What is PSE in networking?

# What is PSE in networking?

> Meta Description: PSE stands for Power Sourcing Equipment—the device that injects power over Ethernet. Learn how PSEs work, the standards they follow, and what B2B buyers must verify.

> Target Keyword: PSE in networking

You are reading a PoE Switch datasheet and a column catches your eye: "PSE power budget: 240W." Or you are troubleshooting a camera that will not power up, and someone on the support call asks, "What PSE are you using?" The term appears everywhere in PoE documentation, yet many buyers treat it as a synonym for "switch." It is not. PSE is a functional role in the Power over Ethernet ecosystem, and confusing it with the broader network hardware leads to mismatched purchases and failed deployments.

PSE stands for Power Sourcing Equipment. In any PoE link, two roles exist: the PSE supplies power, and the PD (Powered Device) consumes it. The IEEE 802.3 standards define exactly how they negotiate, and that handshake determines whether your device powers on, how much wattage it receives, and whether the link is safe. Here is what PSE actually means in networking, and what to check before you buy.

## PSE is a role, not a product category

A PSE is any device that injects DC power onto an Ethernet cable while passing data through. That definition covers two physical forms. An endspan PSE is a PoE Switch—the power sourcing circuitry is built into each port. A midspan PSE is a standalone injector placed between a non-PoE Switch and the powered device. Both perform the same electrical function. The difference is packaging, port count, and management capability.

This is why you cannot simply ask a supplier for "a PSE" without specifying form factor. If you need to power eight cameras in one rack, an endspan switch is tidier and more manageable. If you need to power one rooftop camera from an existing non-PoE Switch, a midspan injector is cheaper and faster to deploy. Same role, different hardware.

The term is also standardised across vendors. When Cisco, HPE, or Ubiquiti documents "PSE detection" or "PSE priority," they refer to the same IEEE-defined behaviour. That consistency is what makes multi-vendor PoE deployments possible, provided the standards match.

## The IEEE standards that define a PSE

A PSE's capabilities are governed by which IEEE amendment it supports. The original 802.3af (2003) defines a PSE output of up to 15.4W per port at the source, with roughly 12.95W guaranteed at the device after cable loss. The 802.3at amendment, commonly called PoE+, raised that to 30W at the source and 25.5W at the device. The 802.3bt amendment introduced two higher tiers: Type 3 at 60W source and Type 4 at 90W source, with 51W and 71.3W respectively available to the device.

Each tier also defines power classes, numbered 0 through 8. During negotiation, the PSE applies a classification voltage and measures the current drawn by the PD to determine which class it belongs to. The PSE then reserves that amount of power from its total budget. A switch with a 240W budget and four Type 4 ports at 90W each would be oversubscribed—it cannot supply all four at full power simultaneously. This is why the total PSE power budget matters as much as the per-port rating.

Pinout also falls under the standard. Mode A delivers power over the data pairs, Mode B over the spare pairs. Most modern PSEs support both and negotiate automatically. Legacy devices, however, may accept only one mode. That mismatch is a frequent cause of "it should work but it doesn't" in the field.

## The detection handshake is the PSE's core job

Before a PSE applies full voltage, it must confirm that a valid PD is connected. The process has three stages. First, detection: the PSE sends a low, current-limited voltage and looks for a 25 kΩ signature resistor across the line. If it does not find one, it assumes a non-PoE device and leaves the port unpowered. Second, classification: the PSE raises the voltage slightly and measures the PD's current draw to identify its power class. Third, power-on: the PSE applies the negotiated voltage, typically 48V, and continuously monitors for overcurrent, short circuit, or disconnection.

This handshake is what makes PoE safe. It is also what separates an IEEE-compliant PSE from a passive injector. A passive injector skips detection and classification entirely, forcing 48V onto the line regardless of what is connected. Plug a laptop into a passive injector and you can destroy the Ethernet port. When sourcing, always confirm that the PSE performs IEEE-compliant detection, not just "48V output."

The monitoring continues after power-on. If the PD draws more current than its class allows, the PSE shuts the port down. If the cable is unplugged, the PSE detects the loss of signature and removes power. This continuous supervision protects both ends of the link.

## Where PSE decisions shape B2B deployments

The most common specification error we see is buying a PSE based on port count rather than power budget. A 24-port PoE Switch sounds generous until you realise its total budget is 370W and your 24 cameras each need 30W—that is 720W of demand on a 370W supply. The switch will power up the first twelve devices and refuse the rest, or shed lower-priority ports depending on configuration. Always sum the device wattages and add at least 20% headroom before choosing the PSE.

Thermal design follows from the same math. A PSE converting 720W of demand dissipates a meaningful amount of heat, even at 90% efficiency. Midspan injectors dissipate heat in a smaller chassis, which is why high-power single-port units often include metal housings and ventilation slots. In sealed outdoor cabinets, that heat has nowhere to go. Oversizing the PSE and choosing units with adequate thermal derating curves is the practical fix.

Management features also vary by PSE type. Endspan switches typically offer per-port priority, scheduling, and SNMP monitoring. Midspan injectors usually offer none of these—they are transparent, unmanaged devices. If your project requires remote port cycling or power consumption reporting, the endspan route is the only option. If it does not, the simpler midspan saves money.

Distance is the final constraint. A PSE cannot extend Ethernet beyond 100 meters. Power and data share the cable, and both degrade past that limit. If your device sits further away, you need a midspan extender or a PoE Splitter at the 90-meter mark, not a more powerful PSE.

## What to verify before placing an order

Start with the standard. Does the PSE support the PoE tier your devices require—af, at, or bt? Then the power budget: per-port maximum and total. Then the number of ports that can operate at maximum simultaneously.

Check the detection behaviour. Confirm IEEE-compliant detection and classification, not passive injection. For high-power devices, verify the PSE supports the correct class and pinout mode.

Request thermal data. For multi-port or high-power units, ask for the derating curve: how much power the PSE can deliver at 50°C ambient versus 25°C. Many budget units quote full power only at room temperature.

Finally, ask for compatibility evidence. As a trading partner serving global B2B clients, we test PSEs against mainstream PDs—cameras, access points, industrial sensors—before recommending them for volume. The most common issue is not the power rating but detection timing: some PSEs probe faster than certain PDs can respond, and the port stays off. Sample kits with a documented compatibility matrix let your team validate the exact pairing before committing to a container order.

PSE is not a product you buy off a shelf. It is a role defined by standards, power budget, and handshake behaviour. Get those three right, and PoE works the way the datasheet promised. Get them wrong, and you will spend weeks chasing a port that never turns on.

If you are sourcing PoE Switches or injectors for a multi-site project, send us your device list with wattage and PoE class. We will return a PSE recommendation with the right power budget, sample units, and a compatibility matrix for your lab testing.