A PLC was built to be fast and discrete: read inputs, solve logic, write outputs, in a few milliseconds, forever. A DCS was built to supervise a continuous process across hundreds of loops with an operator watching it, and to be engineered as one database rather than as a set of programs. Those origins still show in how each one is configured, even though a modern PLC will run a PID loop perfectly well and a modern DCS will handle discrete logic.A PLC was built to be fast and discrete: read inputs, solve logic, write outputs, in a few milliseconds, forever. A DCS was built to supervise a continuous process across hundreds of loops with an operator watching it, and to be engineered as one database rather than as a set of programs. Those origins still show in how each one is configured, even though a modern PLC will run a PID loop perfectly well and a modern DCS will handle discrete logic.

Machine control, packaging, high-speed discrete manufacturing, and anything where the sequence is the product. Lower cost per point. A far larger pool of engineers who can maintain it, which matters more than any technical argument on this list. And you can buy one controller and expand later without re-licensing the plant.Machine control, packaging, high-speed discrete manufacturing, and anything where the sequence is the product. Lower cost per point. A far larger pool of engineers who can maintain it, which matters more than any technical argument on this list. And you can buy one controller and expand later without re-licensing the plant.

Large continuous process, hundreds of analogue loops, batch with regulatory traceability, and plants where the operator interface has to be consistent across every unit because the operator moves between them. The single engineering database is the real advantage: change a tag once and every screen, alarm and historian entry follows. On a large plant that is worth a great deal.Large continuous process, hundreds of analogue loops, batch with regulatory traceability, and plants where the operator interface has to be consistent across every unit because the operator moves between them. The single engineering database is the real advantage: change a tag once and every screen, alarm and historian entry follows. On a large plant that is worth a great deal.

Who fixes it at three in the morning in year seven. If that is your own maintenance team, a PLC and SCADA they already understand will outperform a technically superior DCS they need a contractor to touch. If it is a permanent process engineering group who live in the system, the DCS pays back.Who fixes it at three in the morning in year seven. If that is your own maintenance team, a PLC and SCADA they already understand will outperform a technically superior DCS they need a contractor to touch. If it is a permanent process engineering group who live in the system, the DCS pays back.

Most of the plants we commission are neither pure case. PLC and SCADA on the packaging hall, DCS on the process side, one historian, and a clear boundary with an agreed handshake at it. The mistake is not choosing the hybrid. It is leaving the boundary undefined until commissioning, which is where undefined boundaries are discovered.Most of the plants we commission are neither pure case. PLC and SCADA on the packaging hall, DCS on the process side, one historian, and a clear boundary with an agreed handshake at it. The mistake is not choosing the hybrid. It is leaving the boundary undefined until commissioning, which is where undefined boundaries are discovered.