I can't help myself. I saw a totally broken one for sale and bought it to try help solve the few remaining mysteries.
The brass part is the system for lowering a calibration weight.
This arrangement is starting to get familiar. It's a much less panic inducing way of constructing it though than it being machined from solid.
The fault with this balance was that this part had snapped. It's a double flexure to connect the voice coil lever beam to the main parallelogram (i'm making these names up).
This is the inside of the voice coil. There's a pair of fairly strong opposing magnets that go in the middle of it. The lever beam floats on a pair of reasonably stout flexures and is electrically connected by two extremely thin flat wires. The optical position sensor is just a slot cut in a metal bar in between a transmitter and reciever.
Measuring the resistance of the coil. I'm not sure if this gives me any really useful information but it seemed like a clever thing to do!
The circuit has a bunch of op amps on it, a precision voltage reference in the middle, what I assume to be some sort of thermocouple, and this unidentified blob. It was hidden under an aluminium can filled with rubber, and it's quite well isolated. For thermal stability I assume. Might try dissolving that resin off to see whats underneath.
I need to take a closer look at the main board before I can really take a guess at what does what, but i'm leaving that for now. For the record I have no idea what i'm doing when it comes to electronics and i'm mostly just googling the chip numbers.