Mfcttrf

Just a pure understanding of the workability of this circuit.

It is powered from +/-24V and there are regulators for +/- 12V on top of the schematic.

U4 is wired as non-inverting amp with a gain of 2.

We got 8 CPC1918 OptoMOS relays, presumably activated by sending enough voltage into the "PADx" inputs to light the internal LED. These select feedback resistors between 10R and 10Meg (R6,8,10...). So we can just look at one of the feedback paths, for example the one with 10R resistor, and pretend the others don't exist.

U4 is inside U2's feedback loop, it is a high current opamp, and there is a 10R 50W resistor, this means that U4 is used as a high current buffer with a gain of 2. If we take it out of the picture for a moment, the resulting schematic is a transimpedance amplifier:

In this circuit, the opamp adjusts its output to hold its "-" input at the same potential as its "+" input (0V), and if we consider the input bias current to be zero, then I1 flows into Rf, thus the output voltage is I1 * Rf. The opamp's negative input acts like a virtual ground, able to sink and source current, as long as the opamp is able to adjust its output voltage to make sure I1 flows through Rf.

Your schematic is the same thing, with an extra "reference electrode" input which allows the user to choose at which voltage the other input should be held, and it has selectable Rf via the OptoMOS relays.

It is a selectable gain transimpedance amplifier, able to process quite high currents (a few amps as per LM1875). The voltage output is in the lower right corner of the schematic.

If the circuit diagram makes no sense to you then you need to incrementally approach the circuit and derive the functionality of each block in it.

If the schematic represents something that has been successfully built, then you have to assume it's both workable and designed for the task. If you assume it has not been built or used successfully you are at a very different level of required analysis (and I assume you are not there yet).

The circuit as shown DOES NOT WORK, though I'll cover why later in the answer

Start your analysis like this:

1. Collect the data sheets for the active components and use them to gain an understanding of their use and configuration in any given implementation.


2. With an understanding of the active components involved, try to block out the schematic into logical functions or sections

So for the schematic you show:

1. 
   78M12 and 79M12


2. LT1367


3. LM1875


4. G3VM-61G1

Given the major active components you can now block out the schematic:

We can make some assumptions on the gains for the circuit:

1. Since the HiGain block is powered by +/-12V and the opamp is rail to rail then +/-12V is the maximum output voltage.


2. The Power Gain block is using +/-24V, but can only ever output +/-12V due to the limitation of the Hi Gain block.


3. If +/-12V is the maximum output of the LM1875, the current levels for the input range must be as shown ….1.2uA through 1.2A full scale deflection.


4. This maximum 12V output is divided down in the Output Scale Adjust to 8.4V full scale deflection.

Now we get to why the circuit does not work. It has several mistakes.

1. In the power supply block there is insufficient output current (unless there are other components/loads that we can't see) for the regulators. You need at least 5mA drawn from each supply to keep the regulator output voltage within spec.


2. The input clamps before the Hi Gain block feed back into the power supply. This is OK where you absolutely know that the power supply load is enough to absorb that energy, but in this case since the supply appears underutilized, there is the potential to seriously raise (or lower) one of the supplies.


3. The feedback resistor FET switches are only rated at 400mA. It appears that full scale for this amplifier is beyond this rating on the highest range input. If the maximum fullscale is actually less than 400mA, then the switches don't work because of the leakage currents for the lower ranges.


4. (And this is HUGE) The FET switches are shown incorrectly, the input/output and the LED drive are not on the pins shown. Perhaps the drawing package had no suitable device and the designer made their own package (and screwed up). If you laid out a PCB from this schematic it would not work.


5. One device (U5) is marked with a different designator, and this is poor design if it's just an alternate part. I did not look it up to see if it's the same as the one shown for U6-11.


6. At the lowest detection ranges, FSD is more than the leakage current for the individual FET switches. For example, consider that U11 is turned on with a 10M Ohms series feedback resistor. The leakage currents for U5 through U10 are all in parallel, so the lowest range is unlikely to work at all. In fact it's doubtful that anything below +/-1.2mA would provide any reasonable accuracy.


7. The series resistance of the FET switches is 1-2 Ohms, so for the highest range implies at least a 10% error.

peufeu was right in his answer, that the INTENT of this circuit was that of an accurate transimpedance amplifier used as a current to voltage translator. But missed the errors in the circuit.

Doing current measurements in the uA range is challenging and this circuit looks like it was a first attempt at a viable circuit. And they missed the mark.

Hope this helps.