new findings: forget what I thought before about the drivers all being equal.

When I disconnect the inputs and outputs from the non-used drivers (easy with plug in modules :) ) I find that I only get the malfunctions when using driver2, drivers 1 and 3 can't be made to fault by pressing the jogs on the other axes.

I will go run this simple test on the BOB X and Z outputs in a bit (if driver2 doesn't fault on the X or Z axis then the BOB is also involved in the problem).