How does a telephone hybrid transformer work?

Suppose the following circuit:

I can see how an incoming signal (Kim's voice) wouldn't be coupled with the lower pair, but I can't see how an outgoing signal (Ron's voice) only goes through the lower pair.

When the signal from the upper pair reaches the hybrid then one part of it goes to the 2 wire telephone line (seeing a line impedance of Z1) and the other part (right inner inductors) does the same except it will 180° out of phase when it reaches the lower pair (because the impedance Z0 perfectly matches Z1). As such Kim's voice won't travel back.

On the other hand, I don't understand how a signal coming from the left phone would only travel through the lower pair.

• Kim's voice actually would be coupled into the lower pair. The trick lies in the circuit around Z0, and the dots at particular points on the coils. The dots tell you which way the coil is wound. If the dots on the two coils of a transformer are together then the output is in phase with the input. If they are at opposite ends, then the output is 180 degrees out of phase. I'll have to let someone with a better understanding explain the hybrid transformer in detial, though. – JRE Sep 21 '15 at 12:58

On the other hand, I don't understand how a signal coming from the left phone would only travel through the lower pair.

I've concluded that some of the signal from the telephone (Ron's voice) does reach the transmit amplifier (the source of Kim's voice) BUT this is just not that a big deal because, the important thing is that Kim's voice (port 2) does not get transmitted to port 3 (The receive path for Ron's voice).

Here is a better diagram showing the impedances not shown above: -

Also here is what wiki says about hybrid transformers: -

Double transformer hybrid

When both the 2-wire and the 4-wire circuits must be balanced, double transformer hybrids are used, as shown at right. Signal into port W splits between X and Z, but due to reversed connection to the windings, cancel at port Y. Signal into port X goes to W and Y. But due to reversed connection to ports W and Y, Z gets no signal. Thus the pairs, W & Y, X & Z, are conjugates.

Note 1 - Port Y is the balancing impedance shown as Z$_B$ in the top diagram in my answer.

Note 2 - Port W is the telephone line

Note 3 - Ports X or Z are interchangeable as the transmit output or the receiver input.

• Thanks. I guess that this sets it then, since it clearly says that a signal into W (1 in my figure) does appear in both X and Z (2 and 3 in my figure), as you said. – Ant Sep 21 '15 at 14:33
• @Ant Yes I'm convinced now - I did the theory so long ago I struggled and was pleased to find the wiki page to confirm this. – Andy aka Sep 21 '15 at 14:34