No, I did not look at the oscilloscope because I did not.
The power source I use is from lm2576 adj to create 2.5 volts and an AMS1117 to create 3.3 volts and an icl7660 to create a negative voltage of 2.5 volts.
The 2.5V voltage I make with the LM2576, with its 180mA currents, is dropped by several tenths of its gas sensor, with a lower current flow of about 140mA. But since the regulator is adjustable, with a potentiometer, its value I set it again to 2.5 volts, as a result, the voltage drop is offset, but the current does not exceed 140 mA.
Of course, one time, instead of replacing the schottky diode with lm2576, a tantalum capacitor 10 UF did not have any voltage drops and provided the current without a schottky diode. But when I realized the capacitor was removed, instead of a schottky diode I replaced ss34 with the voltage drop problem!!!
I used for the LM2576 a 100UH inductor and a SS34 diode and an output capacitor of 330UF and an input capacitor of 100 UF.

I put the AMS1117 input and output capacitors of tantalum type.

I do not understand exactly what you mean by these sentence, if you can, explain a bit:
Are you using good decoupling for the amp power right at the amps?
I have connected a positive gain feeder to the earth with a 100 NF capacitor and also with an electrolytic capacitor of 1 UF
I put the 1M resistance input pin apart, but did not care.

Thanks again for a practical offer

 

Decoupling capacitors, also called bypass capacitors are small capacitors connected between each power supply pin of an IC and circuit common. The purpose of these is to help counteract inductance in the power supply connections to the components. A few centimetres of connection length can cause spikes of hundreds of millivolts at the IC pins if the IC causes fast current transients. This usually isn't a problem with low frequency analog circuity, but it is still good design practice to use them. They also help to remove high frequency noise from other souces that is on the DC power connections.

100 nF (0.1 µF) ceramic capacitors are commonly used. The leads should be as short as possible and the capacitors should be as close to the IC as possible.

I think there is a possibility some of your circuits' strange behavior may be due to high frequency noise from the switch mode power supply (SMPS) and possibly for the ICL7660. It is not unusual to see noise of hundreds of millivolts on the output of a switch mode supply. When noise that is higher in frequency than the bandwidth of an amplifier gets into its signals, sometimes it can respond in a strange way that can look like an offset voltage or gain error. Using SMPS in analog circuits often causes problems.

In your circuit at #19 try putting a small capacitor (maybe 10 nF) across R7. Also try putting back the 1 M resistor between pins 1 and 5 and put another small capacitor (10 nF) from pin from pin 5 to the circuit common.

What kind of capacitor is the 330 µF at the output of your switcher?

Can you post photographs of your switcher and the circuit you are experimenting with?

 

I breadboarded the circuit in #19 using a MCP602 and it works, .5 volt in .5 volt out no voltage drop. Vcc is 3.3 volts from a linear regulated power supply and not using any filter capacitors. As ebp said check your power supply.
SG