Use a speaker as a microphone while useing it as a speaker

[Guest Alun]

It is a well known fact that a speaker can act as a microphone becase when sound hits the diaphram it generates a small voltage. I've see circuits that use a speaker as a microphone like a wakie talkie. With these circuits you push a button to talk but it normaly acts as a speaker.
My idea is defferent, it is possible to use a speaker as a microphone while at the same time using it as a speaker.

I've done it before, many years a go (when I was about 11) I connected two small 8 ohm speakers togeather with a twisted pair and made a small telephone. You spoke into one speaker and your voice came out of the other speaker. I discovered it worked best when the speakers were the same impedance so I learned about impednace matching without even knowing it. Of course it wasn't very loud besause there was no amplification. If I added an amplifier it would've been great but I didn't know enough about electronics to design one but now I think I could, whether I get round to it or not is a different matter, maybe this summer when I finish college.

The diagram below shows the model for a speaker:

Where R is the resistance of the voice coil and V is the voltage generated when it acts as a microphone.

[Guest Alun]

Here's a plan of how I plan to connect the following arrangement to an amplifier - it should make this speaker to the two jobs at once.

The red box represents the speaker Rint is the internal resistance, V is the voltage generated by the surrounding sound waves, and Rs is a series resistor I've added to measure the current though the speaker.

If I make R1 and R2 the same ratio as Rs and Rint and attach a differential amplifier to the points A and B the output signal would equal V the voltage generated by the speaker when sound hits the diaphram.

[Guest Alun]

If you had two of these units linked togeather as an inter-com I suppose some sound could go in at one end be reflected of something than be amplified and sent out of the other speaker and reflected back and so on and so forth.

Edit:
This would cause an echoing sound or even oscilation more than distortion, but this wouldn't be too much of a problem so long as the gain isn't too high.

I don't know about resonance of the speaker though, I think the voice coil inductance and parasitic capacitance is more of a problem. R1, R2  and Rs will always have the same impedance regardless of the frequency but the speakers impedance will vary as the speaker is effectivly an R L C network, mechanical resonance will also affect the impedance a bit too. These factors will introduce some distortion but as I only plan to use this for voice transmition I don't need high quality.

Edit:
Hang on a second arn't these factors a problem with any audio amplifier?
They might be a bit worse with this circuit as the differential amplifier's are comparing an LCR network (speaker) with fixed resistors.

[Dazza]

Hi Alun :),

Interesting project, and you've given me an idea ;).

First of all,do you have any idea of what the frequency response would be, for an average headphones speaker being used as a microphone?

I have thought of this before, trying to use a speaker instead of a microphone for the electronic stethoscope project, but haven't tried it yet, I am thinking that it may have very good low frequency response.

And the idea that you have given me, the electronic stethoscope project could do to things, play music to the unborn child, which is supposed to be a good thing ;D, as well as being a stethoscope 8).

[audioguru]

Hi Alun,
A speaker sounds lousy when used as a microphone due to its very strong resonances.
Cheap intercoms at gas (petrol?) stations use them.
They perform not too bad as a speaker because the extremely low output impedance of an amplifier damps the resonances. Try it with any speaker: tap the cone sharply with your finger. Without a load you will hear the resonance. Short it and you won't hear the resonance and it will feel difficult to move quickly.
Therefore you think that a preamp having a low input impedance would damp the resonances? Sure, but it would also kill its output as a microphone, because air vibrations don't have much power.

I have worked with many intercom systems and most have a manual press-to-talk pushbutton switch, or use "voice switching".
When manual, the speaker is a speaker connected through NC contacts of the switch. When you press the button, the switch disconnects the speaker from the amp and connects it to the preamp.
Voice switching uses a level detector in the speaker circuit. If the level is high enough, the circuit switches between transmitting and receiving. Motorola (now ON Semi) makes an excellent speakerphone IC and has a mic preamp, speaker amp, detectors and voice switches inside. Their MC34018 is listed in my latest Newarkinone (Farnell) catalog.

The best sounding intercoms use an electret mic and a speaker. An electret mic sounds so good and costs so little so why not? The mic and speaker must be placed and isolated correctly or acoustical feedback spoils voice switching. You don't want the intercom to suddenly switch into transmit mode because its mic heard its speaker.

I have seen intercoms without manual or voice switching that use a cancellation circuit to avoid acoustical feedback. The cancellation isn't perfect so their volume is very low. ;D

[Guest Alun]

Hi Alun,
A speaker sounds lousy when used as a microphone due to its very strong resonances.
Cheap intercoms at gas (petrol?) stations use them.
They perform not too bad as a speaker because the extremely low output impedance of an amplifier damps the resonances. Try it with any speaker: tap the cone sharply with your finger. Without a load you will hear the resonance. Short it and you won't hear the resonance and it will feel difficult to move quickly.
Therefore you think that a preamp having a low input impedance would damp the resonances? Sure, but it would also kill its output as a microphone, because air vibrations don't have much power.

