1
\$\begingroup\$

I have tried to design a amplifier for an electret condeser microphone. It is basically just a common emitter NPN amplifier.

The problem is the frequency response of the real circuit. I get the expected full response on frequencies higher than roughly 10kHz (scratching, s-sounds, crumpling paper). Lower frequencies are practically non-existent. The response I actually expect is below:

Response

The -3db point (marked) is at around 10 Hz.

What could be the cause? Is there some major design error in my amplifier? Is it possible the microphone response is at fault or my biasing is wrong?

My schematic is below: (V1, C1 and Q1 are all part of the microphone capsule)

Schematic

\$\endgroup\$
  • \$\begingroup\$ Lose C3 and R6, then try again. \$\endgroup\$ – Olin Lathrop Dec 25 '14 at 22:31
  • 1
    \$\begingroup\$ Removing C3 & R6, I get less roll-off and it starts at ~1khz now. Overall much flatter, but still not right. Also there is obviously much less gain - too little to be usable. What was the problem with C3/R6 and how do I get higher gain with better response? \$\endgroup\$ – tehwalris Dec 25 '14 at 22:45
  • \$\begingroup\$ Possibly the rest of my bad bass response is from the capsule. It seems to respond to high frequencies equally from all sides, but low frequencies only straight on and closer up. I guess it is a directional mic capsule, but could the directional response vary so much with frequency? \$\endgroup\$ – tehwalris Dec 25 '14 at 23:32
  • 1
    \$\begingroup\$ How are you generating the acoustic test signal? \$\endgroup\$ – Brian Drummond Dec 26 '14 at 0:41
  • \$\begingroup\$ I used a synthesizer app on my phone to generate sine wave outputs through it's speaker. To monitor I used my scope and my PCs sound card in listen mode. With both the response dissapeared/became inaudible when the phone was to the side of the capsule. \$\endgroup\$ – tehwalris Dec 26 '14 at 11:24
1
\$\begingroup\$

Ratio R(collector)/R(emitter) roughly sets the gain. Without R6+C3 your circuit has very low gain, ~1.5. Make R5 say 300 ohms and put R6+C3 parallel to R4 and set R6 equal to R4. Adjust R1 to have 1/2 Vcc on collector.

The purpose of R6+C3 in the current circuit is to bypass AC and block DC; it necessary to keep circuit output reliably at desirable voltage - no matter what the gain you won't need to fiddle with bias resistors. You may try C3 1000 uF, may help.

\$\endgroup\$
  • \$\begingroup\$ What you suggest - moving R6/C3 parallel to R4 (with R6:=R4) - would mean that for AC signals we see R4//R6 or (1/2)R4 at the collector which according to your text would halve gain for AC. Is this what you wanted? Why? \$\endgroup\$ – tehwalris Dec 26 '14 at 11:29
  • \$\begingroup\$ I suggest it as an experiment: it should cause a bump in the lower frequencies, a kind of bass boost, because the capacitor will conduct worse at lower freqs and gain therefore be higher. \$\endgroup\$ – Ilia Dec 27 '14 at 2:58
0
\$\begingroup\$

A few remarks and suggestions:

  1. The directional behaviour of your capsule seems odd. Commonly directivity increases at higher frequency, in your case it seems the opposite. Do you have a second unit? And (....): why are in your circuit diagram some lines red and others blue?
  2. Did you measure the DC bias at the capsule output/drain? With typically 0.2-0.3 mA the FET drain voltage should be ~2V (3.3 - 0.25mAx5k).
  3. Then, most importantly: did you measure the AC response at the capsules output, feeding the capsule with a decently known acoustic source. Nothing very fancy, but you should make sure to know the audio sound level VARIATION WITH FREQUENCY (so, nothing very absolute) within say +/-6 dB or so. That's enough to decide wheather the AC response is hampered by the capsule or by the NPN-based amplifier!!
  4. Although not of immediate concern: isn't there a problem in your simulation if the capsule circuit is mimicked with the FET without a (very high) gate-to-source resistor (e.g. 1 GOhm)? Many simulators do not like floating sections anywhere.
  5. The NPN stage is a bit inconvenient, with a decent gain (yes) but a uneasy high output impedance (R4=3kOhm); complicates connection to a next object.
  6. NPN Ic~0.4 mA, and so the emitter impedance is ~50 Ohm. At frequencies with Z_C3 being small compared to R6+50 Ohm, the limited transistor beta (say 300) brings Z_base ~ 300 x 0.150 ~45 kOhm, which makes it roughly equal to R1 and R2. As the '45 kOhm' depends on frequency your gain will be affected by it (not that much !!). I suggest to drop R1=R2 to 15 kOhm each and increase C2 appropriately.
  7. Did you check the DC voltages around the transistor?

Hope this is of some help!!

\$\endgroup\$

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.