Im very new to electronics and what Im currently doing is my first project. I want to make a tiny guitar amp using a speaker I salvaged from an old computer speaker. The speaker says 2 watts 6 ohms on it. Ive researched online and found it probably means 2W rms but it could mean peak too. Since the speaker is so tiny itd be really cool to make the amp battery powered with a 9v battery if possible. I did some calculations and I hope theyre good lol.

I think the rms voltage for 2W of power should end up being the equivalent to around a 4.9V amplitude, which is a little more than what the 9V battery can provide.

the current for 2W is around 0.58A, which I read would deplete a 9V battery in an hour.

I dont understand wether the speaker needs 2W of power to even work, or is that a maximum amount of power it can handle? will the speaker always "try" to draw the current needed, or can I adjust how much current it gets with other circuit components (when designing the circuit it says "choose a quiescent current", does this mean I can somehow set how much current goes from the battery?). Is 2W just for the loudest about the blow the speaker volumes and can it run normally on less power, and if yes, how to provide less power?

Edit: Its a common emitter amplifier circuit, R5 is the speaker

enter image description here

Edit: Id like to keep this circuit as simple as possible since it took me a long time just to understand emitter amplifiers haha, Im very much using this project as a beginner learning project.

Edit: Ill use an lm368 op amp instead of this, I see that this sucks now

  • 3
    \$\begingroup\$ Look at the datasheet for the battery. 2 W is way more than you can extract from it. Your average music, save for Enya and dubstep, is about 1/8 of the peak power if you play it just under clip. \$\endgroup\$
    – winny
    Commented Dec 17, 2017 at 20:20
  • 2
    \$\begingroup\$ 1) include the actual circuit 2) single transistor CE amplifiers are crap for powering a speaker. A 200 ohms speaker might work but lower: forget it. All beginners expect miracles from single transistor amps. 3) 2 W is the maximum power that should not destroy the speaker. But treat that 2 W as a guess, usually the distortion at 0.1 W is already so bad that you will stop there. 4) Buy a proper amplifier, like an LM386 based module: ebay.com/itm/… That will give you at least reasonable sound. \$\endgroup\$ Commented Dec 17, 2017 at 20:24
  • \$\begingroup\$ Bimpelrekkie I dont really care about a reasonable sound, I just wanna make something that works and can be used, what im interested in is does that mean that I can use this speaker with a 9V battery if 0.1 watt of power is enough? \$\endgroup\$ Commented Dec 17, 2017 at 20:32
  • \$\begingroup\$ Note that the output impedance of a common-emitter amplifier is much larger than 6 ohms; it is unlikely that you'll get any voltage gain this way. A relatively simple 'fix' is to add an emitter-follower stage between Q1 and the speaker. \$\endgroup\$ Commented Dec 17, 2017 at 20:41
  • \$\begingroup\$ That speaker should work, at least if you have a sane design that keeps DC off the coil. It's more a question if your amplifier tries to push enough power through it that the battery sags. A very efficiently made bridged configuration fed a strong signal probably could overdraw the battery by driving both sides of the speaker in opposition, one with only a single driven side and an output capacitor likely not. But today you might also consider a class D amplifier made for phones and running off 3 AA's or AAA's. \$\endgroup\$ Commented Dec 17, 2017 at 20:43

3 Answers 3


The following schematic uses fewer parts than the one shown at this link as Fig 5.5.3.

So here's my simpler version:


simulate this circuit – Schematic created using CircuitLab


Let me start at the speaker and bootstrap, without getting into details here. You may usually see \$C_1\$ with a speaker tied to ground. But you can also tie the speaker to the plus rail. Either way, \$C_1\$ will wind up with a nearly constant voltage across it. But that voltage is more useful to you if you direct it upwards towards the plus rail, instead. So that's what this circuit does.

The useful trick in making this choice is that since \$Q_1\$ "looks like" an emitter follower here, there is a fixed voltage across the BE junction of \$Q_1\$. With \$C_1\$ also having a relatively fixed voltage, too, this means there is an almost fixed voltage across \$R_3\$. And that means that \$R_3\$ "looks like" a current source. Which is perfect for this application and saves parts while at the same time providing a better way to handle biasing for \$Q_1\$ and \$Q_2\$.

