Power over Ethernet (PoE) enables power transmission through Ethernet connections. In PoE networks, the power sourcing equipment (PSE) supplies power, while the Ethernet network generates an output voltage ranging from 44 to 57 volts. At the receiving end, the powered device (PD) consumes the power. Although work is ongoing to define a higher power Ethernet standard, the current power available to the PD is limited to approximately 13 watts per single Ethernet connection. This limitation is often insufficient for complex applications, leading some high-power PDs to require power from multiple ports, converting it to usable voltage and isolating it from the 48-volt input voltage. Several techniques exist today to provide isolated power conversion from multiple inputs.
PoE has become a widely adopted concept, finding application in products such as VoIP phones, security surveillance systems, and point-of-sale terminals.
One common technique employed in DC/DC parallel power supplies is the so-called “descent method.†By reducing the output voltage and increasing the load current, parallel power supplies can evenly distribute current. This approach eliminates the need for communication between power supplies, avoids a single point of failure, and requires minimal additional components. When current mode control is utilized, limiting the DC gain of the control loop can create a proportional drop in output voltage relative to changes in load current.
Unfortunately, the descent method isn’t highly precise. Without load, the output voltage is typically regulated by the highest-powered supply. If diode-regulated, the lowest-powered supply won’t contribute until the voltage drops to around 5.25 volts. Under worst-case tolerances, the first supply might provide up to 70% of the output power before the lowest-voltage source starts contributing. While this may suffice under certain conditions, it’s not ideal due to its lack of reliability. As the load increases further, the first supply could reach its limit, leaving the remaining supplies to handle the increased current for full power operation.
A power supply architecture with synchronous rectification allows for bidirectional current flow, which can complicate this control method. In extreme cases, a single supply might attempt to regulate both high and low currents simultaneously. If this occurs without a load, some supplies might supply current while others draw it, transferring power inefficiently. Thus, it's advisable to disable synchronous rectification at zero amps.
Another technique for balancing multiple input powers is the interleaving method. Like the descent method, interleaving uses different power levels for each input, delivering them to a common output. However, interleaved power levels (or phases) share a common primary-side controller, reducing costs. Each phase can also operate out of phase with one another, synchronizing their timing. This synchronization minimizes output capacitor ripple currents, allowing for smaller filters. In interleaving, all inputs must share the same feedback loop, making it unsuitable for some applications.
Many PWM controllers are specifically designed for interleaving. For two-phase setups, a push-pull controller can perform interleaving, significantly cutting costs. Figure 1 illustrates a two-phase interleaved flyback power supply using a push-pull controller like the UCC2808. Each phase is limited to a 50% duty cycle and operates 180 degrees out of phase. Peak current mode control ensures both phases maintain similar peak current values. In discontinuous flyback, the output power of each phase is proportional to the square of the primary peak current, naturally balancing power distribution between inputs. The primary MOSFET switching delay is the main cause of imbalance, particularly when input voltages differ. The controller’s peak current limit constrains the maximum power drawn from inputs, while the load cycle clamp manages input current under undervoltage and fault conditions.
A third method to share power across multiple inputs involves a secondary-side load-sharing controller. Independent power supplies with remote sensing capabilities can share the same output. Load-sharing chips are often integrated into power modules, as shown in Figure 2. A shunt resistor measures the current supplied by each converter. Due to tolerances and parasitic impedances, one supply will typically supply more current. This becomes the master supply, setting the voltage on the load-sharing (LS) bus. Slave units use this bus voltage as a reference to control their output current. To adjust a slave unit, a voltage can be injected into its remote sense conductor, enabling the main supply to control the load voltage for optimal load regulation. Master-slave configurations offer excellent current-sharing accuracy, typically better than 3% at full load.
[Figure 1: Push-pull controller driving interleaved flyback]
[Figure 2: UCC39002 load-sharing controller enabling parallel independent power supplies]
Despite advancements, engineers continue exploring ways to optimize power distribution in PoE applications, ensuring efficiency and reliability across diverse scenarios.
Double Pole Rocker Switches,2 Pole Rocker Switch,Dual Pole Rocker Switch
Yang Guang Auli Electronic Appliances Co., Ltd. , https://www.ygpowerstrips.com