Date: Sat, 13 Apr 91 00:10 PDT
From: John Higdon <john@zygot.ati.com>
Organization: Green Hills and Cows
Subject: Re: Phone Audio to RCA Jack

jartel!compsm!rlg@nosc.mil (Randy Gregor) writes:

> In article <telecom11.279.12@eecs.nwu.edu> is written:

> > Has anyone ever come across a device which would allow one to take the
> > line level out from an RCA jack to a telephone line and to take the
> > audio from a phone to a RCA jack?

> Get an old phone with a _carbon_ mic (the pop-out kind).  Remove the
> mic.  Connect the audio to the mic contacts, and adjust the output
> level.

and

"Paul S. Sawyer" <paul@unhtel.unh.edu> writes:

> If your application is as simple as it sounds, the Bogen WMT-1 comes
> to my mind -- bridging transformer, with RCA jack <-> screw terminals.

That bridging transformer is most important. Never, but never connect
any unbalanced audio equipment directly to any telephone or other
device directly connected to the network unless you are positive that
isolation is provided within.

One of the first commandments that we telco experimenters learned was
"thou shalt not unbalance or introduce foreign battery on a telco
line". If you connect unbalanced audio equipment without a transformer
for isolation you will at best probably introduce hum and noise on the
line and at worst introduce a safety hazard with unanticipated leakage
currents. While the connection of a battery operated device MAY work
OK without isolation, it is better to just obey the rule of making
sure that your connection is isolated and balanced.

If you leave such a device connected during the automated testing
(done usually at night), a trouble ticket will be generated and you
may have to deal with a repair man showing up. In the old days, this
would have been tantamount to a visit by the FBI in our minds, but
even though times have changed you do not need the hassle.

Whatever you do, do not indiscriminately connect stuff to the handset
transmitter contacts.


        John Higdon         |   P. O. Box 7648   |   +1 408 723 1395
    john@zygot.ati.com      | San Jose, CA 95150 |       M o o !

From: floyd@tanana.polarnet.com (Floyd Davidson)
Newsgroups: comp.dcom.telecom.tech
Subject: Re: 2wire, 4 wire
Date: 19 Sep 1998 11:46:55 GMT

Grimm <grimstev@bobcat.ent.ohiou.edu> wrote:

>A question for all you telecom gurus.  Concerning 2 wire/4 wire circuits.
>I'm having trouble understanding what voltages flow over the respective
>wires in a 4 wire circuit as opposed to a 2 wire circuit.  Let's take the
>2 wires for transmit in a 4 wire circuit for example.  Standard digital
>transmission says a positive bit will be +/- 3 volts and a 0 bit will be 0
>volts.  I can understand how either 0 or 3 volts can flow over a single
>wire, but what's the point of a second?  Does digital equipment just read
>the data from both and take a "best" reading as to what should have been
>transmitted???  Kind of like error redundancy??
>
>Thanks
>Steve

At least a couple of followup articles have mentioned that on a
given pair of wires the signal is "balanced" in that it is the
voltage difference between the two wires which counts, not the
voltage from either of the wires to ground.  Perhaps I should
expand a little on the significance of that because it is
important to understanding why it is done that way, and the
results a truly dramatic!

The voltage difference is between the two wires, so current flows
down one of them and back on the other.  That is referred to as
"differential mode" as opposed to "common mode" which would be
the voltage difference between either wire and ground, and where
the current flows down the wire but returns through the ground.

If the wires are very carefully "balanced", which means that
electrically they are identical, then any currents induced into
one of them will be identically induced into the other wire too.
The two wires would have, in common, the same voltage to ground,
and no difference voltage.  So the signal, which is a
difference, would not be affected at all!  That is referred to
as "common mode rejection", and for well balanced wire pairs is
very very great.

And that is why telephone wires are twisted pairs, because by
twisting them at short enough intervals it guarantees that whatever
electrical exposure one wire has, so does the other.  They can
be run relatively close to power lines, for example, and pick up
considerable 60 Hz AC from them... but we don't hear it on our
phone line unless it becomes unbalanced for some reason.

