I am currently trying to make a preamp for an electret mic pair, which are placed physically in antiphase. The main idea is to subtract one mic's signal from another and therefore boost the signal amplitude while reducing the noise.

Caution: in the schematics below the nominals of resistors and capacitors are totally wrong, only the topology of preamp is what matters. The FETs in all schematics are just the FETs that are included in the electret mics.

A typical solution for this would be using a differential amp (number 4 on pic). However, I am not sure if this solution is good enough and if it is better to pre-amp the signal of each mic individually before subtraction. For a single-mic preamp I have found three solutions on the internet (see number 1-3 on pic): a simple non-inverting amp (1), a transimpedance amp (2) and a weird schematic which implies getting rid of the DC offset by precisely calibrating the R8 resistor, not by using a capacitor (3).

My question: should I subtract the microphones' signals immediately (using a differential preamp) or after some basic amplification and if the second option is preferable, which one of the preamp topologies (1-3) should I use? What are the pros and cons of each approach?

Different suggested topologies, see text

  • \$\begingroup\$ If the mics are placed anti-phase, then you just need to sum the two signals. Anti-phase is frequency dependent, though. You can only improve a relatively small frequency range this way. \$\endgroup\$ – JRE Aug 19 '17 at 20:28
  • \$\begingroup\$ @JRE, nope. There are placed in anti-phase, but to reduce the electrical noise of the internal FET amps they have to be subtracted. They are placed in anti-phase so that anti-phase summing (that is, subtracting) and anti-phase placement would cancel each other. As my calculation shows, summing the mics placed in anti-phase or subtracting the ones placed in phase is generally a very bad idea, just as you said. \$\endgroup\$ – sx107 Aug 19 '17 at 20:36
  • \$\begingroup\$ Sorry, brain fart. Signals are antiphase, take the difference to get the sum. It doesn't reduce the noise, it increases the signal. Still frequency dependent, though. \$\endgroup\$ – JRE Aug 19 '17 at 20:44
  • 1
    \$\begingroup\$ Ok. So, you are trying to eliminate electrical interference from some other device from the microphone signal. The audio is still going to get messed up. Anti phase for audio is frequency dependent. Anti phase for 1000Hz is about 17cm. Anti phase for 500 Hz is about 34 cm. Anti phase for 250Hz is around 68 cm. If you are anti phase for 250Hz, you are in phase for 500Hz. You will get canceling and reenforcement for different frequencies depending on frequency and distance between microphones. \$\endgroup\$ – JRE Aug 19 '17 at 22:12
  • 2
    \$\begingroup\$ What you can do with antiphase mics is NOT reduce noise ... the main noise sources are uncorrelated so won't decrease summed in antiphase (absent poor design permitting interference). What you can do is generate alternatedirectional patterns ... cardioid, figure of eight etc - a crude sort of beamforming, so you can(e.g.) point a null at an interfering (acoustic) noise source. JRE is right that it'll be frequency dependent. \$\endgroup\$ – Brian Drummond Aug 19 '17 at 22:18

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.