel audio snake.
And just like with an audio snake it is generally not recommended to run extremely low level signals down the same cable as very high level signals.
When using supercat sound cable or cat6a or cat 7, the crosstalk is superior to audio snakes as the perfection of twists on cat cables are much better
Thanks Dave, I was curious if the higher frequency spectrum was causing crosstalk
What's really cool is cat cables are designed to minimize crosstalk it very high frequencies so they work exceptionally well
The audio in your videos are very low. Hard to hear you.
Yeah, working on that. I do the videos spontaneously. Often without planning and try and do a single take and when things are not quite right with the sound or other bits, if the content is there, I just post rather than retake.
@@DaveRat This is common as most audio engineers use plenty of headroom between 12 to 18db, where most UA-cam audio bloggers compress to less than 3 db of headroom, so unless there is post processing on the audio, it is considerably lower than most content. In digital, 0 is the max before clipping.
Could run a 100w amplifier and a speaker down a pair, or down two pairs using each pair connected in parallel, that might get some crosstalk.
Should also be noted that for fast ethernet, ie anything in use today, only four of the eight RJ45 (or more strictly 8P8C) pins are used, so you can use additional pins for anything you want in theory, and you can build all kinds of nifty breakouts/bodge plugs for stuffing things down Cat5 that don't belong down there.
Yeah, I don't see mixing speaker level with mic level as something that would be implemented, but I do like the extremes. Maybe I will do a video on using cat5e for 70 volt systems
@@DaveRat it would be a baaaad idea on many levels, but I do wonder *how* bad the crosstalk leakage would be. Worst case would presumably be mixing unshielded vinyl cartridge phono with... I dunno... an arc welder? PWM power for something? Feed line for a high power radio station antenna? Power supply for a neon light?
Torture / stress tests often reveal features which are difficult to observe under normal operating conditions.
@@DaveRat I have seen people using Cat5 as speaker cable, and have considered hijacking cat5 network wall boxes in office buildings / DIY recording spaces for an analogue talkback mic & speaker, so while it's not a good idea, it might have some applications.
Gig Ethernet and faster uses all the pins, and gig ethernet is likely FAR more common that fast (100mbit) or slower today. But back in the day, yes, you could use those unused pairs for any number of things from a whole second run of ethernet to analog signals.
@@everettchris1 Indeed, bear in mind I started on 10base2 coax. Where segregated networks are used, processing audio, video and internet via different switches, 100baseT is not a bottleneck for email, so there can in theory be available bandwidth for some extra analogue lines. This is what I have found myself thinking while eyeballing fixed sockets in walls, and having to use mobile phones to debug media networks from room to room, or floor to floor - could I make an analogue talkback bodge box? All the switches found everywhere can do 10base or 100base as well as gigabit, and 10/100 switches are still in use, and still on sale. Walkie talkies and mobile phones work fine most places, though, so it's remained a beard-stroke rather than a habit.