Abstract: Power over Ethernet (PoE) is one of the key technologies for Power over Ethernet (EPA). This paper briefly introduces the IEEE 802.3af standard, systematically analyzes the functional requirements and overall design of PoE devices, selects the MSP430F148 microcontroller and MAX5945 PoE power manager, and develops a PoE device compliant with the IEEE 802.3af standard based on the I2-CBUS communication specification. An application example of this power supply device in an EPA system is also given. Keywords: IEEE 802.3af, PoE, Power over Ethernet device, MSP430F148, MAX5945, I2C-BUS, EPA. The MSP430F148 is a Flash-type microcontroller from the MSP430 series of ultra-low-power mixed-signal controllers launched by Texas Instruments (TI). It features a 16-bit RISC architecture, and the 16 registers and constant generator in the CPU enable the MSP430 microcontroller to achieve maximum code efficiency; a flexible clock source allows the device to achieve minimum power consumption; and a digitally controlled oscillator (DCO) allows the device to quickly wake up from low-power mode and activate to active operating mode within 6μs. Applying it to Power over Ethernet (PoE) devices allows for convenient control of the PoE power management chip and enables users to easily monitor PoE devices via terminal monitoring programs. 1. Introduction to the IEEE 802.3af Standard The IEEE 802.3af standard defines a method that allows the transmission of 48V DC power over Ethernet while simultaneously transmitting data. It introduces PoE technology into existing network infrastructure and is compatible with existing network equipment; it can provide a maximum power of 12.95W over a transmission distance of 100m. PoE consists of two parts: Power Sourcing Equipment (PSE) and Powered Device (PD). The PSE is responsible for injecting power into the Ethernet cable and implementing power planning and management. The IEEE 802.3af standard defines two types of PSEs: "Endpoint PSE" and "Midspan PSE". Endpoint Power Switches (PSEs) are PoE-enabled Ethernet switches, routers, hubs, or other network devices that transmit power over the signal or spare pairs of CAT5 cables. Midspan Power Switches (PSSs) are dedicated power management devices that do not perform data exchange. They typically work in conjunction with data switching devices to enable Power over Ethernet (PoE). The main steps of a PSE's operation are: ① Discovery: Before allowing the PSE to power a valid Device Output (PD), it must check the characteristic resistance using a limited-power test source. A two-point detection method is generally used for this. ② Classification: After detecting a valid PD, the PSE uses a probe voltage of 15.5–20.5 V to determine the PD's power level. The PD indicates its maximum power requirement to the PSE by drawing different constant currents from the line (classification characteristic signals). ③ Delivery: After successful detection and classification, the PSE powers the PD normally. During delivery, the PSE also monitors the power supply status of each port, providing undervoltage and overcurrent protection. ④ Power Off (shutdown). When the PD disconnects, the PSE stops supplying power to the line. The PSE can use either DC or AC circuit breaker detection to detect if the PD is disconnected. The PD is responsible for separating the 48V power supply and data signal in the network terminal equipment and transforming the 48V DC power supply to the 5V DC required for normal operation of the terminal equipment. When the PSE detects and classifies the PD, the PD should respond accordingly; simultaneously, during the PSE power supply process, the PD sends a continuous operating signal through the Maintain Power Signature (MPS). 2 Hardware Architecture and Composition In a PoE system, the PSE is the main component. In addition to implementing the power management functions mentioned above, in some special applications, the PSE must also be able to provide real-time operating parameters for each PD and monitor the entire system through a terminal monitoring program running on a PC. The PSE system consists of hardware and software parts. Figure 1 shows the hardware architecture diagram of the power supply system. The system mainly consists of a power supply module, a power conversion circuit, a MAX5945 microcontroller and its peripheral circuits, an MSP430F148 microcontroller and its peripheral circuits, and a CP2102 microcontroller and its peripheral circuits. The 16-bit MSP430F148 microcontroller reads and writes to the MAX5945 via I2CBUS to implement power management functions; it sets the operating mode of the MAX5945 through mode setting signal lines; and it receives error interrupt signals from the MAX5945 through error interrupt signal lines, thereby generating a valid low-level reset pulse signal for the MAX5945 through reset signal lines. Simultaneously, the MSP430F148 communicates with the terminal monitoring program on the PC via its internal UART module, bridged to a USB interface by the CP2102; it can also communicate with the PC via a serial port. The system provides intuitive monitoring and will automatically operate when it detects a lack of connection to the PC. The specific functions of each hardware component are described below. 