Single-Port PSEs Put Power in More Places

Mar 23, 2007 4:06 PM
By Dilian Reyes, Applications Engineer, Linear Technology, Milpitas, Calif.

A highly integrated single-port controller simpli fies PSE design and enables novel implementations such as a wall-plug-powered PSE or a "power forwarder" that powers more than one device off a single port.

Click here for the enhanced PDF version of this article including diagrams and/or equations.

The IEEE 802.3af Power over Ethernet (PoE) standard specifies the delivery of a nominal 48 V at up to 350 mA over Cat 3 or higher cabling. The equipment that transmits the 48-V power is described as power-sourcing equipment (PSE), and the network devices that receive this dc power are referred to as powered devices (PDs). PSE controller ICs based on the standard perform PD detection, classification, port turn on, fault monitoring and power disconnect.

These chips are typically multichannel controllers, most commonly quad controllers, and are conventionally designed into large 12-, 24- or 48-port systems, requiring more than one IC. Some controllers run autonomously, but most applications use a microcontroller to step through the IEEE functions. A microcontroller is also used to read back from the PSE controller the port’s status and PD class. This information is then interpreted by the microcontroller for operations such as user interfacing and power management.

Multichannel controllers are suitable for large multiple-port equipment, either midspans that inject power into systems with legacy switches or, more frequently, endspan equipment such as switches, routers or hubs with the PSE function built in. However, there are many applications that only require a single port or a just a few ports with PoE capability. For example, in environments where PoE is not available, a single-port PSE powered from the ac outlet would allow users to power PDs. And in applications where a PD requires less than the allotted 12.95 W per port, a single-port PSE can serve as a “power forwarder,” taking the unused power and processing it to power a second PD.

PoE Power Classification

Before exploring the single-port PSE applications, it’s necessary to review the IEEE 802.3af specification as it relates to power classification. A PSE port is required to output 15.4 W for a Class 3 PD (Table), so providing full power to each port requires a large supply (15.4 W 3 24 ports = 370 W). To address this potential problem (i.e., the need for oversizing the power supply), the IEEE 802.3af standard gives the option for a PD to provide the PSE with power class information based on the maximum power the PD will ever consume. This allows a reduction in the size of the power supply with the use of a microcontroller performing power management.

Fig. 1 shows an example of a basic power-management operation. After setting the initial available power and detecting a PD, the power class is read by the PSE and then compared to the available power. If there is enough power for the PD, the port is turned on and power is subtracted from the available power. When the PD is disconnected, the port is turned off and power is added back to the available power.

By keeping track of the power allocated at each port according to the PD’s power class, the power remaining determines if there is enough left to power up additional PDs. If a fault occurs at a port or an insufficient amount of power is left, preventing the PD power on, this status can be read back through a user interface if one is available.