0
\$\begingroup\$

I'm trying to make a pre-amplifier for my PC microphone with high gain and low noise.

I tried a few circuits but they had a lot of noise and low gain.

How can I make a preamplifier with high gain and low noise?

  • Frequency range: up to 20kHz
  • Supply voltage: 5V

I tried this circuit:

enter image description here

\$\endgroup\$
6
  • 2
    \$\begingroup\$ What microphone make and model? What DC current does it need? What frequency response is needed for the amplifier? What power supply rails are available? What output voltage level do you need for a given input sound pressure at 1 metre? Specifically, what circuits did you try (you said you tried a few)? \$\endgroup\$ – Andy aka Apr 13 at 9:24
  • 1
    \$\begingroup\$ Does it also have to be low cost, use only components from the seventies, and be constructed on a breadboard? \$\endgroup\$ – pipe Apr 13 at 9:28
  • \$\begingroup\$ this is my microphone en.wikipedia.org/wiki/Electret_microphone \$\endgroup\$ – Macan Apr 13 at 10:59
  • \$\begingroup\$ somehow i wanna make a preamplifier for my sound card \$\endgroup\$ – Macan Apr 13 at 11:00
  • 2
    \$\begingroup\$ This doesn't quite count as an answer: I experimented with this a couple years ago and the enormous noise on the power supply was a major problem. I was drawing power from a USB port, where are you getting yours? The best solution I found was to put a diode in series with V+ and then a decoupling cap to ground after the diode. \$\endgroup\$ – Ian Bland Apr 13 at 11:15
3
\$\begingroup\$

The problem is likely not with your preamp itself (the circuit is basically OK), but with the power supply and gain setting.

Using the internal 5V from your PC (or USB port) for a mic pre is pretty much a no-no. There is almost certainly switching noise in that power rail which will be very audible. (The mic output is typically a few mV, perhaps 10-20mV or a bit more with an electret. This circuit has a gain of 100, so it looks like it expects about 50mV pk-pk as a maximum which seems believable.)

The simplest option is usually to derive a clean voltage rail from a linear regulator and use this for your pre. However if you start with +5V your rail can only be of the order of 4 or 4.5V - now you must select your opamp carefully so that it is suitable for low voltage operation, and also so that the output can swing as close to the rails as possible, so you do not lose headroom. (Pretty sure a 358 won't do this very well. The data sheet says that if you have +30V, Vout will go to 26V, at 5V they aren't admitting anything for the high side swing.)

A better approach is to use a switcher to derive one or more rails from the 5V (ideal would be about +/-16V) and then use linear regulators to get (say) +/-15V which you can use to power op amp circuits using standard opamps (like the venerable and still excellent NE5532). However this is obviously a more complex approach.

The second issue will be gain control. Just knowing the mic sensitivity is not enough. The difference between someone whispering and shouting into a mic can easily be 20-30dB, and on top of that distance of the source from the mic plays a big part (moving from 3 to 12 cm causes a 12dB change in signal). So a fixed gain mic amp is not much use. You need a gain control. (In fact to process a wide range of audio signals, some form of AGC or compression is usually used, but that is outside the scope of your question.)

Given these factors, your application is going to play a big part in your design decisions.

\$\endgroup\$
4
  • 1
    \$\begingroup\$ LM358 (and 324) can handle signals from 0 to 3.5V on a single 5V supply, I used 358s in my own similar experiments with a PC electret mic preamplifier. \$\endgroup\$ – Ian Bland Apr 13 at 11:46
  • 1
    \$\begingroup\$ ok. So you lose at least 3dB of signal compared to a rail to rail device. (Actually in reality its more like 8dB if you centre bias the circuit, as your max symmetric signal is 1.5 to 3.5V, or 2V pk-pk from a possible 5V). It's not really clear whether that is a problem for the OP or not, but it's not a great start. \$\endgroup\$ – danmcb Apr 13 at 11:50
  • 1
    \$\begingroup\$ yes, you do, but 358s are cheaper so swings and roundabouts. I was just pointing out that they are useful at that voltage, not that they are optimal :) \$\endgroup\$ – Ian Bland Apr 13 at 12:52
  • 1
    \$\begingroup\$ yes, I get that. In that case, it would be better to change R5 to 51k in the OP's circuit, as that biasses the circuit close to 1.75V which is really the "mid rail" point (because of the 358's asymmetric voltage swing). Then you get close to 3.5V pk-pk out, which helps a bit. \$\endgroup\$ – danmcb Apr 13 at 13:09
1
\$\begingroup\$

As others have hinted at, the LM358 is ancient and far better modern parts are available. Put "low noise amplifier" into your favourite parametric search. You might select something like the AD8691, a 5V supply low noise amplifier.

Secondly, power supply noise is critical. You're going to need to experiment with more filter capacitors and possibly a ferrite choke to get decent power off USB. You might be better off building the first version to run off 6V of alkaline batteries.

\$\endgroup\$
2
  • \$\begingroup\$ To stay within spec you aren't allowed more than 10uF decoupling on USB and on its own it isn't going to do much against the wall of noise coming off your typical USB power supply. As I said in a comment above, a series diode worked the best for me. That does reduce available voltage further, but you can still easily have up to 3V peak to peak using a cheapo LM358. Op amp noise and quality is an insignificant issue compared to the power supply noise issue. \$\endgroup\$ – Ian Bland Apr 13 at 12:58
  • \$\begingroup\$ far better is a low dropout regulator - you can lose less volts than with a diode and reject much more noise. They don't cost much. \$\endgroup\$ – danmcb Apr 14 at 7:32
1
\$\begingroup\$

A lousy old lM358 is never used for audio because it is noisy (hissss), has a poor slew rate above 2kHz and produces crossover distortion. Your very low value of 1k tor R4 (because the opamp is inverting instead of non-inverting) reduces the output level from the mic to 1/3rd so the amplification from the opamp must be increased which increases its noise.

\$\endgroup\$

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