Audio Op-Amp resistor selection?

[Vince Bafetti]

Hello group.

I'm wondering about selecting resistor values for an audio op-amp
circuit. Say I wanted to make a plain old inverting amplifier with
series resistor Rs and feedback resistor Rf. For a gain of -2, suppose
I pick Rs = 10K and Rf = 20K.

I could also have picked Rs = 1K and Rf = 2K.

In terms of susceptibility to picking up RFI from a cell phone, is
there any advantage to using 1K/2K vs. 10K/20K? I'm guessing the
higher resistances would create greater noise voltage due to a given
noise current...but then, the lower resistances might start adversely
affecting whatever circuit I'm connecting the input to (input
impedance goes from 10K to 1K).

Thoughts?

Thanks,
Vince Bafetti
[email protected]

 

[Stefan Heinzmann]

Hello group.

I'm wondering about selecting resistor values for an audio op-amp
circuit. Say I wanted to make a plain old inverting amplifier with
series resistor Rs and feedback resistor Rf. For a gain of -2, suppose
I pick Rs = 10K and Rf = 20K.

I could also have picked Rs = 1K and Rf = 2K.

In terms of susceptibility to picking up RFI from a cell phone, is
there any advantage to using 1K/2K vs. 10K/20K? I'm guessing the
higher resistances would create greater noise voltage due to a given
noise current...but then, the lower resistances might start adversely
affecting whatever circuit I'm connecting the input to (input
impedance goes from 10K to 1K).

Thoughts?

Assuming that you use voltage feedback amplifiers, I would think that at
audio frequencies you won't see a lot of difference. Of course, at some
point the distributed capacitances and inductances will start to show,
once the resistor values reach a certain point. The main argument
however is the one you pointed out already: Input impedance.

There are other things to watch out for, however, depending on your
actual application:
o Input offset current
o Noise (you mentioned this, too)
o Power dissipation

I assume that you have to protect against RFI pickup anyway, regardless
of resistor values, by using proper filtering, grounding and shielding.

Make your resistor values as low as you can without bumping into other
limits. That should be a reasonable starting point.

Cheers
Stefan

 

[mikem]

Vince Bafetti wrote:

One other big limitation is the ability of the amplifier to source or
sink output current... Some cmos op amps are hard pressed to source or
sink more than a few hundred uA, so the gain setting resistor has to be
sufficiently high impedance so as not to "overload" the amplifier.

Everything in analog design is a compromise...

MikeM

 

[Mac]

Hello group.

I'm wondering about selecting resistor values for an audio op-amp
circuit. Say I wanted to make a plain old inverting amplifier with
series resistor Rs and feedback resistor Rf. For a gain of -2, suppose I
pick Rs = 10K and Rf = 20K.

I could also have picked Rs = 1K and Rf = 2K.

In terms of susceptibility to picking up RFI from a cell phone, is there
any advantage to using 1K/2K vs. 10K/20K? I'm guessing the higher
resistances would create greater noise voltage due to a given noise
current...but then, the lower resistances might start adversely
affecting whatever circuit I'm connecting the input to (input impedance
goes from 10K to 1K).

Thoughts?

Resistors in the range of 10k - 100k should be fine. I would probably try
to keep the input resistance as high as possible. A little known fact
about the inverting op-amp configuration is that that you should not
directly ground the positive terminal, but instead ground it through a
resistor equal to the parallel combination of Rf and Rs. This minimizes
offset errors.

If you enclose the op-amp in a metal box, it shouldn't pick up anything
from the cell phone. You can also add a feedback capacitor in parallel
with Rf to roll-off the gain at higher frequencies. If you are going for
high fidelity, put the 3dB point at 40 kHz or more. If you are going for
voice only, put the 3dB point at 5 or 10kHz. Also make sure you have good
high-frequency bypassing at the op-amp power supply inputs.

Thanks,
Vince Bafetti
[email protected]

Mac

 

[Walter Harley]

If you enclose the op-amp in a metal box, it shouldn't pick up anything
from the cell phone. You can also add a feedback capacitor in parallel
with Rf to roll-off the gain at higher frequencies. If you are going for
high fidelity, put the 3dB point at 40 kHz or more. If you are going for
voice only, put the 3dB point at 5 or 10kHz. Also make sure you have good
high-frequency bypassing at the op-amp power supply inputs.

A feedback capacitor will not do anything to eliminate RFI. Feedback only
works at frequencies where the opamp has gain greater than unity. It is a
useful technique for reducing noise and controlling instability, but it is
not useful against RFI. In fact, depending on the circuit it may only serve
to help couple RFI that is coming in through an output connector back into
the input of the opamp.

 

[Mac]

A feedback capacitor will not do anything to eliminate RFI. Feedback only
works at frequencies where the opamp has gain greater than unity. It is a
useful technique for reducing noise and controlling instability, but it is
not useful against RFI. In fact, depending on the circuit it may only serve
to help couple RFI that is coming in through an output connector back into
the input of the opamp.

The OP has a noise problem of some sort. It may not be RFI. Rolling off
the gain of the audio amp at higher frequencies could help if the noise is
something other than RFI, and is probably a good idea anyway.

The nice thing about a cap like this is that you can leave it off if it
doesn't do any good or makes matters worse. ;-)

Other than that, I defer to your greater experience and knowledge. ;-)

Mac

 

[Tony Roe]

I've generally got best RFI immunity with 3 similar valued caps - from output to
inverting input, across the inputs, and from non-inverting input to ground
(unless it's already grounded). Plus (important) series isolating resistors on
inputs and output.

The OP has a noise problem of some sort. It may not be RFI. Rolling off
the gain of the audio amp at higher frequencies could help if the noise is
something other than RFI, and is probably a good idea anyway.

The nice thing about a cap like this is that you can leave it off if it
doesn't do any good or makes matters worse. ;-)

Other than that, I defer to your greater experience and knowledge. ;-)

Mac

Tony (remove the "_" to reply by email)