I agree, it would sound bad, Rs will increase the output of the amplifier and allow it to act as a microphone as well as a speaker. As you've said the resonances will make a poor quality microphone, the increase in output impedance due to Rs will also increase the affect of resonance when the speaker is acting as a speaker.

I have worked with many intercom systems and most have a manual press-to-talk pushbutton switch, or use "voice switching".
When manual, the speaker is a speaker connected through NC contacts of the switch. When you press the button, the switch disconnects the speaker from the amp and connects it to the preamp.
Voice switching uses a level detector in the speaker circuit. If the level is high enough, the circuit switches between transmitting and receiving. Motorola (now ON Semi) makes an excellent speakerphone IC and has a mic preamp, speaker amp, detectors and voice switches inside. Their MC34018 is listed in my latest Newarkinone (Farnell) catalog.

That sounds very cool, I like the idea of a chip dedicated to this application. :)

The best sounding intercoms use an electret mic and a speaker. An electret mic sounds so good and costs so little so why not? The mic and speaker must be placed and isolated correctly or acoustical feedback spoils voice switching. You don't want the intercom to suddenly switch into transmit mode because its mic heard its speaker.

The chip you suggested uses a separate microphone input, but I agree and I've got loads of those little electret mic inserts lying around and they're not expensive.

I have seen intercoms without manual or voice switching that use a cancellation circuit to avoid acoustical feedback. The cancellation isn't perfect so their volume is very low. ;D

I haven't read the datasheet properly but I haven't seen anything about cancellation ciurcuit so I'll just have to be carefull where I put the microphone.

By the way the circuit on the datasheet calls for a 25ohm speaker but I've neaver seen a 25ohm speaker before and I don't like the idea of using a transformer.

http://www.ortodoxism.ro/datasheets/unisonic/UTCMC34018.pdf

But I'll probably never build this anyway as I don't have any use for it. I've just started thinking about this because yesterday I found the pair of speakers I connected togeather with a twisted pair I made when I was a boy ;D it's still educational and certainly food for thought though.

[audioguru]

Hi Alun,
The MC34018 uses a very sophistocated voice switch, I built one and it worked very well, but your idea won't have any switching between transmitting and receiving. In your case surely the mic will hear the speaker and the other end will amplify it, return it back to the first end and you will have howling sound feedback. A "cancellation" circuit tries to add an inverted signal of its speaker output to the mic preamp but due to the differing wavelengths of different sound frequencies the cancellation won't work very well. As others have said, you will also have echoing.

I worked with teleconferencing and videoconferencing equipment where the problems were much worse. With many mics and speakers in each boardroom, howling feedback and echoes were guaranteed without a method to "break the loop". Less expensive systems used voice switching while the expensive ones used DSP (Digital Signal Processing) acoustical echo cancellation.

The DSP was really smart and analysed the level, phase and delay at every frequency that came out of the speakers and was picked-up by the mics. It also adapted itself to changing conditions like a door being opened (the echo off the door suddenly disappeared) or a user's body motion blocking or revealing a few echoes. Occasionally the DSP "lost it" and made weird sounds that you have probably heard on phone-in TV shows. Large boardrooms have a very long time period for their echoes to fade away and therefore need the DSP to have much higher speed and processing power. The best systems have the quickest "convergence time" that you can actually hear as the echoes are gradually cancelled away.

You should add some DSP to your little intercom. ;D

[ECET0purdue]

you mean a PHONE?
lol just playing

[steven]

:)hmmm i used to use the tiny plug type earphones for a microphone back in the seventies and your voice comes out on the tape recorder like your  talking from a radio and allso useing a speaker as a microphone to makes your vioce come out different again, ive tried all this in me younger days and it was intresting . how such things like that did nearly the same work as a microphone

[mvs sarma]

Alun,
if cost is not a creteria, i suggest , you limit the bandwidth to 3khz, and now you try to switch the active limb of the speaker, to  Input / output at a faster rate(by some FET switches using an osc  ( say 4066). ( i mean chopper)
thus if the switch points spkr as out put you listen and whan as input the signal goes to ip of an amplifier. of course you may have to filter the osc frequency from the audio O/P .

i accept it as a concept  and i am sure you can take it ahead-- all the best-- i shall be happy to get feed back from  you

sarma

[audioguru]

You can't multiplex a speaker as a microphone. A speaker needs the extremely low output impedance of the output of an amplifier across it to stop it from moving after it was playing sound. Because a speaker has resonance, it sounds bad as a microphone.

