I am teaching myself about audio amplifiers and put together a basic circuit to receive the signal from my laptop and then drive a pair of 32 ohm headphones.

The small amplifier is powered by the USB port on the laptop. The sound signal comes out of the laptop and into the little amplifier via a TRS plug. The small amplifier has a rail splitter on the power input to create a virtual ground at around +2.5 volts above the USB ground. The left and right signal inputs to the amplifier have DC blocking capacitors on them.

My quandary has to do with how to deal with the common line on the audio cable going from the laptop to the little amp. On the laptop the common line is connected with the same ground used by the USB ports (0 volts). I checked this with a multimeter. On the amp, the common pin on the input jack is connected to the virtual ground created by splitting the USB voltage, so it is at +2.5V. Clearly this is not right as it creates a ground loop back into the laptop.

So is there a way to implement this setup correctly? Is it appropriate to place a blocking capacitor on the common line as with the L and R signal lines? Should the common line be connected instead to the 0 volt rail (V- on the amplifier) since it will be at the same potential as the source and the inputs are DC-blocked?

Schematic added:

Mismatched ground references

  • \$\begingroup\$ Better if you provide a schematic. Electrical engineers think graphically so much it makes us grumpy to try to follow a detailed description. It inclines us to draw the schematic ourselves, which makes us wonder why it's us doing that work and not you :) \$\endgroup\$
    – scanny
    Oct 10, 2015 at 3:43
  • \$\begingroup\$ Added the schematic for one channel :). \$\endgroup\$
    – codeZonkey
    Oct 10, 2015 at 4:26

2 Answers 2


You probably won't need to worry about the common since it will be already common with the 0V from the USB port. However, this is not a very nice situation- it means that power supply noise is going to be injected directly into your amplifier input. It may not matter so much for a relatively high level signal, but if there is noise in your headphones, this is probably why.

Options include adding a differential front end to your amplifier or isolating the power supply (for example, with a DC-DC converter) or adding signal isolation (for example, with an audio transformer).

  • \$\begingroup\$ Sorry for my ignorance here but do you mean that I can leave the common input line disconnected from the circuit? \$\endgroup\$
    – codeZonkey
    Oct 10, 2015 at 4:34
  • \$\begingroup\$ it will kind of work with common disconnected, you'll get power noise superimposed on your audio signal. \$\endgroup\$ Oct 10, 2015 at 4:45
  • \$\begingroup\$ Yes, that's right. It's already connected inside the computer to the 0V. You can check this with an ohmmeter (computer off). Anything you can do externally will be shorted by this internal connection unless you isolate the supply and/or the audio or use differential input. \$\endgroup\$ Oct 10, 2015 at 5:01
  • \$\begingroup\$ I don't have an isolation transformer at hand which would be the best solution. If I isolate the power supply, wouldn't the common lines still be offset if they are connected together through the audio cable? \$\endgroup\$
    – codeZonkey
    Oct 10, 2015 at 14:15
  • \$\begingroup\$ If you isolate the power supply then the common of the audio will be at the virtual ground potential and you don't even need a coupling capacitor (with the attendant 'thump' at power-up). \$\endgroup\$ Oct 10, 2015 at 15:48

instead of connecting vgrnd directly to the rail splitter connect it via a 10K resistor and connect it to the plug's sleeve via a capacitor

do not connect plug sleeve (audio ground) to the rail splitter.

this will pass the DC signal needed to bias the amplifier whilst also passing the relatively noise-free ground reference of the audio cable.

  • \$\begingroup\$ Ah! I will give this a try. \$\endgroup\$
    – codeZonkey
    Oct 10, 2015 at 5:27
  • \$\begingroup\$ I tried this and it does help with the noise, but it got me thinking that now the laptop amplifier is not fully buffered and has to absorb the current going through the headphones. The resistor helps but the current still has to go there. With my small test setup it's not too bad but probably won't scale up with power. \$\endgroup\$
    – codeZonkey
    Oct 10, 2015 at 14:07

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.