Re: TV and sound over CAT5

[Nigel Orr <nigel.orr@xxxxxxx>]

At 13:03 26/07/99 +0100, you wrote:
>I'm hoping to run output from a soundcard to the input of a hifi
>amplifier over a CAT5 cable pair. Given this is not the full amplified
>power can I expect this to work with reasonable quality, or do I need
to
>be more clever in sending the signal?

The short answer is "quite likely it will work just fine".  The
long answer
follows:

There are 2 likely potential problems, and 2 methods of finding whether
they will occur or not.  The problems are
1) loss of high frequency, due to high source impedance of the soundcard
driving a long cable.  Probability is about 5% or maybe less, with modern
equipment.
2) mains 'hum'- 50Hz noise induced from nearby mains cables or noisy mains
equipment (compressors in fridges etc).  Probability is about 90% that you
will have some sort of audible noise of this kind.

The 2 methods of finding whether these will happen or not are:
1) Years of detailed analysis
2) 'Suck it and see'

I'd recommend 2)- assuming you want whole house audio, you definitely need
the cable, and you can add line drivers, transformers etc at each end to
reduce or remove both problems.  Connect it up, if the noise level is
acceptable and the sound quality is fine, then go with it.  If it sounds
muffled or there is too much background noise, look at solutions to that.
If I was specifying a professional system, I'd assume there would be both
problems and specify balanced line drivers, but at home you can 'evolve' to
solve whatever problems you find.

>Will the other cable pairs suffer
>from induced noise even with low level signals or can I use them for
>other purposes?

Yes they will, and yes you can, if the noise is acceptable on the other
pairs ;-)

> I had thought that the point of using 'twisted pairs'
>was so that they didn't cause/suffer to much interference (but I'm a
>software guy, not a techie!).

It's a bit of a generalisation.  Put (fairly) simply, a wire in a field
(produced by other wires carrying current) will generate a small current.
If you had another wire at the exact same location as the first, and
subtracted its generated current from the first, you would delete all the
noise.  However, 2 wires at the _exact_ same location isn't possible
(although co-ax is remarkably close!), so it is found that by spinning the
wires round each other (a twisted pair), you get a reasonably good
approximation, and it's fairly cheap and manageable.

This all breaks down when the source of the noise is very close to the
cable.  The distance from the source to each wire is now markedly
different, so they don't generate the exact same current, so you will get
some 'left over' noise, and there's nothing much you can do about that.
Does that make sense?

So, anything on one pair in the same jacket will appear to some extent on
all the other pairs on the same jacket.  Question is whether the resulting
noise is acceptable to you.  If one pair is audio Left and one is audio
Right, it'll probably be unnoticeable, if one pair is audio and one is IR
remote data, it will probably be irritating.  If one pair was a microphone
signal and one is speaker (about 60dB difference), the result would
probably be unbearable.  If two of the pairs carry ethernet, the network is
apparently likely to be unreliable if the other pairs are used as well.

As above, it might just work in a particular situation, so at home you can
try it and see if it does, but don't go installing the same system in other
houses because it happened to work in yours!

Nigel

------------------------------------------------------------------------
Was the salesman clueless?
Productopia has the answers.
http://clickhere.egroups.com/click/555



eGroups.com home: http://www.egroups.com/group/ukha_d
http://www.egroups.com - Simplifying
group communications