Index Home About Blog
From: (Floyd Davidson)
Subject: Re: Telephone Hybrid Question
Date: 16 Jun 1996 02:20:02 GMT wrote:

>From 4 to 2 wire
>The signal from the microphone gets induced onto the 2 wire side. However, 
>anything on the 2 wire side gets induced onto the 4 wire side.  This creates 
>echo or "side tone". To control the echo a balance network is put into the 
>circuit.  The balance network matches the characteristics of the 2 wire side, 
>shifts the signal 180 degrees, then induces it back on the four wire recieve 
>side where the two signals cancel each other out (mostly;some side tone is 

The TX side (microphone) splits its power between the 2-wire line and
the balance network.  To the degree that those two match each other,
there is no power sent to the RX (receiver) leg.

>From 2 to 4 wire
>The signal induces itself onto both the receive (speaker) and transmit 
>(microphone) legs.  Due to the characteristics of the coils, the recieve 
>signal will propagate out to the speaker with little resistance (inductance) 
>while the transmit side will meet up with a lot of inductance and be 
>dissipated as heat. 

Resistance and inductance are NOT the same thing at all.  For the
purposes of a basic understanding hybrid circuits, resistance and
impedance are the same thing and inductance can be ignored (but
for a greater understanding of balance networks, both inductive
and capacitive reactance will need to be understood).

Whatever power from the 2-wire line that is sent to the microphone
is dissipated as heat because there is no other conversion
available in that device.  Power that is sent to a receiver will
generate some heat, but the device is specifically designed to
produce motion too (and it moves the air so that we can sense it
with our hearing).  Other types of transducers might generate
light as well.

Let me try to draw one type of hybrid circuit in three different
ways, so that perhaps the functionality is more obvious.  This is
not the best hybrid circuit in the world, or the worst, or the
only one.  It just happens to be easy to visualize what is making
it work.

Basic circuit:

                   *  *
    +--------------)||(---------0  2-wire in/out 0---+
    |              )||(                              |
    |              )||( [B winding]                  |
    0              )||(                              |
                   )||(--+                           |
  output           )||   |                           |
    to             )||   +------0    TX input    0---+
   RCVR            )||*  |                           |
                   )||(--+                           |
    0              )||(                              |
    |  [A winding] )||( [C winding]                  |
    |              )||(                              |
    +--------------)||(---------0   Bal Network  0---+
   (asterisks mark the polarity of A, B and C windings)

Now, redrawn to emphasize input to the hybrid from the TX leg
(from the telset transmitter for example):

    |          |            |*          *
    |          )||          )||        ||(-----0
    |          )||          )||        ||(
    |      [B] )||      [C] )||    [A] ||(   output to RX
    0          )||          )||        ||(
               )||          )||        ||(-----0
  Input        |*           |
  from         |            |
   TX          0            0

    0       2W LINE       BAL NET
    |          0            0
    |          |            |

It is easy to see above that power from the TX leg is equally
split between the 2-W line and the balance net (if they are well
matched to each other), but that transformer windings B and C are
inducing equal and opposite currents into winding A, causing them
to cancel and hence no power from the TX leg is sent to the RX
output leg.

Now, re-drawn again, to emphasize input from the 2-wire line:

    |                 |                |
    0                 |                |
                      |                |
  2-WIRE              0                0
                      TX            BAL NET

    0                 0                0
    |    *   [B]      |  *   [C]       |
                  |  [A]   |
                  0        0

Currents in the B and C legs are in phase (and also unequal) and
therefore input power from the 2-wire line does indeed go to the
RX output (as well as to the TX leg and to the balance net, which
in both cases is wasted power that generates heat).

Hope that helps.

Floyd L. Davidson          Salcha, Alaska

From: Floyd Davidson <>
Date: 14 Feb 2000 22:52:50 -0900

"Harvey Babb (Home)" <> wrote:
>At the telephone instrument, it is a matter of signal economy. The
>signal from the line needs to go to the telephone receiver to make
>sound, not to the microphone to make heat. The signal from the
>microphone needs to go into the line, not back into the talkers ear. For
>the early battery powered, unamplified phones, this was essential to be
>heard. The standards were "set in stone" during this era and will not
>change for as long as there are analog telephones.

Actually it hasn't got much to do with "signal economy".  When a
signal is run through a hybrid, the hybrid has two output ports
and the input power is split between them.  In the typical
hybrid one of those ports is terminated with a resistor, hence
half of the signal power is simply wasted (that amounts to a 3dB
loss, and typically the transformer also has a .5 dB loss, which
means the hybrid loss is about 3.5 dB.  Clearly one could do
better, signal wise, than that!

The reason for a hybrid network in a telephone set is
"sidetone".  That is your own voice being heard in the receiver
as you talk.  There is a problem with levels however.  The
incoming signal from the distant end is expected to have about 9
dB of loss (actually a range of about 6 to 12 dB).  But that
assumes that the transmitter puts out about a 0 dBm send level.

If the receiver has to produce the right volume in your ear when
the distant end signal is received at -9 dBm, and your own voice
is there a 0 dBm, it is literally loud enough to hurt your ears
and make conversation very uncomfortable.  The solution is to
use a hybrid, but instead of allowing it to totally eliminate
the transmit signal from being heard in the receiver, a separate
path is provided around it to give a reasonable level (I'm not
sure, but I think it is either 12 or 15 dB lower than the send
level, hence your own voice will likely be lower in volume than
the distant end.)

The effect without the hybrid can be experienced, and most of us
have heard it.  Just hang two identical phones (so that they
equally share the signal power) onto a phone line and try a
"conference" call with both lines off hook.  Anyone speaking at
a normal volume into one of those lines will be too loud in the
other.  (And note that the level is 3 dB lower, because it is
split between two phones, than it would be if it were one
telephone set with no hybrid, so that would even be louder.)

Floyd L. Davidson                     
Ukpeagvik  (Barrow, Alaska)

Index Home About Blog