There is nothing special there, its a convenient matrix to put in front of a differential amplifier.
I'm afraid that didn't help....
Study the logic diagram in the datasheet.
I did. There is nothing to it as you say.
The convenience is that a0 and a1 control eight switches. You don't need eight logic lines because only certain patterns make sense in this scenario.
Well not really, two address lines can only control four switches, logically speaking. That was the reason for my confusion
I'll rephrase :
Will it ever make sens to provide different analog inputs on one channel with two differential inputs?
If yes, what for?
If no, why does this setup exist, why not just use a single ended analog mux like AD658? (which has 8 channels, but still same pin-count)
cheers
So its very hard to talk to someone who seems to be overcomplicating things. Obviously one control line, being an output, can be an input to any number of switches. Any number. So, the reason, here, is because inputs to a differential amplifier or an instrumentation amplifier come in pairs, V+ and V-. It is *not* generally useful to have any V+ paired with any V-. So they want to be in synchrony. That is what that switch does for you. Maybe it is so simple you are having a hard time with it!
When you do need more flexibility you simply *do* use separate multiplexers and then you need a microcontroller or something to make it work.
Consider the audio analogy and realize that you do not want to have a common ground. You will see that whenever one input is switched in that requires exactly something like the ADG659. (Hint: does it make sense to have an audio input where V+ comes from your cartridge and V- comes from your mp3 player?)
Hope this helps.