Back in the day when microprocessors were mostly nmos and power hungry we designed a data acquisition unit using the RCA 1802 CMOS processor. I was able to power it for a short time using three quarter lemons, three carbon rods and three zinc rods (got about 4V) but the lemon slices had to be squeezed to get enough current out to work the device and a small LED display. Even so, it impressed people.
I think that this could be made to work using a single wire from some steel wool, and by paralleling a number of zinc coated nails and copper boatbuilding nails (roves), then squeezing the lemon to break the cells and free enough acid to get a decent current for a few seconds. But I’m really not going to carry out the experiment.
