It seems I've done a bad job of electrical design. The machine runs a long bundle of cables consisting of power cables for the servo motors, shielded cable for the encoders and unshielded cables for the limit and home switches. All this leads back to a cabinet made from MDF with the internals from a mig welder as the power supply for the servo motors. Next to the power supply is the capacitors for the power supply (cant recall the specs but the are three of them and they are about 2 inches diameter and 5 inches high). That sits next to the PC to drive the whole thing and the servo drives are mounted above the capacitors.
Interestingly, I've not been running the spindle when these issues occur, so I could quite easily have further issues when I get to running with the spindle powered up. The last time I got a false signal, the machine was sitting idle. I assume this would mean the noise is coming, in part, from the power supply. The power supply sits in the original steel case (but without the side panels on it), so I assume as a minimum, a steel panel should be installed to separate the higher voltage equipment from the lower voltage equipment.
Thanks for the advice, I wish I knew all this at the design stage.