9. Powering the system

As noted earlier, the REDAC system imposes minimal requirements on its power supply. Any single-phase safety plug operating on 120V/240V AC mains with a maximum current of 16A is sufficient for proper operation.

The REDAC-specific (mostly analog) components are powered with high quality 24V DC power supplies with a maximum rating of 300W each.

Important

A correctly grounded mains connection is crucial for safe and reliable operation. Most data centers, working offices, or professional laboratories already provide suitable power infrastructure. Ensure that the grounding is verified to avoid potential safety risks and to guarantee system stability.

While the use of a data center power delivery unit (PDU) or an uninterruptible power supply (UPS) is not mandatory, these devices can offer significant benefits. A UPS, particularly one with a battery-backed design, provides additional protection against power surges, outages, and voltage fluctuations. Moreover, a UPS can stabilize the input voltage, which may enhance the performance of the internal power supply. This stability can, in turn, improve the accuracy and reliability of REDAC’s computing operations, making it a worthwhile consideration for critical applications.

Fig. 9.1 240V power distribution in REDAC

9.1. System Startup

REDAC features a 19-inch switchable power socket located at the rear of the rack (see Fig. 9.1). This design allows the entire system to be powered on with a single flip of the master switch, provided the mains connection has been properly plugged in and secured.

Once power is supplied to the system, the following sequence typically occurs automatically (ordered by their time to readiness):

1. The Ethernet networking equipment (router or switch) powers on immediately, establishing the network connections required for system communication.

3. The analog power supply activates, which subsequently powers all connected analog modules and prepares them for operation. This also means that all Microcontrollers power up and wait for an IP address with DHCP calls until the routing switch assigns them an IP address.

2. The server, also known as the supercontroller, powers on, waits for an IP address and starts up all relevant services in correct order. They then begin to initialize their control and managament processes.

Although each of these components can be powered on or off individually if necessary, this is rarely required under normal operating conditions. The automated startup process ensures that the system reaches an operational state efficiently and with minimal user intervention. However, users should verify that the system initializes correctly and observe for any unexpected behavior during startup, especially after relocation or maintenance.

Note

The REDAC system operated by anabrid (see Anabrid Operations) can be monitored at https://status.anabrid.net/status/redac1 in order to see when all systems are up and running after powerup.

9.2. System Shutdown

When shutting down the REDAC system, follow these guidelines to ensure a safe and proper shutdown process:

1. The Ethernet networking equipment (router or switch) can be powered off without any adverse effects on the system. Since the networking components do not store critical data, there is no risk of data loss when they are turned off.

2. The server (supercontroller) requires a manual shutdown sequence to prevent data loss. This is typically done by executing the command sudo shutdown -h now at the Linux command line. This ensures that all running processes are terminated gracefully, and any data in memory is safely written to storage before the system powers down.

3. The analog components of the system can be powered off without any significant effects. These parts do not rely on active processes or stored data and can be safely disconnected from power as needed.

Note

While a sudden power loss is generally not critical for the REDAC system, it is worth noting that the server is equipped with a journaling file system. This feature ensures data integrity and facilitates recovery mechanisms even in the event of an abrupt shutdown. In such cases, the only potential loss would involve volatile database entries, which are typically of minor importance in practical applications. Despite this resilience, a controlled shutdown process is always recommended to avoid unnecessary recovery procedures and ensure smooth system operation.