speaker used as "microphone" #2

Thread Starter

sc07734

Joined Sep 28, 2022
8
All, I found this all thread with the goal to understand and to build a telecom from fundamental components. Thank you for contibuting this info here. After reading the above ideas by all and measurements by #12, I see some important aspects I don't understand and which would help this a more complete discussion:

- Given the speaker's resistance (4 Ohms, 8 Ohms etc), there should be no issues in amplifying the sound by a high-Zin opamp. But as someone mentioned, the speaker may not be as easy to model as a voltage source with Rout=8Ohms. What is the model of a speaker? From theory thevenin/norton equivalent transformations, we can consider a it as either voltage source or current source. But the Ro stays the same in both representations. Let's say Ro=8 Ohms

- #12's testing showed that indeed, the signal output begins to drop if speaker is loaded with >100Ohms, => Zo_speaker (output _impedance_) is somewhere on the order of 10s of Ohms.

- I see that google provides hits with a speaker-to-microphone converter schematics that all use a common-base amplifier stage. If we use a common-base stage, the transistor input impedance (looking into the emitter) =1/gm. If speaker is viewed as a current source, then the Zin_amp needs to be << 8 Ohms, => 1/gm << 8, => gm >> 0.125 S or Mhos. This is because current will flow to a lower impedance node and we want it to flow into the amplifier, not its own Ro. So that could work if transistor is biased with plenty of current for a large gm. Seems feasible, but it's close.

- But why is common-gate current-input amp any easier/better than connecting the same speaker to an opamp? In this case, the speaker is a voltage source with very low output impedance so any basic opamp would do. Opamp Zin >>> 8Ohms (usually 100MOhm or GOhms), so a good power transfer is guaranteed without any design effort.

- What is it about speaker's properties that makes it a better match with a transistor emitter node? Is it its inductiveness, which will make the signal ring/resonate when the speaker sees a capacitive component at the opamp's input? And maybe the low impedance of the common-base will dampen the unwanted ringing?

I would appreciate any feedback. Thank you!

Moderator edit: new thread created from this.
 

Thread Starter

sc07734

Joined Sep 28, 2022
8
This is the thread that I am referencing above:
https://forum.allaboutcircuits.com/threads/speaker-used-as-microphone.73428/

I should also add another reason why the I bring up the question of opamp vs common-base:

* for common base as discussed above:
for a BJT, gm=~Ic/26mV. E.g. Ic=10mA, gm=~400mS. => to achieve gm=10 * 0.125S (see above), we need Ic=~30mA as a bias current, which is quite a lot of wasted, standing power draw. I think it would be even worse with a mosfet. I don't know enough about JFETs I-V characteristics to compare.

* for the case of opamp:
- many typical opamps can have quiescent Ibias= O(1mA) (e.g. TL082 Ib=3.6mA) and deliver current only as needed when amplifying the signal (i.e. using a relatively efficient class A/B stage which has Iout_max >> Ibias), which is much more energy efficient than common-base example above and critical for battery-powered applications. So for this reason opamp is preferred, aside from the concerns I raised in the post just above.
 
Last edited:

MrSalts

Joined Apr 2, 2020
2,350
#12 tested with a 6x8" automotive speaker - a very large surface area. If you use a 3 or 4" speaker, your voice will cause a much greater deflection and higher voltage for an equal resistance.
 

crutschow

Joined Mar 14, 2008
30,768
If you use a 3 or 4" speaker, your voice will cause a much greater deflection and higher voltage for an equal resistance.
That seems counter-intuitive.
A larger cone area should cause a larger deflection, and thus a higher voltage, for a given sound pressure.
I'm quite sure that going down to say, a 1-inch speaker, would not further increase the speak deflection and voltage.
 

MrSalts

Joined Apr 2, 2020
2,350
That seems counter-intuitive.
A larger cone area should cause a larger deflection, and thus a higher voltage, for a given sound pressure.
I'm quite sure that going down to say, a 1-inch speaker, would not further increase the speak deflection and voltage.
Would your body weight deflect one brick more than it would deflect a whole patio?
A person only has so much breath to create noise. One can focus that breath into a 3" cone or across a whole 6x9" speaker. Which will deflect more when you speak the word "hello"?
 

Audioguru again

Joined Oct 21, 2019
5,160
A speaker is driven from an amplifier that has an extremely low output impedance that damps its mechanical resonances.
A dynamic microphone (coil and magnet) has mechanical damping of its resonances.