BJT emitter degeneration

I've added \$R_1\$ and \$R_2\$ so that they provide, at maximum output, voltage drops of about \$100\:\textrm{mV}\$. This helps a little with variations between the transistor parameters.

You could eliminate them, if you want to. But if you have parts of those values (or as much as an Ohm, or so) then it might be nice to experiment with and without them to see if you can tell a difference or not.

The main reason I want them there for now, other than what I already said about it, is that you will need them for later circuit adjustments. So for now, keep them and find something near to those values if possible.

The \$R_3\$ current source

If things are tweaked right (and that's coming up), then there should be approximately \$2.9-3.7\:\textrm{V}\$ across \$C_1\$ and therefore about \$2.2-3.0\:\textrm{V}\$ across \$R_3\$. (The large variation here is due to the fact that a \$9\:\textrm{V}\$ battery varies over its lifetime something like \$7-9\:\textrm{V}\$.) So I set this to source about \$8-11\:\textrm{mA}\$ over the operating life of the battery. The reason I picked that range is that I expect the base currents for \$Q_1\$ and \$Q_2\$ to be around \$1\:\textrm{mA}\$, or so, and I'd like to have nearer 10X that much available near the bases so that variations in base currents can be tolerated well.

\$Q_1\$ and \$Q_2\$ AB biasing

Start with the \$R_5\$ and \$R_6\$ pair, without any input signal. Temporarily replace both of those resistors with a single resistor and don't include \$C_2\$, to start. Just focus on the resistance of a single resistor here. Adjust the value until you see \$\frac{1}{2}\:\textrm{V}\$ more than half your battery voltage at the point shown by the red arrow. So if you have a fresh battery, this means you want to hit about \$5\:\textrm{V}\$. Adjust until you get close. Then record that resistance value.

To complete the biasing, place a voltmeter across \$R_1\$ and plan to make adjustments to \$R_4\$ (blue arrow.)

I've slightly differently arranged biasing of the two output BJTs, \$Q_1\$ and \$Q_2\$, using a series resistor \$R_4\$ here rather than the arrangement used at the link at the top, above. The idea of \$R_4\$, \$D_1\$, and \$D_2\$ is to operate \$Q_1\$ and \$Q_2\$ so that they are both active, but not overly so, when there is no input signal applied. I'd recommend trying to hit about \$5\:\textrm{mA}\$ with no signal applied.

So with the voltmeter in place, adjust the value of \$R_4\$ until you see a voltage drop across \$R_1\$ that is predicted by \$5\:\textrm{mA}\cdot R_1\$ (whatever the value for \$R_1\$ in your case might be.) If you used the values I show, then this would be around \$1.5-2.0\:\textrm{mV}\$.

If it turns out that the current is still higher, even with \$R_4\$ shorted, then just short out \$R_4\$ (remove it and short the nodes.) If that still has the measured voltage too high, then put \$R_4\$ in parallel with the two diodes instead as shown in Fig 5.5.3, start perhaps with \$1\:\textrm{k}\Omega\$, and bring the value downward until that goal is reached.

Now go back and reverify that the node at the red arrow is still where I said you wanted to target, above. Again, adjust your temporary resistor value until that's true. Then record that value, again (if needed.)

Now, you must make arrangements so that the sum of \$R_5\$ and \$R_6\$ (green arrows) is approximately this temporary resistor value, but you need to split the difference between them. You can work that part out later. For now, just divide it about in half and re-add \$R_5\$ and \$R_6\$ and \$C_2\$.


Well, that's about it. Until you apply a signal. That's when you get to worry about how you divided out that temporary resistance value into \$R_5\$ and \$R_6\$. I show an uneven distribution in the circuit because, well, it's likely that the best setting will be uneven.

Putting a higher percentage into \$R_6\$ will means more negative feedback at AC, which may be required. But I'll leave the exact distribution as a task for you to experiment with. Keep a significant portion in each of them, but feel free to play a bit.

Also, experiment with the circuit using a somewhat drained battery, as well. Make sure it all still works reasonably well.

And of course, certainly try out Fig 5.5.3. If it is all equal to you, that one is probably better to go with. I just wanted to give you a flavor of how to approach a case with even fewer parts.

