Caps for AC Coupled Op-Amps

[beaker353]

Started playing around with an LM124 op-amp in single power supply config to act as a signal buffer for an audio based circuit I'm working through.  Going to pass an audio signal with a max level of about 1V RMS with a gain of 0.  I've wrapped my head around the voltage divider idea to bias the signal with half the Vcc and the need for the caps on the in and out to decouple the DC offset.  I am having a hard time however, figuring out the values for the input and output caps.  I imagine a higher value would be better, but wouldn't this roll off the lower end of the AC signal passing through it?  I don't need high fidelity, so some roll off isn't a problem.

I'm still working through learning electronics, but often get pushed projects that are a little above my head that I just need to figure out.

Thanks!  -EM

[ipman]

Caps required are calculated based on what frequency you want to pass. Sometimes you need to parallel some different values to pass more bandwidth if required.
Give us more details.

[ejeffrey]

The AC coupling capacitor will set the low-frequency roll-off while leaving the high frequency signals unmolested.  The capacitor needs to be large enough that the cutoff frequency is below about 20 Hz for full range audio.  The way you calculate the cutoff frequency of high-pass filter is 1/(2*pi* R1||R2 * C).  R1||R2 is the parallel resistance of the resistors in the voltage divider that set you DC bias and C is the coupling capacitor.  If you have a pair of 20 kohm resistors setting your voltage divider the parallel resistance is 10 kohm and a 1 microfarad capacitor will give you a 16 Hz cutoff frequency.

One word of caution: if you end up doing anything more complicated than a single gain stage, it is usually easier to set up a virtual earth for single-supply applications.  If you have cascaded stages of op-amp circuits it is very handy to have them all share the same virtual ground reference so you don't have to AC couple each stage.  This also allows you to heavily filter your ground reference.  Using a simple voltage divider to bias your signal has the problem that supply ripple is coupled into the signal path very strongly.

[Zero999]

Started playing around with an LM124 op-amp in single power supply config to act as a signal buffer for an audio based circuit I'm working through.  Going to pass an audio signal with a max level of about 1V RMS with a gain of 0. 

Well with a gain of 0, the output signal will be 0V so you don't need any capacitors at all. I think you mean a gain of 1.

I've wrapped my head around the voltage divider idea to bias the signal with half the Vcc and the need for the caps on the in and out to decouple the DC offset.  I am having a hard time however, figuring out the values for the input and output caps.  I imagine a higher value would be better, but wouldn't this roll off the lower end of the AC signal passing through it?  I don't need high fidelity, so some roll off isn't a problem.

I don't think the low frequency response is going to be the problem. The LM324 isn't really fast enough for audio, its slew rate is 0.4V/us and at 20kHz the gain is only 34dB (50). Although this probably won't be a problem for your application, you only need a slew rate of 0.355V/us (see calculation below) and a gain of 1 but we don't know what you're using the other op-amps in the package for. If the output voltage is higher than 3.18Vpp, 1.59Vp or 1.25Vrms thie signal will become distorted or the gain might not be high enough to ensure sufficient negative feedback.

sr = 2pi*Vpp
Vpp = 1*sqrt(2) = 2.83V
sr = 2pi*20*10³*2.83 = 0.355*10³/v = 0.355V/us
http://en.wikipedia.org/wiki/Slew_rate

[beaker353]

Sorry for the late reply, most of you know how life get from time to time...  Anyhow...  Yes Hero, I meant to say a gain of "1".  I come from the pro audio world, so I think most natively in dB gain in which "O" is equal in to out.  I don't see any reason I would ever expect to feed or send voltages over 1VRMS, probably more in the 500mVRMS range.  As this is intercom signal and not anything high fidelity,  so some restriction of the bandwidth wouldn't be a problem, it might even be a good thing.  The plan was to add some low and high pass filters feeding the final stage anyhow to remove some top and bottom end.  Jeffrey, I too saw some information on creating a virtual ground to make the circuit more stable. I've attached a screen shot of a circuit snippet from the LM124 data sheet which I am using at the base for my project. Is my understanding correct that this is this a virtual ground referenced circuit?

Thanks!