When a speaker is used as a microphone then it feeds a preamp that has the boomy-sounding low frequency resonance attenuated so that a man's voice sounds like a chipmunk.

EDIT: Use an electret mic instead.
 

Attachments

Last edited:

Thread Starter

sc07734

Joined Sep 28, 2022
8
@Audioguru again : thank you for the explanation! Where did you get these plots? (source)

Follow up question: on the plot the resonant frequencies are exagerrated. Yet you say that if it feeds signal into preamp, the voice will sound like a chipmunk. Won't the opposite be true? Low frequencies will be amplified and 'ring out'?

Secondly: following your logic, is this why feeding the speaker signal into emitter (low Z in) will be better than feeding into a base/gate (hi Z)? To dampen resonant oscillating signals? This is in regards to sample circuits of amps for speaker-as-mic are all common base topologies.


A speaker is driven from an amplifier that has an extremely low output impedance that damps its mechanical resonances.
A dynamic microphone (coil and magnet) has mechanical damping of its resonances.

When a speaker is used as a microphone then it feeds a preamp that has the boomy-sounding low frequency resonance attenuated so that a man's voice sounds like a chipmunk.

EDIT: Use an electret mic instead.
 

Thread Starter

sc07734

Joined Sep 28, 2022
8
@MrSalts : let me break down the logic here.. it takes more energy (here momentum of body at motion) to move many bricks as compared to one. That's because each brick weighs similarly and thus creates a similar friction.

Analogous, with a big speaker, there is a resistance to motion per each square inch of surface, so it would take n-times more energy to displace the membrane of a bigger speaker by same distance as a smaller speaker per each brick.

But I think this logic may not hold through to the resulting output voltages.. A bigger speaker will produce a much higher amount of total charge movement (current flow) when its membrane is moved by same distance as a small speaker (larger coil, larger magnet; more mechanical energy is converted to electrical). Also since the coil is larger, the more current is produced, the higher the voltage will produced as well.

In other words, you get more power out (delivered to a load outside of the speaker) for more energy, which is intuitive. It thus goes back to the more practical question: if we supply the same amount of acoustic energy to the small and big speaker, which one produces a higher output power, delivered to the same load? (i.e. yes the larger speaker's membrane will move less, but the lesser movement is also better at converting mechanical movement to current and voltage). So it's not trivial.

When my work and home life gives me a break in some 'near' future, I'll run experiments at home with some cheap small and big speakers and an oscilloscope.

Would your body weight deflect one brick more than it would deflect a whole patio?
A person only has so much breath to create noise. One can focus that breath into a 3" cone or across a whole 6x9" speaker. Which will deflect more when you speak the word "hello"?
 

Thread Starter

sc07734

Joined Sep 28, 2022
8
@Danko thank you for the reference circuit! This also brings back some sweet memories of when I first learned electronics as a kid, digging through my father's junk boxes which had those pretty plastic orange kt315's. I had an easier time with MP39s and MP42s since they were bigger and easier to handle for a newbie, but I used a couple of kt315 as well. Glad to see people still design with them :)


HERE is example of intercom, where speaker as microphone is connected to emitter of amplifier's transistor.
 

crutschow

Joined Mar 14, 2008
30,768
A person only has so much breath to create noise. One can focus that breath into a 3" cone or across a whole 6x9" speaker. Which will deflect more when you speak the word "hello"?
How do "focus" you voice.
The soundwaves spread out rapidly when you talk.
When you say "hello" everyone nearby can plainly hear it.

So how small a speaker would you use for optimum voice signal?
 

MrSalts

Joined Apr 2, 2020
2,350
Because, for a PA system like this - for example an buzzer system for an
How do "focus" you voice.
The soundwaves spread out rapidly when you talk.
When you say "hello" everyone nearby can plainly hear it.

So how small a speaker would you use for optimum voice signal?
because any time I have seen a circuit like this, where the microphone acts as a speaker - it has been in a. Application like a home intercom system or apartment security system (to buzz guests in) where you have to literally put your mouth up to the speaker to talk (focus the energy). Whether everyone else can hear it is non-consequential since human ears have about 10-orders of magnetude range in sensitivity and any reflected sound will likely be heard by others nearby) - a speaker as a microphone does not have that dynamic range. Proximity (focus of energy is key).
 