2.1 Power Supply Section The power supply section primarily provides operating voltages for the various components in the system. The system requires three voltages: 48V, +5V, and +3.3V. The +5V required by the CP2102 is provided by the PC's USB interface, while other components are supplied with the +48V output from the power module or after conversion. Power Module: A 220V to -48V switching power supply module is used. Since one MAX5945 can manage power supply to four Ethernet ports, and multiple MAX5945s can be connected to I2CBUS, the power of the power module can be selected according to the actual situation. The MAX5945 only requires a separate external -48V power supply, which is provided by the power module after inversion by the conversion circuit. Power Conversion Circuit: The MSP430F148 operates at +3.3V. This design uses the LM2575HVS5.0 power conversion chip to convert +48V to +5V, and then uses the AMS11173.3 to convert +5V to the +3.3V required by the MSP430F148. 2.2 Ethernet Power Management Section The MAX5945 is a four-channel network power controller from Maxim Integrated for power supply devices (PSEs) compliant with IEEE 802.3af. This device provides device (PD) detection, grading, current limiting, and DC and AC load disconnection detection. The MAX5945 can be used in terminal PSEs (LAN switches/routers) or mid-span PSEs (power injection) systems. The MAX5945 can operate independently or be controlled by software via an I2C-compatible interface. Separate input and output data lines (SDAIN and SDAOUT) allow the use of optocouplers. An INT output and four shutdown inputs (SHD_) allow for rapid response from error to port shutdown. A RESET input allows for hardware reset of the device. The MAX5945 is entirely software-configurable and programmable. Graded overcurrent detection enables system power management to detect whether the current drawn by the PD exceeds the current allowed by its grade. The MAX5945 has four operating modes: automatic, semi-automatic, manual, and shutdown. In automatic mode, the MAX5945 automatically performs detection, classification, and power supply functions for standard PDs without microcontroller control. Therefore, in low-cost designs, the MAX5945 can be directly set to automatic mode (in this mode, the MAX5945 uses DC open-circuit detection to detect whether the PD is disconnected). In semi-automatic mode, the MAX5945 repeatedly performs row detection and/or classification as needed, without powering on the port regardless of its connection status. Power to the port must be turned off using software commands each time. In manual mode, superior AC open-circuit detection can be performed, obtaining the voltage and current of each PD in real time. This requires controlling the read/write registers inside the MAX5945 via the I2C bus, thus requiring a program running on the MSP430F148 microcontroller for advanced power supply management. Shutdown mode terminates all activities and safely shuts off the port power. Switching between automatic, semi-automatic, and manual modes has no effect until the device completes its current task. When a port is set to shutdown mode, the port immediately stops all operations and remains idle until exiting shutdown mode. Address setting circuit: The MAX5945 is a slave device, and its four address lines allow for the selection of 16 different I2C addresses. AC circuit breaker detection circuit: In PMM mode, an AC circuit breaker detection signal superimposed on the power supply circuit can be generated by setting the MAX5945's internal registers in conjunction with an external AC circuit breaker detection circuit. Status display circuit: The MAX5945 requires a detection and display circuit to be added to the power supply circuit of each port. This allows the MAX5945 to intuitively display the operating status of each port in all three operating modes. 2.3 Microcontroller Control Section The MSP430F148 is a Flash-type microcontroller in the MSP430 series of ultra-low-power mixed-signal controllers from TI, using a simplified instruction set to operate all functional modules. It features a 16-bit RISC architecture with on-chip 48 KB Flash, 2 KB RAM, USART, and other modules. The CPU's 16 registers and constant generator enable the MSP430 microcontroller to achieve maximum code efficiency. Different clock sources allow the MSP430F148 to meet various low-power requirements. A digitally controlled oscillator (DCO) allows the MSP430F148 to switch from low-power mode to active mode within 6μs. It supports in-circuit emulation, and its development tools provide excellent support for C language development, improving software development efficiency. The MSP430F148's safety fuses protect the program code. The MSP430 F148 microcontroller employs a memory-to-memory architecture, using a common memory space to address all functional modules and a reduced instruction set to operate on them. Its internal structure includes the CPU, memory, oscillator and clock generator, and peripheral modules. Clock circuit: The MSP430F148's clock module mainly consists of a high-speed clock, a low-speed clock, and a digitally controlled oscillator. The digitally controlled oscillator is integrated internally, while the 8 MHz high-speed clock and 32.768 kHz low-speed clock are generated by an external clock circuit. Reset circuit: A resistor-capacitor (RC) reset circuit is used to achieve external manual reset of the MSP430F148. JTAG interface: The MSP430F148 microcontroller has an embedded JTAG interface, supporting the boundary scan technology standard IEEE1149.1. It mainly consists of five control signals: TCK, TDO, TDI, TMS, and RST. Through the integrated IDE development environment, online code debugging can be easily performed. Buzzer: When the program detects an error in the MAX5945, a pulse signal of a certain frequency is sent to the buzzer through the MSP430F148's P3.3 port, thus providing an audible alarm. 