That also explains why any mention of 60 Hz hum or buzz on a
telephone line (all of which run somewhere near some power line)
and the immediate response is to say that the line has something
causing it to be unbalanced.  Common problems are poor or wet
connections at junction boxes, or staples cutting through house
wiring.  Damaged insulation in a damp environment is another
major cause of unbalance.  Whatever the cause, the result is
usually hearing a 60 Hz hum.  But another thing it can cause is
hearing signals from other telephone lines, which are also
induced into the all of the cable pairs in a cable, but are
normally canceled out by common mode rejection.

  Floyd


--
Floyd L. Davidson                                floyd@ptialaska.net
Ukpeagvik (Barrow, Alaska)                       floyd@barrow.com

From: floyd@tanana.polarnet.com (Floyd Davidson)
Newsgroups: comp.dcom.telecom.tech
Subject: Re: 2wire, 4 wire
Date: 20 Sep 1998 18:06:11 GMT

jim <jimw@nospam.mcsi.net> wrote:
>Wow,  that was a good explanation.  That was the first time Iv heard of a
>reason for the twists.  I always thought that maybe they where twisting them
>for the magnetic fields.  Never thought about equal exposure.
>
>
>Floyd Davidson wrote in message <6u05jf$472@enews4.newsguy.com>...
...
>>
>>If the wires are very carefully "balanced", which means that
>>electrically they are identical, then any currents induced into
>>one of them will be identically induced into the other wire too.


The key words for telephone lines are really "transmission
line"!  And the best place to find a good theoretical discussion
of how they work is any good book on transmission line theory,
which is usually found in books on radio rather than telephony.
For non-engineers the ARRL "Radio Amateur's Handbook" for any
year one can find is probably the easiest to read and understand
description.

Twisted pair cable is a variation on open wire transmission
lines, much the way twinlead for TV is also.  The amount of
twist changes the characteristics, and the most significant
difference is that the tighter the twist the higher the
frequency where good balance is maintained.  Typically telephone
cable has a twist every 18 to 30 inches or so (I don't remember
the exact numbers).  But twisted pair used on patch panels for
either digital signals or the lower frequency levels in analog
carrier systems are twisted much tighter, at as much as several
per inch!  The most significant difference between telephone
drop wire and CAT3 and CAT5 is the increased twist for higher
frequencies.

However, a tighter twist takes more wire and is physically
larger, and therefore telephone cable uses as little twist as
possible.  Different pairs are also twisted a little more or
less too, to prevent them from laying exactly alongside each
other and negating the effect of the twist.  And, in large
cables, there are bundles... and the bundles are also twisted!
(Actually, it is more of a swirl for the bundles.)

  Floyd

--
Floyd L. Davidson                                floyd@ptialaska.net
Ukpeagvik (Barrow, Alaska)                       floyd@barrow.com

From: floyd@tanana.polarnet.com (Floyd Davidson)
Newsgroups: comp.dcom.modems
Subject: Re: Local loop parameters
Date: 30 May 1999 11:26:08 GMT

Icky Pic <7dkhuff@ttc-cmc.net> wrote:
>alcazar3@my-deja.com wrote:
>>
>>
>> Why does 60 hz  need attenuation ?
...
>current. If this induced AC voltage can find a path to ground there will be
>an AC current flow. If there is an induced AC current flow and you pick up
>the phone, that same AC current will flow through your phone receiver
>(speaker in the handset) and presto, you have a very annoying 60 cycle hum.
>I guarantee, without the C Message filter, you would hear the hum allot
>louder than you could hear the person you are talking to. Imagine having 5
>to 30 volts of 60 cycle induced current and the normal voice voltage is .5
>to 2 VAC (I think).
>I wish there was a way to hook a speaker up to a cable pair or remove the C
>Message filter (or whatever it is, Floyd) and see how loud the hum would
>be.

