My first task is to figure out how the connections can safely interface with the microcontroller. Why first? Because if I can't get this working, I can't go any further.

All the standard functions are needed, and possibly one or two more. So I've got to be able to measure current, voltage, and resistance. Each has its own challenges, but I think I've worked them all out. What makes it difficult is that any combination of terminals must behave itself if presented with the full rated voltage (in this case 600 V because I'm aiming for compliance with CAT III 600 V, and this is one feature the standards require).

Current:
It is normal for current measurement to have dedicated terminals, probably because relay contacts can have too much resistance (e.g. 0.1 ohm). I'll go this route too. The voltage isn't going to be a problem because we're just measuring across a very small resistor. And if it is, some diodes can clamp it – they'll just make the fuse blow faster, because it definitely was going to anyway.

Voltage:
Basically a voltage divider with suitably high resistance. The only detail is switching the resistors in and out (for auto-ranging), but an analog switch / multiplexer can do that OK (its resistance must be accounted for though, so I'm going to look at using MOSFETs instead). Once again, some diodes can easily clamp the voltage within range. And there is no need for a fuse, because any current they draw is coming through about 10 megohms. But still use some MOVs or similar to suppress transients.

Resistance:
This is the one where I've had the most problems. It has to be relatively low resistance because we need to provide an output of around a milliamp with a reasonable voltage left for diode measurements (in other words, about 2k to allow 1 mA into a load with a 1 V drop]). But it still has to cope with 600 V on its input without passing that on to the micro. Clamping doesn't work on its own, because that would be a power dissipation of at least 180 W. But it can be disconnected immediately if the voltage exceeds the limit. I intend to use a solid state relay for this. The clamping will prevent a high voltage being applied to the micro for the millisecond or so that it takes to turn off.