Ken,
If I'm not mistaken there are two coils inside this switch machine (DZ-1000), which in this instance is not being used to swing switch points, but instead to direct the movement of the contacts in the relay module (DZ-1008) attached to it.
There are two states, one to power the rail path through the crossover in one direction, one for the opposite. When it's working properly you set one position by activating one of the coils briefly, by rolling a set of wheels over an insulated rail section on one rail path. Alternatively you set the other by briefly activating the other coil by rolling through the opposite rail path. Although both can be activated at the same time (cars on both tracks approaching the diamond) when you remove the activations (wheel sets at Z and Y), both coils will become unpowered and will settle in one of the two states based on whichever coil loses its power last.
It appears here however that one of the two coils is stuck on, probably due to on unintentional wiring issue that you introduced when you made the swap. Normally the coils are activated only for a short period of time in order to swing the switch machine, and the relay attached to it, to the opposite position from where it was before the activation.
If it swings back by itself then there's would likely be a coil energized continuously, pulling it back. This should be generating some heat if it's stuck on. Is the switch machine hot to the touch while this is happening?
Do you have a means to measure voltage, for example a voltmeter? If so are either of the voltages, from Z to common or Y to common, zero, or close to it? Neither should be around zero when everything is just sitting there, i.e. powered up but with no rolling stock on the rails.
Mike