The above figures for induced 60 Hz AC voltage on telephone
cables are correct; however, the reason you don't hear a 60 Hz
hum on a good telephone line is due to 1) the use of twisted
pair cable, and 2) an extreme effort is made to maintain balance
on telephone circuits. THAT is what eliminates the 60 Hz from
what you can hear on the phone line.  In a plain old telephone
set there simply is no filter at all to get rid of or even lower
60 Hz tone levels.

The 60 Hz voltage is induced into both sides of the cable pair
equally.  (The use of twisted pair cable is the main reason that
is true.)  Since any voltage induced into the ring is also
equally induced into the tip, a telephone set connected between
tip and ring will not hear the 60 Hz induced voltage at all.

"Longitudinal balance" is the telephone industry's measure of
how effectively diminished an induced voltage will be (it is
called "common mode rejection" by the rest of the electronics
industry), and generally cable pairs are better than 60 dB
(minimum spec is 58 dB).  When that is combined with the measure
of insensitivity of the telephone receiver and the human ear
(C-Message Weighting) to low frequencies, which is about 50 dB
at 60 Hz, the result is that 60 Hz noise induced into a
telephone cable has the effect of a tone which is reduced in
volume by about 110 dB.

The level at which an induced 60 Hz tone would be audible then
is about +50 dBm, which is about 135 VAC!  Induced voltages of a
little less than half of that are not unheard of (55 VAC).  That
would be perhaps 4 dB less, at about 46 dBm.  Such lines work
well, as long as the longitudinal balance is maintained...

  Floyd




--
Floyd L. Davidson                                floyd@ptialaska.net
Ukpeagvik (Barrow, Alaska)                       floyd@barrow.com
     North Slope images: <http://www.ptialaska.net/~floyd>

From: floyd@tanana.polarnet.com (Floyd Davidson)
Newsgroups: comp.dcom.telecom.tech
Subject: Re: Category 5 wiring tecnique questions
Date: 10 May 1999 03:23:50 GMT

John Fricks <jfricksNOTelvis@bellNOSPAMsouth.net> wrote:
>What you're saying is contrary to my expectation. Please explain.
>
>If pairs are individually twisted, and if the twist is maintained to the
>connection point, then common mode fields cancel and there isn't any energy
>left to induce current into a nearby pair.

Your perception of what happens is slightly flawed.   The field is not
cancelled, but is in fact induced into each and every cable pair.  Every
field is, in fact.  So there is noise from the other cable pairs and from
power lines, and from whatever else happens to exist too.

But each field is induced equally into each wire of a pair.
When a receiver looks for a signal on that pair it looks for a
difference between the two wires, and since noise is induced
equally there is no difference between the pair of wires as a
result of the induced "noise" field.

The equality of course varies!  And noise is not detected to the
degree that the equality, or balance, exists.  The higher the
frequency the tighter a twist must be to help maintain the
balance because the twist must be a very small percentage of
the wavelength of the induced field.

At voice frequencies a twist every 20 inches or so is just fine.
At 1.544 Mbps T1 speeds (778 Khz) the twist has to be tighter or
the noise is going to be greater (and since voice quality cable
is commonly used, the noise is in fact greater).  For 100 Mbps
ethernet a very tight twist is needed, which you can clearly see
with CAT5 cable.

>If these tightly twisted pairs are further separated from each other (as
>would be the case if the pairs are no longer tightly bound by a common
>sleeve) the induced currents are further reduced.
>
>So, removing the sleeve (while maintaining the twist) is good!

Well...  inside the sheath those pairs are also twisted around each
other, which helps prevent loss of that equality (it maintains balance).
Removing the sheath will in fact improve the balance *if* an extreme
amount of care is taken to avoid close proximity to metal objects that will
capacitively couple to the cable and if all sharp bends are avoided.
Generally it is easier to maintain better balance by not removing more
of the sheath.

>Show me the numbers (the near end crosstalk numbers)!

I certainly don't have any figures, but the theory has been in
existance for a lot of years.

  Floyd

--
Floyd L. Davidson                                floyd@ptialaska.net
Ukpeagvik (Barrow, Alaska)                       floyd@barrow.com
     North Slope images: <http://www.ptialaska.net/~floyd>