John, I've looked at the link to that board and I can't work it out - but it looks probable (PCB underside image) that the 0V reference for the 0-10V analogue output is common with the other 0V references on the board. In that case - I understand your concern with effectively raising the 0V reference to some hundreds of volts that the Spindle Controller operates. If that 0V cannot be isolated (between the logic and the analogue output) then, yes, you have a problem. I understand your argument of adding an isolation transformer to allow the 0V rail to be raised to that of the spindle-controller, but - of course - at that point you're raising the common 0v reference, back to the logic supply and ergo the short-circuit hazard remains at the point that you connect to another system - the connecting PC. It crossed my mind whether you could do the opposite and provide an isolation transformer (1:1) to the supply to the spindle controller, but you would wonder if there's any internal bonding of earths to line neutral anywhere on the spindle controller - either direct or indirect through EMF caps. Considerable care is needed going in this direction.

These https://www.amazon.co.uk/Isolation-Q.../dp/B082NXR4PR look interesting and useful. I've no experience of them, but understand the description rather better than the BOB.