  • \$\begingroup\$ Thank you, I appreciate the effort youve put into this, although i have to say that i have trouble wrapping my head around it now since im really a beginner. Ill get back to this later probably. \$\endgroup\$ Commented Dec 18, 2017 at 2:20
  • 1
    \$\begingroup\$ @TwoheadedFetus I'd probably recommend starting with trying to wrap your head around \$Q_1\$ and \$Q_2\$ as the output pair, rather than the approach you showed in your own example circuit. You really need to understand why your circuit isn't so good and the BJT pair is so much better. If you can't get past that, then the rest never really will make sense. So start there. \$\endgroup\$
    – jonk
    Commented Dec 18, 2017 at 2:24
  • \$\begingroup\$ ok, first i want my own circuit to work and then ill see what can be changed \$\endgroup\$ Commented Dec 18, 2017 at 2:52
  • 1
    \$\begingroup\$ @TwoheadedFetus Your own circuit simply won't work (well, without excessive/destructive heating anyway.) But even that will be a good lesson, I suppose. Have at it. \$\endgroup\$
    – jonk
    Commented Dec 18, 2017 at 11:21

The 2W rating for the speaker is just the maximum it can handle without damage. You can put a lot less power into it without any problems. For an everyday room setting, a few hundred milliwatts can sound quite loud, and even a few tens of milliwatts can be enough to listen to.


If the amplifier is powered by a single 9 V source, then 9 V is the maximum signal amplitide from the negative peaks to the positive peaks.

If Vrms = 1 V, then

Vpeak = 1.414 V (Vrms x root2)

Vp-p (Vpeak-to-peak) = 2.828 V

For 2 Wrms (the speaker continuous power rating) of output power,

P = E^2 / R

12 = E^2

E = 3.46 Vrms

Therefore, the peak-to-peak output voltage swing for 2 W into 6 ohms is 9.8 V. Even with a 100% efficient amplifier circuit, a single 9 V battery is not enough for full power.

Separate from that, a single common-emitter stage is going to have a lot of distortion even at low volumes, and the transistor will run pretty hot. You are starting in a reasonable place, but the performance will be so poor that little will be learned. There are zillions of circuits and kits for low power audio amplifiers, many of them all discrete. have you considered starting there?

  • \$\begingroup\$ To be honest part of the fun was gonna be the usage of salvaged parts, using kits and premade circuits...isnt as appealing haha. I think i can find a smaller speaker in an old toy electronics kit i had as a child and maybe that could work, but im thinking maybe this speaker even if not on full power could be enough idk. If nothing works i guess ill just buy a cheap one. And im fine with this not being the best pr most efficient circuit, tbh ill be amazed and feel great if it just works at all, thats a start, the main thing for me is understanding exactly what im making. \$\endgroup\$ Commented Dec 17, 2017 at 22:41
  • \$\begingroup\$ @TwoheadedFetus Do you want a design that can't entirely be designed, but will require some tweaking once built to "calibrate" it, so to speak? Or do you want a design that can be crafted so that it will work straight out, without adjustment later? (There are quasi-designs that do depend on the BJT and need tweaking, but work fine once tweaked. And others that use another BJT or so and don't need the tweaking.) I'm keeping in mind the fact that a 9 V battery has 1-2 Ohm series impedance and droops to 7 V over use, or thereabouts. \$\endgroup\$
    – jonk
    Commented Dec 17, 2017 at 22:46
  • \$\begingroup\$ @jonk im fine with having to tweak it, wouldnt expect it to be perfect straight out, id rather start with a simpler design first and move on to more complex ones later. \$\endgroup\$ Commented Dec 17, 2017 at 22:51
  • \$\begingroup\$ @TwoheadedFetus Okay. But the problem with that is that the tweakable one isn't really good design practice. So you'd be learning to do things WRONG, first. Which isn't always the best way to start. But it is one way, I suppose. Are you sure you'd rather just keep it simple but have to make measurements and tweak? \$\endgroup\$
    – jonk
    Commented Dec 17, 2017 at 22:55
  • 1
    \$\begingroup\$ learnabout-electronics.org/Amplifiers/amplifiers55.php \$\endgroup\$ Commented Dec 17, 2017 at 23:01

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.