I read a bit about audio power amplifiers and struggle to understand the power rating of the emitter (i.e. current sharing) resistors in power amplifiers.

I refer to this project.

Schematics and BOM below: all credits and copyright go to buildaudioamps.com

According to the BOM, the resistors R31-R34 are chosen to have a 5W rating (wirewound type). I have done some simple Spice simulation and I see that at the maximum (reasonably usable) input voltage, the power dissipation across a single emitter resistor is only about 500mW, which is to be expected because of their low value.

Somehow I have the feeling, that the guy has an idea of what he is doing.

What is the reason for the chosen power rating?


enter image description here

BOM: enter image description here

  • \$\begingroup\$ So now we need to know values and the maximum permitted power rail voltages. \$\endgroup\$
    – Andy aka
    Mar 8, 2019 at 13:13
  • \$\begingroup\$ I have also added the complete BOM as presented on the author's website. The author further notes in the project description: "The measured power output of the prototype is about 110 watts into an 8Ω load with ±50VDC power supply rails" \$\endgroup\$
    – Junius
    Mar 8, 2019 at 13:25
  • \$\begingroup\$ Those are NOT current sharing resistors (and are also clearly NOT emitter resistors!). As to the power rating, you need to look at fault currents, especially as that design is completely lacking in any sort of over current protection so far as I can see, and you will hit 5W in a 0.22R resistor at about 5A, so 10A total in one leg of the design. 10A into 4R would be 40V so is not unreasonable, and allowing for the 50% duty a 5W part seems sane here as it is mounted in an area of the circuit prone to run hot and so will need derating. \$\endgroup\$
    – Dan Mills
    Mar 8, 2019 at 15:26
  • \$\begingroup\$ Thanks for pointing that out(that the transistors are on the collector side). Can you tell me what that output stage topology is called for further reading? \$\endgroup\$
    – Junius
    Mar 8, 2019 at 17:23
  • \$\begingroup\$ Of course the output transistors are current-boosted emitter followers. Lookup Sziklai Pair which is an improved darlington. The max average output voltage is 35.4V which is divided by the 8 ohm speaker down to 0.48V. Then each 0.22 ohms resistor gets (0.48V squared)/0.22 ohms= 1.05W A 5W resistor will survive 4W when somebody uses a 4 ohm speaker. \$\endgroup\$
    – Audioguru
    Jul 8, 2022 at 22:22

1 Answer 1


110 watts into an 8 ohm load implies an RMS current of 3.71 amps. That current, through a single 0.22 ohm resistor implies a power of 3.025 watts. Considering that this power is shared by 4 resistors, the average power per resistor is 0.76 watts.

But then one has to also consider the biasing of the output stage and this will cause a certain amount of DC current to pass though the resistors so maybe 1 watt per resistor might be dissipated if it's a class AB stage and possibly a couple of watts if it isn't properly biased i.e. it is closer to a class A stage.

The zobel network (C16 and R33) on the output might also add a few tens of milli watts to each resistor at lowish frequencies and a few hundred milli watts at high audio frequencies.

So you could be looking at maybe 2 watts to 2.5 watts for each resistor and therefore choosing a 5 watt resistor seems like a little bit of overkill considering that I would opt for more like a 3 watt resistor.

  • 3
    \$\begingroup\$ On the other hand, a 5W rated resistor will run cooler at the given power output and that may be a consideration for stability and reliability. Where cost is not paramount, it's often better to over-engineer. \$\endgroup\$
    – Ian Bland
    Mar 8, 2019 at 13:49
  • \$\begingroup\$ First of all, I had a typo in my question where i mention 5mW per resistor - should have read 500mW in the first place (per resistor). I have corrected that The simulation actually shows about 400mW (rms) per resistor on 8Ohms load and +/-50VDC supplies. Still seems a bit of over-engineering. I will check the temperature derating curves for various resistor types -> is the inductance of "standard" wirewound power resistors a problem in that audio application? (i doubt it) \$\endgroup\$
    – Junius
    Mar 8, 2019 at 13:59

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.