I have an audio signal of these characteristics coming out of an instrumentation amplifier:-

Frequency: white

DC offset: 2.5V

Amplitude: 3Vp-p

I also have some 1/4W 8 Ohm and 64 Ohm speakers pilfered from old transistor radios. I would like to be able to hear the audio component of my signal. I can just about hear it if I plug the 64 Ohm speaker directly into the in amp output, but of course that's not the proper way. The speaker is to be mounted permanently to act as a verification device that all the circuitry is working. It is not a temporary thing. I also need to maintain the signal levels for further processing down stream so I can't suck all the voltage level out of the in amp.

I've considered a single transistor voltage follower, a simple discrete component class A amplifier and the venerable LM386 audio amp. I also have a 13.8V power rail. Is there a simplest option to hear my noise?

  • \$\begingroup\$ AC couple it into your Mic input on your computer \$\endgroup\$ Commented Sep 28, 2016 at 3:38
  • \$\begingroup\$ High impedance headphones sound a good option. \$\endgroup\$
    – Andy aka
    Commented Sep 28, 2016 at 9:42
  • \$\begingroup\$ @Andyaka Sorry, I wasn't clear. I want the solution to be permanent with a mounted speaker, not just a one off thing. I've edited... \$\endgroup\$
    – Paul Uszak
    Commented Sep 28, 2016 at 11:46

2 Answers 2


A low-count buffer amp might consist of a class AB emitter follower:


simulate this circuit – Schematic created using CircuitLab

Q2, a NPN general-purpose transistor will run a bit warmer than the lower PNP transistor. Something that can dissipate a bit more heat like TIP29 or MJE180 might be more robust.
C1 will be an electrolytic capacitor of greater than 6v rating, with positive end facing the two transistor emitters.
Note that this buffer depends on your instrumentation amplifier to provide +3v bias - "input" must be connected directly to op-amp output.

  • \$\begingroup\$ Glen, I've just messed with this and I think that it needs another diode in series with D1. You get some cross over distortion otherwise which kinda ruins the point of an AB circuit. The part count is getting a bit high. Any chance a more basic class A might work? I can't quite perfect the arrangement though... \$\endgroup\$
    – Paul Uszak
    Commented Sep 30, 2016 at 4:13
  • \$\begingroup\$ @PaulUszak - your criticism is well founded - biasing out that crossover is one of the most difficult parts of AB operation - was hoping that as a "monitor" this would be adequate. Class A will necessarily run hot. You can ease Class A power dissipation with an audio step-down transformer, perhaps from that pilfered transistor radio - are you willing? That LM386 is looking more attractive. \$\endgroup\$
    – glen_geek
    Commented Sep 30, 2016 at 13:12
  • \$\begingroup\$ You're absolutely right. I asked for a monitor and now having gotten that stupidly expect hi-fi quality from it. Touche. \$\endgroup\$
    – Paul Uszak
    Commented Sep 30, 2016 at 17:55
  • \$\begingroup\$ @PaulUszak , Moving closer to that critical class AB area will reduce crossover distortion. Its a bit risky, but you can do this with one extra resistor...replace D1 (1N4148) with a series combination of a 100 ohm resistor + 1N4148. Maybe even as high as 120 ohms, but that's riskier. \$\endgroup\$
    – glen_geek
    Commented Sep 30, 2016 at 18:35

I'm not familiar with the concrete circuit design issues, but I note that your use case (driving a small speaker from a similar-level signal) is very much like a headphone amplifier. You could look at DIY headphone amp designs for inspiration, and possibly use something even simpler because you aren't looking for audio fidelity — though the once popular "CMoy" design is just a single op amp plus supporting components.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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