Precision op-amp configurations

[ricko_uk]

Hi,
few op-amp related questions...

1) when designing precision instrumentation using single supply, if signal polarity (i.e. inverting or non-inverting the signal itself) is NOT an issue which op-amp topology is preferred, inverting or non-inverting? And why?

2) unrelated to question 1, if you try to make any phase measurements, again which is best, inverting or non-inverting? And why?

3) with reference to the attached picture (single supply operation), if I want amplify/pass extremely low frequencies like 0.1Hz do I still need the capacitor highlighted in red? And generally how do I calculate its value?

4) still with ref to the attached picture, can I safely remove the red capacitor or would it create issues in time (perhaps drifting)?

5) when mixing signals (i.e. summing amplifier) again in single-supply low-noise designs, is it preferred to use inverting or non-inverting?

Thank you

[TimFox]

If the signal is positive-definite, such as a voltage reference, then the non-inverting configuration is much easier to use with a single supply.
For the inverting configuration, you need a voltage divider to establish a middle voltage on the non-inverting input, reference the input signal to that voltage level, and the output voltage can then go negative with respect to that terminal.  In that case, I find it better to use a true split supply.
AC errors, such as phase shift, depend on the "noise gain", which can be lower for the non-inverting amplifier.  Also, lower resistors can be used for the feedback network, giving lower voltage noise from the non-inverting amplifier.  In the inverting amplifier, there is always a reasonably-large resistor in series with the signal, increasing the voltage noise.
The red capacitor is almost mandatory for the non-inverting circuit shown:  if the feedback network gives high closed-loop gain, with a small resistor from the inverting input to ground, the output voltage will go very high (quiescent conditions).  Even for reasonable closed-loop gain > 1, the capacitor will reduce the DC output drift.  Since there is also an input capacitor, due to the DC bias at the non-inverting input, your signal gain will fall to zero at DC, anyway.

[ricko_uk]

Thank you Tim!

How do I choose the value of the red capacitor? I am interested in passing frequencies as low as 0.1Hz possibly even lower.

Thank you

[TimFox]

To first approximation, as you go down in frequency, the gain will start to decrease when the reactance of the red capacitor equals R2 in series with it.  Below that frequency, the impedance of that branch increases with the reactance of the capacitor.  At zero frequency, where the impedance of that branch is infinite, the asymptotic value of the gain is +1.

[Vovk_Z]

How do I choose the value of the red capacitor? I am interested in passing frequencies as low as 0.1Hz possibly even lower.

It seems to me sometimes it can be easier to make a fully DC amplifier but not 0.1 Hz high-pass if there are some phase requirements etc.

f3 = 1 / (2*pi*R*C), where f3 is the frequency with a -3dB drop of an amplitude, and 90 degrees of phase shift if I'm not wrong.
So, if you are interested in low phase shift at 0.1 Hz you have to move f3 frequency 10-20 times lower than 0.1 Hz. That is 0.01-0.005 Hz.

[ricko_uk]

Thank you Vovk_Z!

[ricko_uk]

With ref to the original picture (attached below again), if the op-amp (still in the same non-non-verting configuration) is powered from a dual supply, do you still need the red capacitor?

[David Hess]

1) when designing precision instrumentation using single supply, if signal polarity (i.e. inverting or non-inverting the signal itself) is NOT an issue which op-amp topology is preferred, inverting or non-inverting? And why?

The inverting configuration is inherently more accurate because it lacks error from the limited common mode rejection of the operational amplifier.  This becomes a consideration when the highest accuracy is required but an inverting amplifier is not the only way to handle it.

2) unrelated to question 1, if you try to make any phase measurements, again which is best, inverting or non-inverting? And why?

The non-inverting configuration has a higher bandwidth because of lower noise gain but this will only be significant at low gains.

3) with reference to the attached picture (single supply operation), if I want amplify/pass extremely low frequencies like 0.1Hz do I still need the capacitor highlighted in red? And generally how do I calculate its value?

The resistor in series with the capacitor determines the break frequency.  The capacitor serves to remove DC errors from the amplifier.

4) still with ref to the attached picture, can I safely remove the red capacitor or would it create issues in time (perhaps drifting)?

Removing it would allow the amplifier's DC and low frequency error to contribute to the output

5) when mixing signals (i.e. summing amplifier) again in single-supply low-noise designs, is it preferred to use inverting or non-inverting?

The inverting configuration has the advantage of isolating the inputs from each other which may be important.

[voltsandjolts]

Recommended reading:

https://www.analog.com/en/analog-dialogue/articles/avoiding-op-amp-instability-problems.html#