Just an idea, I may be bonkers......

[Kitwn]

But I'm not sure I can be arsed to undo all the "conventional" wiring I've already done

If it ain't broke, don't fix it!

Kit

[Doddy]

The machine design is all about balancing compromise. From an electrical power drive to the motor what you propose (locating the drivers close to the motors) is a good idea - you're removing parasitic resistance and impedance from supply lines to the motor - which will impact the performance of the motor at speed. You're having to balance this with driving the digital step/dir signals over a longer length of cabling, and these are more susceptible to noise, but it's not rocket science to switch these (non-linear) resistive loads of a few volts / milliamps at 200kHz - differential drive is nice but not necessary unless you have long cable runs and a high noise environment (and you're removing a good proportion of the emissive lines by removing the high-current AC signalling to the steppers). I'd invest in a heavier gauge power cabling to the driver.

The other compromise is the mechanical aspects, and as others have said - vibration, dust etc. You are making life difficult for yourself here but this is just part of the overall design compromise.

[Neale]

On my machine, and I suspect many others, the motor cables aren't nearly long enough anyway, so need extending. This being the case, my motor cables (i.e. the ones actually coming out of the motor itself) terminate in a little plastic junction box somewhere near the motor and therefore quite a long way from the "sensitive" cables (step/dir, etc) which are retained within the screened control box and use differential signalling anyway. Junction box to control box is via screened CY cable. My guess is that the combination of screening and, effectively, differential motor connection work to significantly reduce emitted interference. The long "signal" wires on my machine (limit switches and e-stops) use 24V for better noise rejection. The drivers also stay away from vibration and dust in the control box with a light waft of air from a couple of low-power fans to keep them cool.

However, as Doddy says, engineering is all about compromise and there is a whole bunch of trade-offs to take into account. I went for the easy and most common configuration but that's not to say that in other cases, a different solution might not be better.