If the Fault LED on the controller goes solid red, all efforts to configure communications over any port or protocol are futile.
The fact that the controller shows a solid red fault light under any circumstances, as well as the fact that the controller previously lost the user program when the power was cycled, both indicate a faulty controller.
It could be a tiny wisp of conductive material across a processor pin, it could be a colony of Madagascar Silicon Mites breeding between the memory chips, it could be a deep strike to the backplane ASIC by a high-speed muon burst from the hadron supercollider.
I'm willing to be shown to be incorrect; in most cases I've seen the solid red fault light is a permanent occurence, not a conditional one.
You can continue to diagnose using PING, ARP, and HTTP on the Ethernet port, and carefully holding the user's hand about COM ports and DF1 drivers.
But odds are good that the controller is damaged and must be replaced.