3

Is it possible to connect multiple electret microphones in order to increase sensitivity (gain) of the microphone(s)? If yes, how do I connect them?

Update:
Obviously one solution is to increase the number of amplifiers, but I am curious if multiple microphones can be serial connected - so this is more like a theoretical question (unless it is possible to do it - then it becomes a practical question) :)

Gabriel
  • 793
  • 2
  • 14
  • 33
  • When you say "sensitivity", are you referring to the **gain** (in dB) of a microphone? –  Aug 17 '17 at 16:31
  • Do you want louder, or clearer (less background noise?) Louder is easy, just amplify - no need for multiple microphones. Less background noise is harder, and more mics can help. – JRE Aug 17 '17 at 16:31
  • @JRE-I know that I can add yet another amplifier block. But this will increase the volume, battery drain and it will add additional noise (I try to build a miniature long distance spy device). I thought that adding additional microphones will increase the gain. Plus, a microphone introduces less noise in the system than the amplifier. – Gabriel Aug 17 '17 at 16:40
  • @KingDuken - yes. I want to get a higher voltage from the microphone. – Gabriel Aug 17 '17 at 16:40
  • 4
    No. The sensitivity is a matter of microphone design (physical and electrical.) Piling more of them in parallel (or series) will apply them in ways they weren't designed to operate and it will decrease performance (or worse.) You can look into microphone beam forming or other methods of combining arrays of microphones to improve overall signal or other characteristics. But they are non-trivial and you seem to be asking for only trivial solutions. You might also consider an "ear trumpet" addition. – jonk Aug 17 '17 at 16:45
  • 2
    Shure actually has a small article about connecting microphones in parallel. You can see that [here](http://www.shure.com/americas/support/find-an-answer/can-i-parallel-microphones-into-one-input-of-my-mixer). Funny, because they just updated the article yesterday. –  Aug 17 '17 at 16:46
  • @KingDuken - Thanks for the link (and +1). Unfortunately, the say "Only dynamic microphones may be used", So, your link kinda answer my question. – Gabriel Aug 17 '17 at 16:49
  • "and you seem to be asking for only trivial solutions" - Yes. – Gabriel Aug 17 '17 at 16:50
  • Somewhere here I have a copy of a plan for a super sensitive spy microphone. It is rather large, but the electronics are trivial. One microphone, an amplifier, and a bunch of mechanical stuff. Have to see if I can find it. – JRE Aug 17 '17 at 17:11
  • 1
    [Found it. The shotgun sound snooper.](http://gbppr.dyndns.org/~gbpprorg/mil/shotgun/Build_the_Shotgun_Sound_Snooper.pdf) – JRE Aug 17 '17 at 18:02
  • 1
    By the way: a microphone doesn't necessarily add less noise than an amplifier. And an amplifier doesn't have to be a battery killer. – JRE Aug 17 '17 at 18:53
  • the shure article refers to parallel mics in a paging setup, where only one is used at once - to save mic channels. Not using the mics simultaneously. – danmcb Dec 12 '22 at 09:47

2 Answers2

11

I am a technician at a boutique mic company. I like to build electret mics in my spare time.

You can increase sensitivity, gain, and signal to noise ratio by running multiple electrets in PARALLEL-not series. This works because:

1) each capsule has an onboard JFET. When combining noise from multiple FETs, some of it cancels. Two amps= -3dB, four amps= -6dB of noise reduction [see Small Signal Audio Design by Douglas Self, Routledge 2010].
2) Multiple elements will have lower output impedance, which can mean less noisy makeup gain down the line.
3) More diaphragm area means more sensitivity, which makes sense. It also means less self noise, for reasons that I don't really understand - but that's the principle behind the $3500 Audio Technica AT5047 (FOUR diaphragms!).

If you end up with a signal that is too hot and distorts the next stage, put a 10ohm resistor on the audio output of each capsule before wiring them in parallel. This averages rather than sums the outputs.

Zachary Alan Stewart
  • 146
  • 2
  • 5
  • Welcome to Electrical Engineering. You are starting with good answers. – Grebu Nov 20 '18 at 21:18
  • 1
    will some of the "good" signal (not the noise) cancel also itself? because not all microphones are perfect parallels (I mean to the physical surface they are mounted on), they will generate slightly different signals. – Gabriel Feb 04 '19 at 10:50
  • 2
    @Rigel Yes, comb filtering. It'll mostly be in the ultrasonic, not worth worrying about too much. Also audible-range comb filtering is so common in everyday life, that it's often basically unnoticeable, so even when you suspect there might be issues, it's worth trying regardless. – Siana May 02 '19 at 10:57
  • Given that the output of each FET amp is a voltage source (with some small but real source resistance), I don't see how putting them in PARALLEL is going to leverage the increased sensitivity. You are basically resistively mixing two almost identical signals. The output cannot be greater than the mic with the largest output unless you have active gain in the summing stage. You will not achieve greater voltage output for a given SPL. It isn't possible. – danmcb Dec 12 '22 at 09:19
0

Identical(!) electret microphone capsules in 2-wire configuration can be wired in parallel since they are essentially acting as variable current sources. Note that they will then also draw the added supply current which may or may not become a problem.

In 3-wire configuration (needed for some high sound level applications), the capsule-internal FET is wired as a source follower and acts as a (unidirectional) voltage source. Wiring the capsule outputs (the FET is used as source follower then) in parallel will not form a sum or even average of the combined outputs but rather a maximum. You need to sum them through resistors, and the most robust way is to sum them to a virtual ground point like the negative input rail of an opAmp in inverting configuration. That way, the microphone outputs stay independent.

Problem, of course, being that electret microphone capsules stay operative (albeit not linear) down to quite low operating voltages and currents while opAmps do not. So that tends to end up a rather fuzzy option. If you are operating at higher signal levels, you are likely better off picking comparatively insensitive 2-wire capsules and wiring them in parallel than dealing with the headaches the 3-wire configuration with the capsule FETs as source followers create.