Can I use an INA128 as a diferential amplifier?

Question

I need to measure the voltage from a reference solar panel. We've calibrated a shunt resistor to produce a mV output, and we're using an 8-channel datalogger to record these values.

The solar panel has an internal connection between its solar cells in the array as shown in the image. I cannot remove that internal connection.

Due to this internal connection, the datalogger is unable to read the values correctly because its channels are not independent; the negative pins are all connected. To address this issue, I am considering the use of an INA128 op-amp as a differential amplifier with independent power supplies to isolate the signals. The recorded values are inconsistent, and the gain is not 1 (Rg = NC.)

Any idea what I might be missing here?

---

I wanted to share my power supply scheme with you. I used two Zener diodes to create a symmetric power supply because I was concerned that it might be the source of the error. As you mentioned, the input voltage has a range, which could indeed be the issue since I am trying to measure a signal between 0 and 35mV.

The datalogger screenshot is below

Answer 1

What you are proposing could work. There are a few conditions.

Shunt resistor voltages should be within the input range of the in-amp. For INA128 it's from (V-)+2V to (V+)-2V. If the positive supply rail is +36V, and the negative supply rail is GND (i.e. single supply operation), then your input range would be from +2V to +34V.

Your bank of in-amps shouldn't be completely floating with respect to the solar panel.

The Ref terminals of in-amps should be connected to a low impedance node. Usually it's ground.

update:
Channel 2 on the logger shows 16.7V . I'll make a working assumption that it's the common mode voltage of the highest shunt (and differential mode is between 0 and 35mV). The in-amp is powered with ±12.5V . A 16.7V signal is outside of the input range of the in-amp.

Answer 2

Your zener diodes are directly across the 25V supply, which will break the diodes. You need some resistance to operate them as shunt regulators:

^{simulate this circuit – Schematic created using CircuitLab}

The above will afford you about 25mA of current in each rail, enough to power several INA ICs. The INA128 requires inputs that never get closer than 2V to each supply potential, so you can't apply potentials outside the range −7V to +7V, relative to the central 0V ground. This shouldn't be an issue, since three shunted PV cells in series produces a potential difference of \$35mV \times 3 = 105mV\$, and the common mode voltage for all of them will therefore be very close to zero. Even eight of them will never produce a common-mode potential over 0.3V.

The output will have similar constraints, being unable to swing beyond ±7.5V. If we use a gain of 100 (\$R_G=511\Omega\$), each INA should produce +3.5V maximum.

So, with input common mode potential near zero, and an output swing well within the supply limits, the INA128 should perform perfectly well. The arrangement might look like this:

^{simulate this circuit}

I hope it goes without saying that the 25V supply source needs to be isolated from all other measuring equipment, since the ground being used here has a very different potential from that source's own terminals.