What Changes When Control Moves from Distributed to Centralized

At the beginning of the solar system project, control was distributed across three Arduino boards. Each of them handled a local task (solar collection, heat pumping, and cooling) with minimal coordination. This approach was simple and modular, but it shaped how the system behaved in subtle ways.

With distributed control, timing becomes implicit rather than guaranteed. Each controller runs its own loop, communicates asynchronously, and reacts to local measurements. While this works for isolated subsystems, it becomes limiting when loops are thermally coupled through a shared storage tank. Decisions made in one controller can unintentionally interfere with another.

Upgrading to a centralized STM32 controller changed this dynamic fundamentally. All sensors, actuators, and timing logic were brought into a single control core, making system behavior more explicit. Every control decision was made with access to global state information, and all actions were synchronized in time. The system shifted from loosely coordinated reactions to deliberate, system-level control.

Centralization also simplified reasoning about the system. Fault handling and safety constraints could be expressed coherently rather than duplicated across boards. Delay became measurable and predictable. Most importantly, the system became ready for higher-level control strategies that require coordination across loops.

This transition highlighted an important lesson: control performance is not only about algorithms, but about architecture. Moving from distributed to centralized control not only made the system faster and cleaner, but changed what kinds of control strategies were even possible.