1
\$\begingroup\$

I'm trying to prototype a hobby project on a breadbaord.
I'm using an electret microphone + opamp to feed MCU ADC with audio signal. When I'm using a regulated power supply I get good signal with relatively low noise.

Now I wanted to power it with battery using a step up boost converter - but got a serious noise problem on the amplified signal.
Since it's a hobby project I don't have the equipment (scope) to find the exact characteristis of this noise, I only know it's bad enough to completely ruin my audio input, and that it comes from the boost converter since I don't get it when I use regulated power supply or when connecting the circuit directly to batteries.
Tried to add decoupling caps, but it didn't help much. I even tried adding another low dropout regulator (s81230ag) before the opamp and adc analog power but the noise is still there.

How would you recommend to proceed?

  • Is there a simple way (small number of components, on a breadboard) to filter this noise with the existing components this and this?
  • Should I try a differnt boost converter, perhaps with switching frequence high enough so it could be filtered by the mic amplified circuit?
  • Pehaps I should use a differnt amplifying circuit for the mic which is less susceptible to power supply noise?
\$\endgroup\$
  • \$\begingroup\$ Tried a LC filter yet? \$\endgroup\$ – Ignacio Vazquez-Abrams Nov 9 '13 at 9:07
  • \$\begingroup\$ As a first step, try a really large decoupling cap, say 1000 uf or more. You don't say what caps you already tried. \$\endgroup\$ – Brian Drummond Nov 9 '13 at 10:30
  • 2
    \$\begingroup\$ a circuit would help. \$\endgroup\$ – Andy aka Nov 9 '13 at 11:22
  • \$\begingroup\$ @BrianDrummond The largest I tried was 220uf. I'll try to get a larger one and give it a try. \$\endgroup\$ – Amir Gonnen Nov 9 '13 at 13:01
2
\$\begingroup\$

There are two potential problems with the booster you've chosen. Firstly, it's specified as having 30 mV ripple and if your Mic circuit is producing output signals close to the supply rails then you'll get this ripple superimposed. Secondly your amp has a power supply rejection spec that may be poor and this will produce noise on the feed to your ADC. Thirdly, the switcher is a type that works best under medium load conditions and if your amp is only taking 10mA or less the switching frequency will be quite low and possibly ripple will be higher. Try loading the 3v3 with 100 ohms and see if this improves things.

Fourthly, your electret Mic power feed should have two resistors from the power rail and a 10uF cap down to ground to further reduce supply ripple superimposed on the Mic signal.

\$\endgroup\$
  • \$\begingroup\$ Loading the booster with 100ohm did not help. However, I added, as you suggested, two 100ohm resistors on the mic-module power and ground and added 220uF cap between the mic-module power and ground and problem was solved! Cap alone was not sufficient... Could you explain why the resistors are needed? \$\endgroup\$ – Amir Gonnen Nov 9 '13 at 19:04
  • \$\begingroup\$ @AmirGonnen I said try loading it with 100 ohms not 1000. Any way it works now. I'm assuming that you have 3V3--->1K---1k----mic with the junction of the two 1ks having a cap to gnd. The 1k at the 3v3 and the 220uF capacitor make a good low pass filter with frequencies on the 3v3 above about 1Hz progressively attenuated at 6dB per octave. Glad it works dude. \$\endgroup\$ – Andy aka Nov 9 '13 at 20:25
  • \$\begingroup\$ I did not load it with 1000, I loaded it with 100-ohm (100ohm). I'll try to remember spacing the "zeros" and "ohm"... all the resistors I've added were 100-ohm so no 1k there. \$\endgroup\$ – Amir Gonnen Nov 9 '13 at 20:43
  • \$\begingroup\$ @AmirGonnen oops my mistake - hey it works!!!! \$\endgroup\$ – Andy aka Nov 9 '13 at 21:29
  • \$\begingroup\$ Yes it works - Thanks!! All I really needed was a low pass filter on the mic module power. \$\endgroup\$ – Amir Gonnen Nov 10 '13 at 7:25
1
\$\begingroup\$

Another thing you can try is to use a different LDO (Low Dropout Regulator). I know you said you tried adding a s81230ag LDO in series, but looking at the datasheet for this part, I could not find a measurement for the PSRR (Power Supply Rejection Ratio), which leads me to believe that it is not very good.

From a quick look at the datasheet of your switcher, it looks like the switching frequency is 125 KHz. Strictly speaking this should be outside of the hearing range, but maybe some division of that frequency is getting into your signal. Since your circuit is for audio, you can try recording the signal into your computer and giving us a screenshot. You shouldn't be able to see the switching noise directly (assuming your PC soundcard has a low-pass filter that works), but you should be able to see the noise you are actually hearing. If you can't record it on your computer sound card, you won't be able to hear it anyway.

There are LDOs which specify very good PSRR, but you have to be sure they can reject the frequencies that your noise appears at, so you need to have some idea of the characteristic of the noise. For example, the TPS79901 has good PSRR (<-50dB) within the audio passband, but the PSRR gets worse with higher frequencies. I suspect that using the correct LDO in series can significantly reduce the noise (at the cost of having lower output voltage due to the drop of the LDO).

\$\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.