Thread Starter

sc07734

Joined Sep 28, 2022
8
Very good point. Please note that a focused beam of sound waves (here: talking with a mouth closed to the intercom) which e.g. has a 3-inch diameter cross-section on a mic-speaker, will have the same incidental total power on a speaker with d=3" as it will on a speaker with d=10". Yes, the same power will move the smaller speaker's membrane with higher amplitude vs. bigger speaker (since larger speaker's membrane weighs more and has higher resistance to motion), but the larger speaker has proportionally larger coil and magnet. Larger coil/magnet will generate more power for smaller movement amplitudes.

So the end result will come down to whether the larger or smaller speaker is more efficient/effective at converting the same incident sound wave power to electrical power. The answer is unclear to me. But it is interesting.

Because, for a PA system like this - for example an buzzer system for an

because any time I have seen a circuit like this, where the microphone acts as a speaker - it has been in a. Application like a home intercom system or apartment security system (to buzz guests in) where you have to literally put your mouth up to the speaker to talk (focus the energy). Whether everyone else can hear it is non-consequential since human ears have about 10-orders of magnetude range in sensitivity and any reflected sound will likely be heard by others nearby) - a speaker as a microphone does not have that dynamic range. Proximity (focus of energy is key).
 

MrSalts

Joined Apr 2, 2020
2,350
Very good point. Please note that a focused beam of sound waves (here: talking with a mouth closed to the intercom) which e.g. has a 3-inch diameter cross-section on a mic-speaker, will have the same incidental total power on a speaker with d=3" as it will on a speaker with d=10". Yes, the same power will move the smaller speaker's membrane with higher amplitude vs. bigger speaker (since larger speaker's membrane weighs more and has higher resistance to motion), but the larger speaker has proportionally larger coil and magnet. Larger coil/magnet will generate more power for smaller movement amplitudes.

So the end result will come down to whether the larger or smaller speaker is more efficient/effective at converting the same incident sound wave power to electrical power. The answer is unclear to me. But it is interesting.
Theorize as much as you want, the question is, have you tried it? I did and we got many more mV from a small 8 ohm speaker than a large 8 ohm speaker.
 

Thread Starter

sc07734

Joined Sep 28, 2022
8
@MrSalts : I appreciate that you are specific here. This is the first time you mentioned that you have actually tested it. And that's a very important point. (also I'm sure you see that I haven't tried it, otherwise I wouldn't have asked these questions).

Btw, I assume all else was equal among the two speakers you compared? (Such as paper cone vs paper cone, not paper vs. rubber - that alone would make a huge difference in output)

Finally, I hope no one gets too bothered by theorizing. Theories drove the creation and improvement of most of the items you and I use and enjoy in everyday life, hand in hand with practical brute-force testing. The Loudspeaker is no exception. I appreciate your patience.

Theorize as much as you want, the question is, have you tried it? I did and we got many more mV from a small 8 ohm speaker than a large 8 ohm speaker.
 

ericgibbs

Joined Jan 29, 2010
16,362

Audioguru again

Joined Oct 21, 2019
5,160
@Audioguru again : thank you for the explanation! Where did you get these plots? (source)
The graph of the frequency response is from a Sony headphone speaker. The response of a cheap intercom speaker is probably worse.

Follow up question: on the plot the resonant frequencies are exagerrated. Yet you say that if it feeds signal into preamp, the voice will sound like a chipmunk. Won't the opposite be true? Low frequencies will be amplified and 'ring out'?
The 12dB peak level of the low frequency resonance of a little speaker is not exaggerated. 20dB is common.
Instead of passing the huge resonance that would probably overload the preamp causing severe distortion or having a tuned bandpass circuit to reduce the level of the resonance, they simply reduce the levels of ALL low frequencies.

Secondly: following your logic, is this why feeding the speaker signal into emitter (low Z in) will be better than feeding into a base/gate (hi Z)? To dampen resonant oscillating signals? This is in regards to sample circuits of amps for speaker-as-mic are all common base topologies.
Yes. But a low input impedance preamp reduces the "speaker as a mic" level a lot. Today an inexpensive but good-sounding electret microphone is used.
 

MrSalts

Joined Apr 2, 2020
2,350
Yes. But a low input impedance preamp reduces the "speaker as a mic" level a lot. Today an inexpensive but good-sounding electret microphone is used.
Agree on the $0.25 electret vs the circuitry and complexity of trying to use the speaker as a microphone. The time we tried it was more as a challenge to understand how a signal could be generated with a class of 11-12th grade students in Engineering Club back when my kids were in the local high school.
 
Top