0

I have a few concerns with the audio amplifier part of my circuit which I hope somebody can help me with...

It's part of a larger circuit but the schematic below shows the parts of the circuit which I'm concerned about.

Transistor Amplifier Schematic

Short Version:

  1. Does the headphone socket need a capacitor and if so where?
  2. Does the speaker need a separate capacitor?
  3. Does a value of 10 Ohms for R10 seem suitable?
  4. Could this part of the circuit be responsible for draining the battery so quickly?
  5. And for a bonus point, am I within the 20ma continuous current limit of the MCU pin?

Longer Version

As it is, the circuit appears to be working fine, and I get reasonable audio from both the speaker and headphones. However, I have two devices on my bench. One which I've made myself and has a 250mah battery, and an off-the-shelf solution which has an 80mah battery. The off-the-shelf solution will run on it's battery for about 7 days. Whereas mine, with a substantially larger battery will only run for about 4 hours. I've ruled out everything else I can that might be consuming power so now I'm looking at the audio part of the circuit. Looking at my schematic, I made my best attempt at doing some calculations and came up with a value of 10 Ohms for R10. Does this seem correct?

My second concern is that I understand speakers don't like DC current so normal practice is to include a capacitor. As it stands, the headphone socket is receiving DC current. I tried putting a series capacitor between the digital potentiometer and the headphone socket but the circuit stopped working.

Datasheets Speaker Transistor

I've got no attachments to this model of transistor so am perfectly happy to swap it for a different one. However, I must use this specific speaker.

Glyn Davidson
  • 25
  • 5
  • If you are happy with your amplifier's topology, I'm not sure I've anything to add. I can well imagine why a COTS solution works longer. But then they probably have a few more parts, too. – jonk Jul 10 '20 at 08:37
  • Maybe add a buffer, and a DC blocking cap behind it, between the digi pot and the headphone jack/base of the transistor. Be sure to bias your signal of the digi pot at half the supply voltage. || Why go through the jack (pin2 and pin4) before going to the transistor? – Swedgin Jul 10 '20 at 08:56
  • This is the first time I see a digi pot used to make an audio signal. Most of the time it's just a PWM signal with a low-pass filter behind it. – Swedgin Jul 10 '20 at 08:57
  • Do you know *why* the circuit stopped working when you put a capacitor in series with the headphone socket? – user253751 Jul 10 '20 at 09:56
  • @Swedgin - the purpose of going through the jack is because the jack has a normally closed internal switch. When you insert some headphones into the jack it pushes the switch open and cuts the signal to the speaker. (No point playing audio from the speaker when you are using headphones). Could you explain a bit more about the buffer and DC blocking cap please? In simple terms for dummies please :) The signal is being created by an Atmega1284p. The digi pot is there to control the volume. – Glyn Davidson Jul 10 '20 at 17:54
  • @user253751 - No idea. I presumed it had something to do with the relationship between the capacitor and the transistor but I don't have the expertise to problem solve it. – Glyn Davidson Jul 10 '20 at 17:56
  • 1
    @GlynDavidson I think you have a great deal to learn about amplifier structures. But that said, I can't argue against simplicity, either. If you are satisfied with your topology and not interested in anything a little more complex but less power-hungry then I've little to add (as I said, already.) If you are seriously interested in lower-power designs, then you'll need to let loose of the topology at hand and move beyond it. But that's a whole other question. Just FYI. – jonk Jul 11 '20 at 04:26
  • @jonk - You're not wrong there; I do have a great deal to learn! The point is, I'm not satisfied with my current circuit, hence reaching out to the community for help. I'd be grateful if you could explain what makes my current topology power hungry and point me towards a better solution. – Glyn Davidson Jul 11 '20 at 07:26
  • @GlynDavidson Start [here](https://electronics.stackexchange.com/a/503581/38098) and tell me if that helps any. For a discrete design using BJTs, it's relatively straight-forward and there's enough there that you should be able to make a custom design. It's a class-A, so not very efficient. But it will be likely better than yours and it won't distort the signal nearly as much. Modern class-D amplifier designs will likely be more so. But they aren't entirely simple to design and build, discretely, either. I also have pages on class-AB, which is much better in efficiency than the above class-A. – jonk Jul 12 '20 at 01:23
  • @GlynDavidson At the bottom of [this page](https://electronics.stackexchange.com/a/368660/38098) I provide a class-AB design that is at a good hobbyist level using discrete BJTs. I don't describe all the details getting there, though. So just look at it for a sense of "complexity." Now that this is out of the way, there's another thought. You are driving this from an MCU pin. You could (it takes work) consider getting rid of the digipot and going to class operation beyond C (D, for example.) But what's best for you is likely what you can understand. So the class-A I provided is likely better. – jonk Jul 12 '20 at 01:26

0 Answers0