2-way analog intercom

[elecfan]

Dear all
I am working on a 2-way analog intercom,
I need a suitable 2-channel pre-amplifier with AGC.

I am trying to use a half duplex communication, with channel priority,
in order to avoid laud noise of the speaker

your help will be appreciated
Best regards

 

[KlausST]

Hi,

given informations: Analog, communication, 2-way but half duplex
speaker.. thus I assume we are talking about audio range. But not whether reduced bandwidth for an alarm tone, or voice, or high fidelity audio...

My first idea: good old analog telephone. It can even do full duplex and does not necessarily need amplifiers. Over long distance.

****
I guess you need to give more details.


Klaus

 

[crutschow]

Wireless phone sets often have an Intercom feature and you don't need any wires between them.

 

[Audioguru]

I worked with Philips, Stentofon and TOA large high quality intercom systems. They were all half-duplex that used sound level for communication direction and none of them used AGC.
Years ago, Motorola introduced two intercom ICs that were also half duplex but they averaged then cancelled continuous background noise so that it did not affect communications direction.

Later I worked with high quality boardroom telephone-conference systems. They began to use an echo-cancelling IC for full duplex and sometimes it got confused and made funny noises.

Today, Polycom intercoms use an echo-cancelling IC for full duplex but they are VERY expensive.

 

[elecfan]

The system is going to be designed for over counter communication .
In one side close distance of speaker and microphone is required.

 

[KlausST]

Is this all info you can give?

Klaus

 

[ghasemghasem]

Hi,
I am designing a two-way intercom similar to what you have discussed, and I am facing a problem that there is unwanted feeback from microphone to speaker on the side that microphone and speaker are close together(customer side). Basically my design is very simple, I have two pre-amplifiers in which the microphones are connected at one side, and the other side goes to power amplifier which is a TPA3110. In between I have placed a volume control which actually brings down the gain, so that I can use two watt speakers on the output of TPA3110. I have a very big problem that I cannot stop the microphone-speaker squeal when I turn up the volumes. And if I do't turn up the volume the system is barely audible and thus not usable.
Any suggestions on how to fix this problem, or if you can think of a better affordable system ?
Best regards

 

[dick_freebird]

An alternative to AGC is "compander" (audio compressor /
expander), back when coms were still analog these used
to be common in the mic signal lineup. Not sure whether
any such function chips survive today but worth a look.

Limting bandwidth to useful speech range could tamp
down high frequency feedback (if the feedback-tones
are far enough above speech range of interest).

 

[elecfan]

Is this all info you can give?

Klaus

for the moment I am using MAX9814 as preamp and TPA3110 as amp.
is almost similar to what ghasemghasem is explaining, in his/her post above.

the idea was to see if using a dual channel preamplifier with channel priority, can help to avoid the loop in the circuit.
this should help to turn off one channel when the other is working.

but I do not know if such a chip does exist.

--- Updated Nov 7, 2020 ---

I worked with Philips, Stentofon and TOA large high quality intercom systems. They were all half-duplex that used sound level for communication direction and none of them used AGC.

Years ago, Motorola introduced two intercom ICs that were also half duplex but they averaged then cancelled continuous background noise so that it did not affect communications direction.

Later I worked with high quality boardroom telephone-conference systems. They began to use an echo-cancelling IC for full duplex and sometimes it got confused and made funny noises.

Today, Polycom intercoms use an echo-cancelling IC for full duplex but they are VERY expensive.

Many thanks
could you please write the ICs' numbers from Motorola?
or any other you might be aware of?

 

[betwixt]

That would be the MC33219A which I believe is still in production.
They were used in telecoms hands free circuits so whoever shouted loudest got through and the back channel was muted.

You can also use a ring modulator to add a very small shift in frequency to one of the paths to break the dominant pole in the transfer but it isn't simple.

Brian.

 

[Audioguru]

Most of the Motorola speakerphone ICs are obsolete. I used two MC33018 ICs and they worked very well.
Then they made the MC33118 then later they made the MC33219 and maybe others.
NXP is listed as making some of them and many (fakes?) are sold on ebay.

In PA systems I tried shifting the frequencies to reduce feedback but the talkers and singers said it sounded awful to them to hear it.

