The system that watches all ten inputs at once — and tells you in time to act.
Growing has always been two things: paying attention, and deciding what to do. SCADA is the industrial name for doing both at a scale no person can reach alone — every room, every input, every minute, including two in the morning when the heater quits and you are asleep. SCADA stands for Supervisory Control and Data Acquisition. In plain terms: it senses what's happening, shows it to a human, raises an alarm the moment a number falls off a cliff, and — where you allow it — acts on its own. This site already calls that loop Collect → Have → Use. SCADA is that same loop built to industrial discipline, and a growing operation needs its own version of it.
Here is the catch that makes CEA different from anywhere else you'd put a SCADA system. The plant does not experience the room. It experiences the leaf and the root. The thermostat on the wall reads 25 °C; the leaf under your lights is closer to 29. A system that only watches room averages is watching an abstraction, not a reality. The whole point of this section is to specify what to measure, where to measure it, what the numbers mean, and what to do when they drift — so the system reports what the plant is living through, not just what the air in the middle of the room happens to be.
The three tiers, in plain language
Industrial SCADA is built in three layers, and they map cleanly onto the Collect → Have → Use frame this site already uses.
Tier 0–1 — the field. (Collect.) The sensors and local controllers out in the growing room: the pH probe in the reservoir, the CO₂ sensor at canopy height, the speed controller on the exhaust fan. These are the eyes and hands. A PLC — a programmable logic controller — or a small computer runs the local logic (turn the fan on, dose the acid, fire a local alarm) and keeps doing it even if everything above it goes dark.
Tier 2 — the supervisor. (Have / Use.) The screen where a human sets targets, watches trends, and receives alarms. The industrial word for that screen is the HMI, the human-machine interface. A setpoint is simply the target you give it — "hold VPD at 1.1 kPa" — and the system's job is to hold it and to shout when it can't.
Tier 3 — the memory. (Have.) Long-term storage and analysis — the historian — where every reading is kept, stamped with when and where, so you can lay this crop cycle against your best one and find the patterns no person has time to see. This is "Data is King" made operational.
What makes CEA SCADA its own discipline
A generic building-automation system was not built for a fertilizer-soaked, high-humidity room full of high-power lights. Three things bite here that don't bite in an office building:
- Electrical noise. High-frequency LED drivers and VFDs — variable-frequency drives, the speed controllers on fans and pumps — throw interference that can corrupt an analog sensor signal. Critical sensors want shielded cable or a digital connection.
- The boundary layer. Without air movement, the thin film of still air against a leaf lets the leaf's temperature, humidity, and CO₂ drift far from the room's. Your sensor reads one reality; the leaf lives another.
- Coupling. The ten inputs are not independent. Move one and others move with it. A SCADA built for CEA has to understand those linkages, or it will alarm on a symptom while the real cause goes unseen.
These are exactly the failure modes the rest of this section is built around.
Where this fits with what you already know
You met the ten inputs in Growing — the science of what each one does and why it matters. You'll find the generic how-to in the Fundamentals: how a sensor works, what MQTT is, how to power a node in the field, when to let the system decide and when a human has to. SCADA is the connective tissue between them. It takes the ten inputs from Growing and turns them into instruments, alarms, and a memory, using the building blocks from Fundamentals. It re-teaches neither side. It joins them.
If you want the hands-on, walk-up path, the Hort Assistant approach and the Home Assistant device library are where to start building. This section is the architecture that sits above them.
The rest of this section
- The instrumentation map — the ten inputs, one cell each: what to measure, which sensor, where to put it, the alarms that matter, and how to keep the reading honest.
- Alarm management — how to build an alarm system people actually trust, instead of one they learn to tune out.
- Cascade detection — how a single drift becomes a crop loss, and how to catch the chain early.
- The Plant's Eye View — the room reading and the estimated leaf reading side by side, so you can see when the room is lying to you.
- The historian — what to log, and how a season of clean data becomes next season's edge.
- The architecture — how a fail-safe, edge-to-cloud system is actually built, layer by layer, from commodity parts.
- Commissioning & maintenance — how to stand it up so you trust it, and how to keep it honest over time.