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

[Voicecoil]

It's a cunning plan. There could be problems with dust and vibration for the drivers mounted on your gantry and don't forget that the power cables feeding the drivers, which will have to come from the control box alongside the control cables, will not have a smooth DC current flow, though you could add some filtering at each driver to reduce that.

Aye, I can see that dust etc. might be a problem and was thinking of a cover to minimise that.With a suitable cable (starquad or high current coaxial) I don't think the current flow in the power line would be too much of an issue.

The biggest problem I can see is that you are replacing the nice short signal cables between your controller and the drivers by several metres of 'antenna' wiring which will have to run alongside the still noisy stepper driver power cabling and the VERY noisy spindle power cabling from your VFD. It might be case of frying pans and fires in spite of your best efforts at screening. But please don't stop thinking out loud!

I was thinking of using a 3 x twisted pair screened cable running to the (Lichuan) drivers with have an iso-optalator input so are effectively floatting diff., so ought to be OK. But I'm not sure I can be arsed to undo all the "conventional" wiring I've already done

In my controller the break-out is mounted inside it's own screening box and has short screened cables running to the drivers leaving only about 20mm of exposed wiring. The output cables are deliberately routed away from the inputs to minimise coupling. I've cut the motor's own wires quite short and use screened cable to connect back to the drivers with all the cable screens connected only at the controller end. Limit switch inputs have 1uF capacitors across them to prevent noise-induced false triggering. I haven't had any problems since upgrading the controller to this design.

Kit

Caps across the inputs are cool, been using them for years with great results.