Do I always need an output coupling cap?

[Dominique]

My problem:

Given my simple microphone preamp meant to be plugged into usual "line-in": After the signal leaves the op-amp there's nothing more but the output coupling cap...

Now if I use a very low offset voltage OP-amp (e.g. an OPA 2228 P op-amp with a typical input offset voltage of 5 uV !! (75uV at max.). Now the typical mic signal I get from my mics is about 1mV (in my applications), so the offset-to-signal-ratio (1:200) should be good enough to leave the cap away, isn't it? (Given the offset isn't at it's max of 75uV!) Or is there something else to consider in addition?

In any case I like the idea of leaving the cap away which would make my circuit even simpler&smaller!

Any comments are greatly appreciated!
Cheers!
Dominique

[Leolabs]

If your OPAMP is running on a balanced supply voltage(ie +15V:0V:-15V),then coupling cap is not neccesary,BUT please submit your schematic for conformation.

[Dominique]

thanks for replying

Yes, my op-amp runs on dual supply.
The circuit is pretty much like the one at the bottom.

The amp is powered by batteries. And in my previous version, I left out the caps C3 to C6, still the amp worked great (it's housed in a closed metal box so not prone to RFI at least), but I wonder if I better include at least C3 and C4 (which I'd place next to the op-amp pins like everyone's telling me). Sorry if that's a different subject!

Cheers
Dominique

[Leolabs]

What kind of mic are u using???Wondering what is R7 doing there???

C3&C4,no harm to include them

[Eva]

Remember that OP-AMP offset specifications are only achieved when both inputs are driven from the same DC impedance

Your circuit will allways show high offset on the output, no matter what OP-AMP you choose, because one input is seeing less than 270 ohms DC resistance and the other sees 470k

[janusz]

Hi,

Just a few humble comments. Firstly, I'd keep R3 as small as practically possible mainly because it generates some imput noise. Check what is the minimum acceptable load for your mic and use that value plus 25% just case. If it is possible to reduce it substanially then the DC offset will also be reduced.

The output of the OP-AMP has its connection to the ground via R2 and R5. No problem here. The question is what happens after the OP-AMP output. If it is connected to another input via some input cap the only thing you have to check is the value of Zin as with the C2 removed it runs in parallel with R2-R5. Its value may need to be increased but not necessarily.

But first of all measure the DC offset of your OP-AMP.

Cheers,

[Eva]

R3 only contributes to the noise at very low frequencies [rumble]

At audio frequencies the source impedance of the microphone dominates

Resistors show a theoretical noise voltage proportional to the square root of the resistance but this noise voltage source appears in series with the own resistor, so the noise voltage suffers a strong attenuation when the noisy resistor is shunted externally with a much smaller impedance

[Jennice]

I don't know if you need to consider it (depending on what's after the mic-amp), but remember that your output off-set == [input off-set] * gain. In your case, it would probably be more like 0,5mV to 7.5mV. Could this cause a problem with the circuit connected to your mic-amp output?

Jennice

[janusz]

Hi Eva,

You right to a point. Voltage noise Vn contribution of a resistor equals square root of: 4*k*T*R*df;
while current noise (In) equals square root of: 4*k*T*df*1/R;

where T is temperature in degrees Kelvin and k is Boltzman's constant

and df is the relevant bandwith

so the noise contribution of a resistor is accross the relevant bandwith (here that of a microphone) not only at low frequencies.

But it is true that R3 is in parallel to the source so our R = R3 || Rs and if Rs is small than R3's contribution is marginal. My fault I did not inspect the diagram carefully.

Assuming that if Zmic (500ohm) is the source impedance and constant then R3 does not contribute much noise accross the mic's bandwith. As source impedance is not really a constant R3 will start to contribute near the resonance level of the mic which is probably well beyond the bandwith and well beyond the issue of keeping C2 in or not.

If C2 is removed then 100ohm may be placed in its place. Looks that Zin denotes input impedance of the next stage and not a resistor in the mic preamp circuit. If it was a circuit resistor then it is not needed with C2 gone. Zin however, looks like a input impedance of the next stage so it may need an input cap or not subject to what Jennice said about DC offset magnitude.

Cheers,

[Dominique]

Now you got me all mixed up!

I'm very thankful for all of that information!
It's all very interesting and reminds me doing my homework and read Horowitz/Hill more often!

I'm not sure if I understood the Rs (=Zmic) / R5 thing right.
Thinking practically:
If I feed the preamp with a 1kHz wave and the mic impedance would be about 270 Ohms @ 1kHz, that 1kHz wave should only have the OP-amp's minimum offset, but at frequencies with a different (mic) impedance, the offset would be larger...? But well, there's always a signal with virtually all of the frequencies going into the amp, because of the mic's noise, isn't it?

Oups, I remember that I wanted to realize R5 as a potentiometer or two or three resistors which are selected via jumper, so R5 can change in order to adapt the gain to various recording purposes ("quiet talking" up to "concert")
Instead I could adapt the feedback resistor, but I remind having read that the feedback should be kept short and simple. Oh no! The more information I get, the more problems appear

Regarding the input impedance of the next stage: Yes, I forgot to say: The impedance of the next input (normal line-in of e.g. MD recorder, mp3 player/recorder, possibly tape recorder) is represented by Zin. It ranges from 10k (my soundcard), often 12-14k (many mp3- and MD-recorders) to 47k (often seen). I want the circuit to work fine with all of them.

@Jennice
I don't think that a few mV of offset could cause a problem here.

@Leolabs
That mic is an electret mic and needs to be fed. Btw, I'll have to lower R7, because 10k was just fine for the 9V battery.

Well, I'm totally sunken into texts about op-amps now and forgot my sourdough bread in the oven... it's all your fault! lol

Just to warn you: Possibly I'll come up with some further questions!

Cheers,
Dominique