I've been busy on other things for the last while, so don't really have much progress to report.
First, a few days ago I was going to order some MCU chips, and found the place I was buying them from was sold out – with expected delivery in October or November! That kind of put me off...
But I decided that there were two things I needed to prototype / investigate, the MCU (particularly the ADC) and the input section. I don't have the analog switches or actual op-amps yet either, so I couldn't do a proper prototype. So I just did a quick test with a different op-amp (TLC274) as to how the input impedance affects the output. Result: a 1 megohm pot acting as a voltage divider applying approximately 2 – 10 mV to the op-amp (configured as a voltage follower) had a very similar output voltage whether connected directly to the positive input, or connected via 20 megohms – and connecting another 10 megohms to ground as a voltage divider produced the expected voltage as well. So looking good so far.
Today I discovered two things: one, I had a major mistake in my design; and two, the MCUs are back in stock! So I'm about to order them, the analog switches, the LCD, and a programmer for the MCU.
The mistake in my design was simple: U103, the reset chip, was powered by 3.3 V instead of directly from the battery! As it forces the MCU into reset when its supply is below 3.5 V, it would never have come out of reset. Fortunately, it was easily fixed as +BATT was already available nearby on the PCB.
Unfortunately, this probably never going to get past a prototype. To get the PCB made for me would cost nearly as much as the rest, and put the total cost far higher than I'm willing to pay. I could possibly make it myself, but I'm not particularly confident in my abilities to make one that large, at least not unless I can improve the quality of the masks for exposing the photoresist – which would probably require a better printer to prevent the light bleeding through.