DC offset in microphone output? Really? Why?

[analogo]

In this video by GreatScott he shows a simple amplifier circuit made with an opamp. What I do not understand is: why does he say that applying a 2.5v offset to the non-inverting input of the opamp will create an output offset of 2.5v around which the waveform will oscillate around? Direct link: https://youtu.be/ts-JqEVzvDo?t=6m59s.

IIUC, the 2.5v input offset will _remove_ 2.5v from the inverting input, making a 0 -- 5v input voltage into a -Vrail -- +Vrail output waveform.

And even if he is right and there is a 2.5v offset in the output, isn't that a bad thing? I thought that any DC offset in bad in an audio setting.

[mikerj]

The op-amp in question is running from a single supply rail, so the output can not swing below ground.  In order to cope with an audio signal in this case you need to impose a DC offset, and for a rail-rail output amp running from 5v, you'd want this offset to be close to half the supply rail to maximise the output swing.

Since the inverting input to the op-amp is AC coupled, the DC gain at the non-inverting input is 1.  Therefore applying 2.5v to the non-inverting input gives 2.5v on the output.

[ebclr]

Nice site

https://www.elv.de/journal

[tooki]

The op-amp in question is running from a single supply rail, so the output can not swing below ground.  In order to cope with an audio signal in this case you need to impose a DC offset, and for a rail-rail output amp running from 5v, you'd want this offset to be close to half the supply rail to maximise the output swing.

Since the inverting input to the op-amp is AC coupled, the DC gain at the non-inverting input is 1.  Therefore applying 2.5v to the non-inverting input gives 2.5v on the output.

That makes sense. But I thought that a DC offset at the output is bad (for the input of the computer or mixer, etc.)? Wouldn't you then want a DC-blocking capacitor at the output (after the op-amp, but before the signal reaches the input of the other device)?

Nice site

https://www.elv.de/journal

It is (especially old issues), but it's only available in German, which doesn't help most people here, and what does this have to do with this thread?

[rhb]

That makes sense. But I thought that a DC offset at the output is bad (for the input of the computer or mixer, etc.)? Wouldn't you then want a DC-blocking capacitor at the output (after the op-amp, but before the signal reaches the input of the other device)?

Probably.  It depends upon what comes next.  Lots of stuff is AC coupled, but some isn't.   This can be a great source of confusion.

[analogo]

That makes sense. But I thought that a DC offset at the output is bad (for the input of the computer or mixer, etc.)? Wouldn't you then want a DC-blocking capacitor at the output (after the op-amp, but before the signal reaches the input of the other device)?

Probably.  It depends upon what comes next.  Lots of stuff is AC coupled, but some isn't.   This can be a great source of confusion.

What comes next is a standard PC audio chipset. Aren't there specs that tell you "The audio signal should swing between -X volts and +Y volts, the maximum tollerable DC offset is Z volts"?

For example, the output of this amp is never going to go under 0, is this OK as an input for an audio card?

[ogden]

What comes next is a standard PC audio chipset. Aren't there specs that tell you "The audio signal should swing between -X volts and +Y volts, the maximum tollerable DC offset is Z volts"?

Specs of audio chipset indeed tell maximum signal swing, but they also say that decoupling of line-in using capacitors must be implemented. Try to google (pictures) "ac97 codec reference design circuit" and you will see - they all have AC decoupling. Yes - circuit shown in the video is not quite correct, but it will work with most PC's/mixers/amps, if not all. Only seemingly bad thing that can happen - jack detection may not work.