Audio DC Isolation

[electronicsfreak]

Hi All,

Hopefully a simple enough question for discussion. I am building an audio project with a single 5V DC supply, so I plan to add 2.5V to the original audio signal from a laptop, mp3 player, etc. before it is amplified and sent along it's way.

What would be considered proper isolation in this case, I.E. preventing any of that DC (particularly from that 2.5V reference) from ending up feeding the input of the audio source? is a single, simple first order capacitor HPF considered well enough? Would a stronger filter be required for something like this? (I plan to put the knee ~20Hz). Is there generally a defined frequency where "DC" (note the quotations) to that device is observed (say 1Hz, 5Hz, etc), where potential damage could occur?

Anyways, curious on your thoughts here.

-EF

 

[bountyhunter]

There is virtually no music content below about 30 Hz. Put the HPF anyplace below that.

 

[AnalogKid]

This usually is done with a single capacitor between the two stages being coupled. Also usually, the output impedance of the source stage is much lower than the input impedance of the receiving stage, so the output impedance dominates the filter corner frequency. The output impedance and the capacitor form a highpass filter, so you do not need an explicit resistance. The frequency response of a single pole filter is down 3 dB at the corner frequency and 1 dB one octave away, so to have "flat" bandwidth from 20 Hz, the corner frequency needs to be set to 10 Hz or below.

ak

 

[Nigel Goodwin]

Hi All,

Hopefully a simple enough question for discussion. I am building an audio project with a single 5V DC supply, so I plan to add 2.5V to the original audio signal from a laptop, mp3 player, etc. before it is amplified and sent along it's way.

What would be considered proper isolation in this case, I.E. preventing any of that DC (particularly from that 2.5V reference) from ending up feeding the input of the audio source? is a single, simple first order capacitor HPF considered well enough? Would a stronger filter be required for something like this? (I plan to put the knee ~20Hz). Is there generally a defined frequency where "DC" (note the quotations) to that device is observed (say 1Hz, 5Hz, etc), where potential damage could occur?

Anyways, curious on your thoughts here.

-EF

You're VASTLY over-thinking this - you simply need a DC blocking capacitor, it's also highly unlikely to cause any problem even without it, as any input 'should' already be DC blocked. But it's standard practice to include such a capacitor on both inputs and outputs.

 

[Rich D.]

I'll back up what Nigel Goodwin stated. Virtually all reasonable audio outputs and inputs should have a DC blocking capacitor included in the device. You should be able to confirm that by measuring the DC resistance and (after a suitable delay of seconds or so) see that the DC resistance is ∞ or "OL" or at least in the very high Mega Ω region. These devices will likely have the right value capacitor for its designed low-frequency response.

In some rare cases you could have a transformer coupled output or input, in which case you may see a much lower ohm reading. Not very likely with a laptop or MP3 player though. If that is the case you will need a blocking capacitor to prevent damage to either the 2.5V power source or the transformer depending on which part looses the battle. You sound like you are able to determine the capacitor value in Farads to use. Best results would be from a metal film capacitor, polyproplyne, polystyrene...something like that. Ceramic caps are best avoided for audio use. Beyond a μF or so, it may be hard to find, so select an electrolytic that is "non-polarized" or "bipolar". Not only will you not have to worry about DC polarities on the cap, but non-polarized electrolytic capacitors are known to have less distortion than any polarized capacitors in the audio range. I would err towards getting a voltage rating on the cap as high as practical, of at least 16 volts, 25 even better.