[TowerTalk] 600 Ohm Line

[Jim Brown]

On Thu, 29 Jul 2004 19:02:31 +0100, EUGENE SMAR wrote:

>> Turns out you can't consider the phone line as a lumped component
>> even to get audio to the CO; moreover, the characteristic impedance
>> is complex.  This is why the phone company has to use loading
>> coils.

>Rick is correct.  The copper twisted-pair (TP) is considered a distributed
>reactance element.  The loading coils are actually series inductors
>(remember the 88 mH toroidal coils we all used in the 70's?)  

They go way back before the 70's, but they are LUMPED elements, and they are 
compensating in a LUMPED model. 

>(Remember the
>70's?)  These series inductors cancelled a bunch of the capacitive reactance
>of those parallel conductors on long (> 1 mile) lines so that T-carrier
>devices (1.544 Mbps) data signals could pass from customer premises (usually
>businesses) to the CO.  Otherwise, the phase shift and attenuation of the TP
>would obliterate the T-1 signals.

What's the old line -- "if you remember the 70's you weren't there?" 

>When I worked for a local phone company (AFTER the power company), we were
>using ADSL over TP (duh) to attempt to deliver video.  The biggest hurdles
>we encountered were not so much with the ADSL modems, but with the condition
>of the outside plant (OSP) itself.  

I haven't worked in that part of the industry, but common sense and Transmission 
Lines 101 says that those modems had better be treating that line as the 60-100 
ohm line that it really is of they are going to work at those data rates!  That means 
matching each end to that characteristic impedance. There is, of course, another 
wrinkle -- not all the cable in the loop has the same Zo, so there can be bumps at 
each transition. And, of course, there the loss in the line at the higher frequencies, 
and the noise. 

The AES/EBU Standard for transmission of digital audio specifies 110 ohm cable 
for  balanced transmission and 75 ohms for unbalanced transmission, and both 
follow transmission line principles to the letter. I'm a member of the Working Group 
that wrote and maintains that Standard. A 48 kHz sampling rate AES/EBU digital 
audio signal needs 6 MHz of bandwidth, 192 kHz sampling rate needs 24 MHz. 
(Note that this is uncompressed 24 kHz stereo audio with no data reduction). 

The CAT5 (and higher spec) cable used for Ethernet is 100 ohms. 

Jim Brown K9YC