Tarmo:

[snip]
"Tarmo Tammaru" wrote in message
...
I never thought about coax having a low frequency effect, but makes sense.
Twisted pair telephone wire is considered to be 600 - 900 Ohms with a
large
capacitive component at voice frequencies, by 25KHz it is about 135, and
at
T1/E1 frequencies, which are over 1 MHz, about 100 Ohms.
[snip]

Depending upon it's length and how it is terminated, telephone twisted pair
is
approximately 5 - 10K Ohms and highly capacitive at the very low end near
100 - 200 Hz and only reaches 600 Ohms somewhere in the range of
1 - 2 KHz and then as you noted it continues downward to around 100 Ohms
at over 1MHz. In the usual lengths in which it is deployed by the telephone
companies,
i.e. [1500 Ohm design rules which "reach" to a maximum of 18,000 feet] it is
quite lossy usually having a DC resistance in the neighbourhood of 1500
Ohms.

For modern xDSL systems which span from a few hundred Hz up into the MHz,
the telephone twisted pair can fairly be considered a very lossy broad band
channel with
essentially no constant Zo. Zo varies all over the map! Cecil would have a
problem
calculating his "reflections" without a constant Zo, heh, heh... To
complicate things, such lines
often contains bridged taps, which are open circuit stubs hanging off the
line for "sparing"
purposes, and these stubs causes 1/4 wave "suck out" notches in the broad
band
transmission bandwidth.

Full duplex echo cancelled data transmission by xDSL techniques over such
channels
is problematic to say the least, but occurs in millions of instances daily.
In some xDSLs
the transmitters are operating simultaneously in the same bandwidth from
both ends while
the receivers on both ends have to operate in the presence of those local
transmitters over
18,000 foot distances with high attenuation and lots of cross-talk without
errors.

All of this is accomplished with dedicated DSP processors.

--
Peter K1PO
Indialantic By-the-Sea, FL