X-PoE Explained: How It Works, What's Different, and Why It Matters

Every other post in this series has referenced X-PoE — as the foundation that prevents automation decay, as the infrastructure that turns the ceiling into a network layer, as the architecture that scales to device-dense buildings. This post explains what X-PoE actually is, how the system works from power supply to light engine, and what makes it architecturally different from both traditional AC lighting and standard Power over Ethernet.

What X-PoE Is (and Isn't)

X-PoE is an extension of the IEEE 802.3 Power over Ethernet standard — the same standard that has powered IP phones, cameras, and access points for 25 years. It's not a proprietary protocol built from scratch. It's an evolution of proven technology, optimized for a specific use case: powering and controlling lighting loads.

Standard PoE (802.3bt) tops out at 90W per port. X-PoE delivers up to 120W per port — 40% more power — by using two individually regulated Class 2 circuits on a single Cat5e/Cat6 cable. Each cable contains 8 conductors split into two sets of 4, and X-PoE treats each set as an independent, individually controlled power channel.

That distinction matters: X-PoE isn't just "more watts." It's two circuits per cable, each with independent current regulation, dimming control, and power monitoring.

And critically, every X-PoE switch is backward-compatible with IEEE 802.3af/at/bt. Connect a standard PoE camera or access point, and the port auto-detects it and delivers IEEE-standard power. Connect an X-PoE lighting fixture, and the port delivers X-PoE power. Same port, same cable, different modes — automatically.

The Components: How the System Fits Together

An X-PoE system has four core components. Understanding how they connect is the key to understanding the architecture.

1. Power Supply

A standard AC-to-DC power supply converts building AC power to 48–57V DC. This is the only point where AC power enters the system. Everything downstream is low-voltage DC.

2. X-PoE Lighting Controller (Switch)

The XS-108H is the brain of the system. It takes the 48–57V DC input and distributes it across 8 auto-sensing ports, each with two independently controllable channels. Here's what lives inside the switch that doesn't live anywhere else in the system:

• LED driving circuitry — constant current (up to 1,150mA per channel) and constant voltage (24V–57V) regulation. This is the LED driver. It's not in the fixture. It's here.
• Per-channel power metering — over 95% accuracy, every channel, continuously
• Auto-detection — the switch identifies whether a connected device is IEEE PoE, standard X-PoE, or custom X-PoE, and configures power delivery accordingly
• Dimming control — down to 1%, per channel, via software
• Tunable white support — two channels per port means one channel can drive warm LEDs while the other drives cool, enabling human-centric lighting up to 2x65W
• Network interfaces — REST API and MQTT for integration with any IP-based controls system
• Failure detection — autonomous load detection flags anomalies at the port level

The switch ships in a 1U half-rack form factor — two units join to form a full 1U. It mounts in an electrical closet, an IDF, or on a wall. Not in the ceiling.

For larger deployments, the X-PoE Panels (XSP-2-1600 and XSP-2-2000) house two switches with an integrated power supply in a single enclosure — up to 3,200W of capacity, 16 ports, 32 channels.

3. Powered Devices (PDs)

PDs are the adapters that connect the Cat5e/Cat6 cable from the switch to the actual light fixture (or other device). They convert the RJ45 connection into one or two pairs of +/− conductors that power the LEDs directly.

PDs come in multiple configurations:

PDs can be housed inside the fixture, panel-mounted, or installed in a junction box. Some include adaptive signatures that optimize the detection handshake for faster, more efficient startup.

The critical point: PDs are passive adapters, not drivers. They don't regulate current. They don't dim. They don't fail the way LED drivers fail. The intelligence is in the switch.

4. LED Light Fixtures

X-PoE fixtures are driverless. They contain LED light engines — arrays of LEDs designed to operate at a specific forward voltage and current — but no driver electronics. The switch provides the precise current the LEDs need, and the PD delivers it.

