Op amps with single supply (again)

[elki]

Hi everyone, I am trying to understand the behavior of the inverting op amp configuration (x10) when using with single supply. I tried several op amps (LM7171, TLE2081, etc.) with the same observation, so I believe that the issue is of a general type. Naively, I would think that one should fix the feedback resistor to one of the recommended values (e.g. 1k) and then set the input resistor for the required gain. However, I see a very nonlinear dependence of the output gain on the input resistor. When varying the feedback resistor instead, I get a much better linear gain behavior, however there is a substantial distortions of the output shape. Does anyone know if there is a conceptual problem with this circuit?

[Doctorandus_P]

At the moment the system gets powered, the non inverting pin of the opamp jumps from "GND" to "half supply voltage", and all capacitors are discharged. It takes some time for the capacitors to charge and the system to stabilize, and this is normal and expected. For audio amplifiers for example, they usually have a relay at the output that only connects the output to the speakers (or the line out if it's a pre-amp) after a delay and the circuit has stabilized.

[tszaboo]

Also, you should try some opamps from this century. There are plenty of them that will work nicely with 5V systems, with rail to rail inputs and outputs, small offset voltages, output drive to small capacitances and inputs that can handle multiplexing and other transients.

[newbrain]

However, I see a very nonlinear dependence of the output gain on the input resistor. When varying the feedback resistor instead, I get a much better linear gain behavior

What's your input signal? Have you considered that the 5 pF input capacitor will have a very high impedance at, say, normal audio frequencies?
The total gain will be affected heavily by its impedance, in series with the input resistor.
OTOH, changing the feedback resistor will affect the gain as expected.

[elki]

Thanks a lot - this is indeed what seems to be happening.

However, I see a very nonlinear dependence of the output gain on the input resistor. When varying the feedback resistor instead, I get a much better linear gain behavior

What's your input signal? Have you considered that the 5 pF input capacitor will have a very high impedance at, say, normal audio frequencies?
The total gain will be affected heavily by its impedance, in series with the input resistor.
OTOH, changing the feedback resistor will affect the gain as expected.

[CaptDon]

Can't imagine where you ever came up with the idea to fix the value of the feedback resistor and adjust gain with the input resistor. Kind of like holding the gas pedal on the floor and adjusting speed with the brake pedal.

[newbrain]

fix the value of the feedback resistor and adjust gain with the input resistor

Considering the OP, it seems to be done for learning purposes, and in that context it makes sense.

In an actual design, the input resistor dictates the input impedance, so there's much less leeway in defining it.

As a special case, for current feedback amplifiers the feedback resistor must be in a very specific range for optimal performance, so the input resistor needs to be chosen to obtain a specific gain.

[Vovk_Z]

Does anyone know if there is a conceptual problem with this circuit?

An inverting amplifier itself with LM7171 and with 10k/2k gain is fine and linear (10/2 = gain is 5). It is just not clear what you wanted to achieve with a 5 pF capacitor at its input (together with a 2 k resistor in series). That's a high-frequency filter with f(-3) = 16 MHz.  Quite high frequency.
What is your task? Do you want to understand the behavior of the inverting op amp configuration or do you want to investigate high-frequency staff (by working with tens of MHz)?

[TimFox]

Although the LM7171 is fast enough to work with a 16 MHz hpf at the input, it is indeed strange to use 5 pF there.
Is that a typo for nF or uF?

[elki]

Thanks. It is not a typo. The idea is to calibrate the amplifier on a small input charge.

Although the LM7171 is fast enough to work with a 16 MHz hpf at the input, it is indeed strange to use 5 pF there.
Is that a typo for nF or uF?

[TimFox]

A charge-input amplifier normally doesn't have the series input resistor (to the inverting input), but uses an integrating capacitor in the feedback path to convert the input charge to an output voltage.