Wiring at the 3ways, main breaker panel, grounds, neutrals are are tight and check out.
When I removed the ZL7431 the “ghost” voltage on the “open” traveler leg of the 3ways also vanished. The “extra” voltage was not coming from the AC/DC power supply/transformer, it looks like it is coming from the ZL7431.
It is now clear and the Smartest House somewhat confirms that the wiring withing ZL7431 is a little more complicated than I might have thought. While it does have a simple AC relay to drive the “load” it is also clearly using a low voltage signal to sense the load and more important any change in wall switch status and that voltage is on the same circuit as the relay coil within the ZL7431
In my case, and I confirmed probing the wall switches, with the wall switch off (open) and the relay off (open) the ZL7431 sends a low voltage signal (around 8v) back to the “open” wall switch. On long runs where inductive voltage is possible it only take a couple of volts for the ZL7431 relay to sense a closed switch and confuse the device enough to turn on (and hold on) the lights.
Given that a normal “signal” from the wall switch is “loud” (110-120V) you would think that vision would build a little larger window into their trigger voltage… I’m still wondering if this unit (which had different pigtail wires than the other three) might have been “old stock”. Others on this forum have posted this problem and I thought I had read the Vision had addressed it.
I’m going to back probe my other Vision units to see if they have the same low voltage signal.