This is the architectural inversion that changes everything:

continue
access** | Lift + ceiling access per fixture | Rack height in electrical closet | | Dimming method | 0–10V analog signal or DALI bus (separate wiring) | Digital, per-channel, over the same cable that delivers power | | Power monitoring | None at fixture level | Per-channel, >95% accuracy, continuous |

The Signal Flow: Power Supply to Photon

Here's what happens, end to end, when an X-PoE lighting system powers a fixture:

1. AC → DC conversion Building AC power enters a standard power supply and converts to 48–57V DC.

2. DC → Switch The DC power feeds into the XS-108H (or X-PoE Panel). The switch distributes it across 8 ports.

3. Switch → Auto-detection When a cable is connected to a port, the switch runs a detection sequence:

• First, it checks for an IEEE PoE device (standard handshake)
• If not IEEE, it checks for an X-PoE signature
• If a known X-PoE PD class is detected, the port configures automatically to the correct current output (e.g., 300mA, 500mA, 1,000mA)
• If a custom load is detected, the port defaults to 20% output until manually configured

4. Switch → Cable → PD Regulated DC power travels over Cat5e/Cat6 to the PD at the fixture. Two channels travel independently on the same cable — each on its own set of 4 conductors.

5. PD → LED engine The PD converts the RJ45 connection to +/− conductor pairs and delivers current directly to the LED light engine. No conversion. No regulation. Just delivery.

6. Photons The LEDs emit light at the brightness level set by the switch — anywhere from 1% to 100%, controlled digitally, adjustable in real time via the Luum platform, REST API, or MQTT.

The entire path from power supply to photon has one active intelligence point (the switch), one passive adapter (the PD), and one cable type (Cat5e/Cat6). Compare that to traditional lighting: panel → breaker → conduit → AC wire → junction box → LED driver → 0–10V control wire → fixture. Six or seven components, three cable types, two trades.

What Makes This Architecturally Different From Standard PoE Lighting

X-PoE isn't the first system to put lighting on Ethernet. Standard PoE lighting exists. But there's a critical architectural difference that's easy to miss:

Standard PoE lighting uses IEEE PoE switches to deliver power to fixtures that contain their own "PoE nodes" — small embedded computers that negotiate power, manage dimming, and communicate with the network. The intelligence is distributed into every fixture. Each fixture is essentially a networked computer with an LED attached.

X-PoE lighting centralizes the intelligence in the switch. The fixture is a passive LED engine. The PD is a passive adapter. The switch does the driving, the dimming, the monitoring, and the communication.

This is why X-PoE documentation describes the switch as a "lighting controller" rather than a network switch. It's both — but the lighting control function is what makes the architecture distinct.

Safety and Standards: The Class 2 Question

A common question: "Is 120W on a single cable safe under Class 2?"

Yes — because each X-PoE port is actually two individually regulated Class 2 circuits. The NEC Class 2 limit is 100W per circuit. Each Cat5e/Cat6 cable carries two sets of 4 conductors, and X-PoE treats each set as a separate circuit — each independently regulated, each under the 100W limit. Two Class 2 circuits on one cable, not one Class 4 circuit.

X-PoE switches are UL-certified Class 2 power supplies (UL 62368-1, UL 2108-1). Safety and reliability are covered by UL and NEC — the same bodies that certify every other electrical system in the building.

And for context: the IEEE standards that codify PoE power levels have historically been published 6–8 years after the technology was already in market. Cisco shipped 60W UPoE in 2011; IEEE 802.3bt didn't formalize it until 2018. X-PoE follows the same trajectory — certified, deployed, and operating in buildings today while the standards process catches up.

What X-PoE Powers Today

While lighting is the primary use case, the same infrastructure supports a growing range of endpoints:

One switch. One cable type. Multiple device categories. The infrastructure doesn't care whether the endpoint is a light, a camera, or a charging station — it detects, configures, and powers it accordingly.

Where the Switch Lives — And Why That Matters

The XS-108H mounts in an electrical closet, IDF room, or on a wall — not in the ceiling, not above a drop tile, not behind a fixture. It's a 1U half-rack form factor, accessible at rack height.

This isn't a minor installation detail. It's an architectural decision with cascading consequences:

• Maintenance happens where people can reach it — no lifts, no ceiling disruption, no tenant impact
• Thermal management improves — the switch dissipates 200 BTU/h in a ventilated closet instead of a sealed plenum space
• Upgrades and replacements are centralized — swap a switch and 16 channels are updated at once
• Emergency override is physically accessible — a dry contact closure on the switch enables emergency operation during communication loss

The closet becomes the building's lighting infrastructure hub — the same way a server room is the IT infrastructure hub. Centralized, accessible, maintainable.

TL;DR

X-PoE is a low-voltage power architecture that centralizes LED driving in the switch and delivers precisely regulated power over Cat5e/Cat6 to driverless fixtures.