 

[ghasemghasem]

I have used the microphone amplifier Max9814 with AGC and for amplification this time I used the TDA2030 which it seems powerful enough to act as a twoway intercom. The system seems to whistle again as I bring the volumes up, since as I already mentioned before, one side of this system is the remote side, which comes in a small tiny box. I have't been able to remove the whistle out of this system, even though I used atleast 5 different speakers and more than that number tiny electret microphones. Kindly anyone can pointout what I can do to remove the whistle, or if not, what system may I use so I do't encounter this problem ?

 

[KlausST]

Hi,

The "whistle" is resonance: amplifier --> speaker --> mircophone --> amplifier -->...
It depends on gain and phase:

* reduce gain: reduce speaker volume, block/reduce sound from speaker to get to the microphone, install a system that feeds back the electrical signal (narrow bandwith) to the microphone signal in order to cancel the (audible) feedback signal from the speaker, place speaker and microphone more far from each other. Mounting method, sound absorber...

* phase: install a phase shifter, try to just interchange speaker plus and minus, install a audio delay line.

Klaus

 

[Audioguru]

Of course you have acoustical feedback when a microphone can hear the speaker.
Without seeing your wiring then a very simple intercom has the mic on the first end amplified and its sounds come out the speaker on the second end but the sounds go into the mic at the second end are amplified and come out the speaker at the first end. Then the mic at the first end hears the sound from its speaker and the sound goes around and around and around.

A cheap old intercom used a Press-To-Talk push button at both ends so that when no button is pressed then no mic is active. A button is pressed at one end activating its mic then the sound is heard in the speaker at the other end but with no feedback because the mic at the other end is not active.

An electronic circuit can produce half-duplex by activating a mic when a signal is sensed from a mic. But it can also produce feedback or cutoff spoken words.

Here is a way to use only 2 wires between ends and it uses cancellation circuits at both ends so that the communication is in full duplex:

 

[Easy peasy]

usually a local opamp is used to subtract the local microphone signal from the line signal leaving only the remote signal into the local speaker - of course if there is too much audio connection from the local speaker to the mic at each end you will form a loop which can squeal if the total loop gain is too high ...

 

[Audioguru]

usually a local opamp is used to subtract the local microphone signal from the line signal leaving only the remote signal into the local speaker - of course if there is too much audio connection from the local speaker to the mic at each end you will form a loop which can squeal if the total loop gain is too high ...

The circuit I posted is old so it uses simple transistors instead of opamps. The trimpots R7 are adjusted for perfect cancellation so there is no feedback howling or squealing.
A similar cancellation circuit (hybrid) can be used in a speakerphone but the various resistance, inductance and capacitance of a telephone line needs the cancellation circuit to be re-adjusted for each telephone call. An echo-canceller circuit does it well.

 

[Easy peasy]

@A_guru, I note that when SW1 is closed the base is AC grounded such that there is no mic amplification, such that a push to open switch is needed for PTT. In the physical world, if the mics and speakers are too close at each end, there there would still be a loop path for an audio signal if the gain is high enough - turning R7 to its lowest point merely increases the mic signal to the amp.
R7 is really a mix of local and remote - local (mic ) content is really useful if head-phones are used - so you can gauge an idea of how loud you are into your mic - of course worn headphones remove the loop squeal effect almost entirely ... as the physical audio path from speaker to mic is usually very low ineed for worn headphones ( but take them off and put on a hard surface - a different matter... ! )

 

[Audioguru]

The switch is not PPT because the circuit is full duplex (both speakers and both mics are active).
Instead the switch is a Mute-The-Mic switch.

R7 is not adjusted to the lowest point. It is adjusted for perfect cancellation of the mic signal to the speaker amplifier because the collector and emitter of the transistor are 180 degrees apart. With perfect cancellation then there is no feedback squealing.

 

[Easy peasy]

One can seethe full duplex, , in each case though, there is a positive reinforcing loop in each local mic and speaker - if the gain is high enough,

again, in a headphone scenario it is useful to have some mic content in the headphones

I like the idea of a muted mic and PTT, as it keeps the channel quiet most of the time - and allows more than two users

 

[ghasemghasem]

Dear All,
As it was suggested, I thought also it is a good idea and designed a delay circuit using PT2399 IC. As it is given on this IC datasheet, a potentiometer can increase and decrease the delay. The circuit worked fine, but it did not reduce the whistle on the channel. I am at this point thinking of putting two echo circuits, one on the forward audio channel and the other on the reverse audio channel (full duplex). Any suggestions are welcome on perhaps redefining the usage of PT2399(maybe a different circuit?), or using a different IC? Does the bucket-barrel delay type circuit work instead? Anybody has had luck with those?