single supply op amp signal bias error

[electrolust]

When running an op amp with single power supply, I'm reading that you need to bias your input signal.  Makes sense.

Is there a problem due to the bias being wrong by a few percent?  Because now the zero crossing of an AC signal is offset from the virtual/effective ground.  Wouldn't this cause that part of the signal that is within the offset error to be amplified in the "wrong direction"?

I'm looking at a datasheet that suggests to use a resistor divider to add bias to the input signal.  However due to resistor tolerance the bias isn't going to land right at +Vs/2.

[homebrew]

Yes, that would definitively be a problem. And that's why you don't do it that way :-)

On the bottom of the page you will find the appropriate schematic:
http://www.radio-electronics.com/info/circuits/opamp_non_inverting/op_amp_non-inverting.php

The whole idea is to use the capacitor (C2 in this case) as a way to auto-adjust the virtual ground.

[dan3460]

Interesting, could someone explain how this is calculate:

Resistors R3 & R4:   These two resistors form the bias for the non-inverting amplifier, setting the input voltage and operating point. Typically they are set to provide half the supply voltage and therefore they will be equal in value. Values of around 100k? are typically chosen because they determine the input impedance. To incoming signals they are in parallel with each other and the input impedance of the op amp itself is normally much higher and is often ignored - although check for the given op amp. Assuming this and each resistor being 100 k?, then the resiatnce in parallel is 50?

This quote comes from the end of the article when describing the values for each part.
My ohms law calculate that two 100K resistors in parallel will be 50k and not 50ohms. I'm missing something?

[electrolust]

Seems pretty obvious that the missing 'k' is just a typo.

I am more confused about other seeming typos (or not?) on the page.  EG the discussion on AC coupling.  They talk about the possibility to DC couple the circuit and describe how to handle that, but proceed to show a circuit that is AC coupled.

I'm also lost on how to select C2 but I suspect that is more my lack of knowledge than some error.

[John Coloccia]

Seems pretty obvious that the missing 'k' is just a typo.

I am more confused about other seeming typos (or not?) on the page.  EG the discussion on AC coupling.  They talk about the possibility to DC couple the circuit and describe how to handle that, but proceed to show a circuit that is AC coupled.

I'm also lost on how to select C2 but I suspect that is more my lack of knowledge than some error.

Well, the way that particular schematic is drawn is not exactly how I'd do it, but nevermind that. The way it's drawn, C2 is there so that you don't amplify the DC offset. It determines the low frequency response of the circuit. With the resistors, it essentially creates a high pass filter. Choose it large enough that it gives you a linear enough response through the range of frequencies you need to amplify.