2.4 USB Bridge CP2102 The CP2102 is a highly integrated dedicated communication chip. The chip's function is to convert data between UART and USB formats. It integrates a full-speed function controller conforming to the USB 2.0 standard, EEPROM, buffer, and UART data bus with modem interface signals, as well as an integrated internal clock and USB transceiver. The CP2102 can easily bridge UART to USB, thus adding a USB communication interface to the system. 3. Software Design and Implementation The PSE software implementation mainly consists of two parts: the PSE operation control program running on the MSP430F148 and the PSE terminal monitoring program running on the PC. They communicate through the USB interface constructed by the CP2102. 3.1 PSE Operation Control Program The PSE operation control program mainly implements system initialization, control of the MAX5945, communication with the PC, and data encapsulation and parsing functions. As shown in Figure 2, when not connected to the PC, the MAX5945 is set to work in AM mode, and the MAX5945 will run autonomously; at this time, the specific operating data of each power supply port cannot be obtained, and the operating status of each port can only be displayed through the LEDs in the status display circuit. When connected to a PC, the system will set the MAX5945 to the corresponding operating mode according to the user's requirements. At this time, the system can collect the operating parameters of each port. In SAM and PMM modes, the system can partially or completely control the power supply of each port according to the user's settings. The monitoring process is achieved through read and write operations on the registers of each port of the MAX5945. 3.1.1 System Initialization System Clock Initialization: Select 8 MHz clock XT2 as the master clock source and DCO as the slave clock source. I/O Port Initialization: Set P3.3 as an output, serving as the signal to drive the buzzer; set P4.0 as an output, serving as the mode selection signal for the MAX5945; set P4.2 as an output, serving as the reset signal for the MAX5945; set P4.1 as an input, serving as the error interrupt input signal for the MAX5945. Serial Port Initialization: The MSP430F148 communicates with the CP2102 via UART1. UART1 is configured as follows: 8 bits for transmitted characters; transmit/receive rate of 9600 bps; select auxiliary clock ACLK as the clock source for the baud rate generator; enable serial port receive and transmit operations; set the function selection registers P3.6 and P3.7 to serial port transmit/receive mode. 3.1.2 I2C-BUS Implementation The MSP430F148 does not have a standard I2C-BUS communication module. Therefore, SDA and SCL in the I2C-BUS communication specification need to be simulated in software using P3.0 and P3.2 to complete the I2C-BUS read and write operations. (1) I2C-BUS Write Operation The I2C-BUS write function "void WriteI2C(char Addr, char Reg, char Ctr)" consists of the formal parameters Addr (address of MAX5945), Reg (address of MAX5945 register), and Ctr (control information); the write function is composed of the sub-functions I2CInit(), I2CStart(), I2CSent(unsigned char data), I2CReceiveAck(), I2CReceiveAck(), I2CStop(), and delay(). The I2C-BUS write function is used to write control information to the specified MAX5945 internal register. The specific I2C-BUS write operation flow is shown in Figure 3. (2) I2C-BUS Read Operation The I2C-BUS read function "void ReadI2C(unsigned char Adr, unsigned char Rg)" consists of the address of the formal parameter Adr-MAX5945 and the address of the Rg-MAX5945 register. The result of this operation is to read the information in the Rg status register of the MAX5945 at address Adr and store it in a char global variable. The read function consists of the sub-functions I2CInit(), I2CStart(), I2CSent(), I2CReceiveAck(), I2CSent(unsigned char data), Rec_dat(), I2CSentNAck(), I2CReceiveAck(), I2CStop(), and delay(). These sub-functions work together to complete the I2C-BUS read timing. The specific I2C-BUS read operation flow is shown in Figure 4. 3.2 PSE Terminal Monitoring Program The PSE terminal monitoring program mainly implements the real-time monitoring function of each power supply port. Because of the use of the USB bridging chip CP2102, the monitoring program only needs to handle serial communication logically. Various control data are set through the terminal monitoring program, and the real-time operating parameters of each power supply port are also displayed intuitively on the monitoring program. The terminal monitoring program implements advanced power supply management functions. 4. Conclusion With the maturity and development of Power over Ethernet (PoE) technology, it will gradually be widely adopted. This design uses the MSP430F148 microcontroller and the MAX5945 PoE manager to develop a monitorable advanced PoE management system compliant with the IEEE 802.3af PoE standard. This system can also be simplified to reduce costs according to actual needs. This system has been applied in an EPA (Ethernet for Industrial Automation) system with good results.