Is it crazy to run a toolchanger with one shared extruder and quick-connect wiring?

I ran a similar setup on a 350 build, three hotends on a rail with one shared extruder, and the connector question is the one that bit me first. For a 40W heater at 24V you're looking at under 2 amps, so honestly almost anything with clean mating surfaces will handle the current fine. The real problem is micro-movement at the contact. If the connection can flex or vibrate even slightly while hot, you get intermittent resistance and that shows up as temp noise or the heater PWM hunting. I ended up using gold-plated spring-loaded pins on solder-in pads rather than pogo pads on bare PCB, because the pads oxidize and the contact pressure drops. Whatever you pick, make sure the connector body itself takes the mechanical load, not the solder joints. XT30 passthroughs work but they're bulky on a toolhead and the retention is all in the friction of the shells, wich loosens with heat cycles.

On the docks, yes, printed parts creep. I printed mine in regular PETG originally and after a couple months of long ABS sessions the pockets had opened up just enugh that the tools sat a half millimeter lower each time. Not enough to fail, enough to mess with pickup repeatability. I reprinted in ABS with higher infill around the magnet pockets and added a hard mechanical stop the tool rides against, so the magnets only locate, they don't carry the alignment. That was the fix. A detent or stop taking the docking force means creep just doesn't matter anymore. For the macros, with one shared extruder and swappable hotends you're basicly doing what the index approach does. Klipper doesn't care which heater exists where, you just define each hotend as its own heater and sensor in the config even though they physically share the carriage. Your tool select macro turns off the old heater and fan, turns on the new ones, and applies the saved gcode offset for that tool. Treat the hotend as dumb hardware and keep a variable per tool for offsets and temps. Works fine as long as every tool's thermistor is wired to its own pin, don't try to multiplex those.

Agree on the dock creep, that matches what I saw. Mine held fine cold but after a week of long ASA runs the rail had walked sideways a bit, the screws into printed brackets had just eased off from the heat cycles. I ended up bolting the dock rail to a strip of aluminum angle screwed into the frame instead, and added a quick touch-off in the select macro after pickup so it complains if the offset drifted. Caught a loose magnet that way before it turned into a dropped tool. Curious how heavy your three tools are on that monolith, mine started parking sloppier once I pushed the accel up.

Reading this while I plan basically the same thing on a smaller build, only two tools instead of three. I've been testing some magnetic spring pins I had left over from a probe project and they held fine on the bench, but mine were only carrying a fan, not a heater, so now I'm rethinking it. One thing I keep wondering about: if a thermistor wire gets flaky mid-print while the heater is on, does Klipper catch that fast enough before anything gets silly, or do people add thier own checks in the select macro? Total newcomer to the firmware side so I genuinely don't know how forgiving it is there.

Total beginner question here but how does the printer actualy know which little hotend it just picked up? Like if I mix up the order on the rail one day, does it just start heating the wrong one and melt everything? I'm guessing there's some kind of pin or magnet trick to identify each one but I have no idea how that works. Also curious if grabbing a hot tool off the dock messes with the little springy connector pins over time, since they'd be warm every time they slide out.