[eurobuskers]

one should NOT say "one cant do this or that."  :o
I think its a cool idea, who cares if it sounds like shit, the Idea has not been done yet, so it should be tried. If it works then work on the quality. I think the first battery was a lemon with two strips of metal, once it was found to work it was improved on.
maybe some kind of filtering with diodes may be a good start, the signal into the speaker is going in one direction so the signal made by the speker as a mike can be taken by a diode looking in the other direction. some kind of negating cct would be needed in order to remove the output from the amp and leave only the signal from the mike/speaker. even if it doesnt work, it's fun to imagine ideas of how it could be done. good luck, I'll keep an eye on your progress. ;)

[audioguru]

In my electronics career, I worked with many intercom systems from cheap and lousy ones to excellent ones. The cheap and lousy ones used the speaker as a microphone. The resonance of the undamped speaker made speech very difficult to understand. When electret microphones were used, the sound quality was excellent.

You cannot use a diode to damp the resonance of a speaker that is trying to be a microphone because resonance is AC.

It was fun for me to try out the very 1st full-duplex boardroom tele-conferencing system and add improvements to it. Then other competing systems were even better and again it was fun to program its dsp so that it cancelled echoes and howling feedback.

[eurobuskers]

I use skype, the speaker is just next to the micro and there is no feedback. there is a distance where a micro ans speaker will howl, therefore maybe the fact the sound I make has to go through the net and back before it is fed-back stops the feedback getting to the micro as it has to treat other sounds. with a delay the same thing could be done.Maybe your friends will then answer the questions before you have asked them ;D

[audioguru]

What is a "skype"?
If you put a speaker next to a microphone and it is connected to the same thing at the other end, then they will make long distance acoustical feedback howling and very strong echoes.
Here is a full-duplex intercom that uses a phase-shift network to cancel acoustical feedback howling:

[mvs sarma]

i understand audioguru and thanks to Eurobuskers--
i really feel like realising it-- in telecom we have something called 4port hybrid-- ofcourse a pasive network. converting 4 wire to 2 wire -- if one can recall the principle involved there-- it is not a problem realising a speaker receiving signal from an adjecent port and the same working as a mike pumping into the adjecent ports as looked from the speaker-- let me analyze-  call the four ports N, E, S, &  W

same as north,east etc

now the recived  output may be fed from west port -- it will be distributed to N&S but minmum (null if balanced) to E port.
the speaker could be imagined on N port  -- next when speaker ,working as a mic will output small signal from N port to equally distribute to E &W ports.

the signal towards W is wasted --while the one towards E port could be amplified and with necessary feedback and still be used with minimum disturbance from in coming signal-- ofcourse it calls  for careful balancing of the hybrid network-- detailed explanation could be found in transmission technology of telecom fundamentals--

these things are like challenges to youth and few of them do realise- the immposibles (so far so)

all the best

sarma

[audioguru]

Hi Sarma,
I have adjusted the balance of telecom hybrid networks many times. It is a big compromise with the capacitance, inductance and resistance of each telephone line often being different and with the conditions changing. You are lucky to obtain a max attenuation of unwanted signals of only 20dB which isn't enough to stop howling in a speakerphone or a boardroom tele-conferencing system.

[mvs sarma]

dear audioguru,
being a retired engr of telecom from india, i recollect that we were able to achieve a balance of say 40 db-- i do understand -- but i was only trying to think of a way loudly -- i do realise that we have to do a lot towards perfection. if something works and with a lot of zeel and patience  it coud be possible -- how ever there are other straight forward and easier models of desogning -- one need not of course stick to a single speaker doing bothe the jobs -- ok  thanks for the reply please

sarma

[audioguru]

Hi Sarma,
Most of the intercom systems, speakerphones and boardroom teleconferencing systems that I worked with used a fixed hybrid 2-wire to 4-wire converter plus half-duplex voice switching to avoid acoustical feedback. The newer more expensive systems used a fixed hybrid 2-wire to 4-wire converter plus DSP echo cancellation for pretty good full duplex.
Almost all of them used electret microphones.

[eurobuskers]

if the circuit between the 2 speakers (that is one for each person communicating) is switching between transmitting and recei  8)ving at a very high speed (10k for example) the problem could reduced assuming we know the distance between the 2 people?

[audioguru]

If the speakers and/or microphones at each end were in rooms without any echoes and the speakers don't have any resonances, then switching back and forth at a high audio frequency will stop acoustic feedback howling.
But in our world, rooms and speakers have echoes and resonances.
Distance adds delay which also must be cancelled.
An acoustic echo canceller circuit is perfect for the job of eliminating howling and allowing each person to speak at the same time (full-duplex).

[gogo2520]

Hello
      Interesting read. I was wondering what are Q1 and ic1 in audioguru circuit
                                      thanks you
                                        gogo

[audioguru]

I was wondering what are Q1 and ic1 in audioguru circuit

Q1 can be any little NPN transistor.
The project uses a little bridged amp IC but an LM386 amp would work fine (different pinout).
The project is here: http://www.redcircuits.com/Page78.htm

[gogo2520]

Thank You
      audioguru
       
            gogo