The M580 is an ePAC β Ethernet runs right through its backplane, so the CPU talks to its I/O over the same network that spans the plant. Local rack, RIO drops on a deterministic ring, and DIO devices for everything else. Build the network, watch the cyclic data flow, and pull a drop to see the diagnostics fire.
The M580 CPU sits in the local rack. RIO (Remote I/O) drops hang on a redundant Ethernet ring for deterministic, time-critical I/O; DIO (Distributed I/O) devices connect for less time-critical data. Cyclic frames sweep the ring every scan. Press β Pull RIO drop 2 and watch the ring report exactly which drop dropped.
Break a single ring cable and notice nothing stops β the ring heals around the break (that's why RIO uses a ring, not a line). Break a second cable, or pull a drop entirely, and then you lose that drop's I/O and the diagnostics light up.
The CPU and any I/O in its own rack β the fastest, addressed topologically (Track 02). Where the brain and its closest senses live.
Remote racks on a deterministic, self-healing Ethernet ring. Scanned in sync with the CPU each cycle β for time-critical distributed I/O.
Distributed devices (drives, islands, third-party gear) on standard Ethernet. Flexible, not ring-deterministic β for the rest of the plant.
All three ride the same Ethernet β that's the ePAC idea. The M580 also speaks Modbus TCP and EtherNet/IP over it, so one network carries control, I/O and supervision together.
Your RIO drops are wired in a ring rather than a daisy-chain line. A maintenance tech accidentally cuts one ring cable. What happens?