2.4 Hz pulses are probably from autozero. Switch it off to check.
You're right, they do go away when I disable autozero.
The Ohms current source is in deed confusing. As far as I understood it they use the opto-coupler U537 in analog mode for the current regulation loop with the higher current ranges. U324 does the loop control.
U333 seems to be the regulating amplifier for the very low current source (Hi_ohm). There is a separate current source for the lowest current, that used R354 as it's reference resistor.
Thanks for all the information, it's starting to make more sense now. Yeah, the manual mentions that the current source for the mega/gigaohm ranges uses R354 through OHM ISOURCE, instead of the regular OHMS path through the K300 relay and protection circuit.
I've managed to repair the unit, but it wasn't very satisfying. I ended up connecting a logic analyzer to A/D IN, A/D_DATA, D_OUT, R1_STB, R2_STB, and DC_STB, in order to figure out what the unit was doing/measuring in the open circuit calibration phase that was failing, since the manual provides the full bit patterns for the shift registers. I happened to do a four-wire resistance calibration this time, instead of a two-wire resistance calibration, as I had previously gotten lazy trying to reproduce the problem, since it only happened on the last phase of comprehensive calibration. This time, the calibration succeeded with no errors, so I removed my logic analyzer/probes and re-attempted calibration, which again succeeded, so I fully re-assembled the unit and finally performed actual calibration. I can only surmise that perhaps the fault was indeed in one of the analog switches, which I fixed when I replaced them, but didn't notice because I got sloppy with the calibration.