Audio Feedback - Why doesn't it occur?- being Mic and Speaker signal in same line ate same time?

Thread Starter

Adrenaline

Joined Apr 4, 2018
2
I am NOT talking about feedback using "the air" as medium but directly on the (electrical) line - MY QUESTION IS:

How (or why) doesn't feedack occur, since the microphone and speaker of a door answering system or even on a telefone land line, have the microfone/speaker signal travelling in the same (electrical) line at the same time?



(PLEASE TAKE IN MIND that telefones exist since way before complicated electronics like "real time feedback attenuators" existed - How was this kind of feedback - direclty on the electrical line which the signals travell on -avoided in the early phones since in those days no complicated electronics was used as well)

Although I am an Enginner my area as nothing to do with electronics nor audio - but you may (please) anwer in an "enginner kinda way"


PS - One idea, for instance, could the Mic and Speaker signals be of phase in any sort of way that is really simple and may be a simple rule of physics that we usually don't talk about?
 

crutschow

Joined Mar 14, 2008
23,502
It's actually fairly simple.
The POTS telephone used a hybrid coil to convert the 4 wires from the mic and the receiver to the 2 telephone wires, which prevented the mic from causing feedback oscillations to the receiver/speaker.
Rather ingenious.
No electronics needed. ;)
 
Last edited:

Audioguru

Joined Dec 20, 2007
11,251
Loudspeaking (mic and speaker) intercoms and speakerphones have a long distance problem where you speak, the sound comes out the speaker at the other end and is picked up by the mic over there and comes back to your speaker producing feedback sounds. It is prevented by a Half Duplex method:
1) Press to talk, when your speaker is disconnected and release to listen when your mic is disconnected.
or
2) Voice switching, when a circuit disconnects your speaker when you talk and a circuit at the other end disconnects their speaker when they talk. This causes problems when noises cause inappropriate switching and when both ends talk at the same time.

Modern speakerphones use a digital echo canceller circuit to prevent feedback and sometimes ( I laugh each time it screws up) people at both ends can talk and listen at the same time which is called Full Duplex.
 

AnalogKid

Joined Aug 1, 2013
8,138
One idea, for instance, could the Mic and Speaker signals be of phase in any sort of way that is really simple and may be a simple rule of physics that we usually don't talk about?
That is it, and in telecom circuit design we talk about it all the time. Search for 'telephone hybrid circuit" to see both the transformer version (post #2) and solid state versions using either transistors or opamps. Solid state versions are in all modems. I did a deep dive on this and found patents going back to the 30's.

The basic idea is that the outgoing audio is available in both normal (0 degrees) and inverted (180 degrees) inside the instrument. In a modem the two are used to (try to) achieve 100% cancellation at the input of the receiver circuit, so the receiver sees only the incoming audio. In a phone the cancellation is less than 100% - on purpose. The receiver circuit sees the bidirectional audio on the line, and combines that with about 80% of the outgoing audio signal that is 180 degrees out of phase with what is being transmitted. This lets the talker hear himself in his own earpiece. Because people naturally adjust their speaking volume so what they hear sounds "right", the phone company has to do very little automatic gain adjustment because the talker does it for them. The small amount of fed back audio is called sidetone. It is in all land line phones, but in almost zero cell phones. Cell phones do not use full duplex audio, so there is no need for cancellation of things that never are combined.

The phase shift also prevents the handset from completing a feedback loop because there is no in-phase reinforcement of the audio signal.

One way to do the in-phase / out-of-phase trick is with one transistor, called a phase splitting circuit. Here is an image from the 40's. As you adjust the pot (item 15) from bottom to top, the signal goes from 100% in phase, through zero signal, to 100% out of phase.

US2441334-0.png

Here is a transistor version in a hands-free intercom:

Full-duplex Intercom Circuit diagram[6].gif

In a data modem, near-perfect cancellation is important. National Semiconductor (or course) had something to say about that.

ak
 

Attachments

Last edited:

Yaakov

Joined Jan 27, 2019
1,601
It's actually fairly simple.
The POTS telephone used a hybrid coil to convert the 4 wires from the mic and the receiver to the 2 telephone wires, which prevented the mic from causing feedback oscillations to the receiver/speaker.
Rather ingenious.
No electronics needed. ;)
The hybrid was one of those really amazing things I had to figure out when I was a kid. It was very mysterious, and really, it's quite clever and made telephony very, very successful.
 

AnalogKid

Joined Aug 1, 2013
8,138
One thing Western Electric *really* understood was wire. For grins, check out the circuit for the original Touch Tone keypad:

16 tone pair combinations
1% frequency accuracy
No long-term or temperature drift
Two multi-tapped inductors
*Zero* calibration or production adjustments
Exactly ***one*** transistor.

ak
 
Top