Showing 0 products

Frequently Asked Questions

When should an injector be used instead of a PoE switch?

For one or two devices, or where a device needs more power than the switch can give it - not as a way of avoiding buying the right switch.

Legitimate cases are genuinely common: a single camera at the end of a run from an existing non-PoE switch; an access point requiring a higher power class than the switch supports; a temporary installation; or a device beyond a switch that is fully committed on its power budget.

Where it becomes wrong is accumulation. Each injector is an unmonitored power supply in an awkward place, adds two connectors to the link, and cannot be power-cycled remotely. Once there are more than a few, a PoE switch is cheaper in total, tidier, and gives per-port monitoring and remote reboot - which is worth a great deal on cameras and access points.

A reasonable threshold is three. Beyond that, price the switch.

What distinguishes passive and standards-compliant injectors?

A standards-compliant injector negotiates before applying power; a passive one applies voltage regardless, and can damage a device that was not expecting it.

802.3af, at and bt injectors detect the connected device, determine its power class and only then energise the pairs. Connecting an ordinary computer to one is harmless, because detection fails and no power is applied.

A passive injector puts voltage on the cable continuously. If the device expects exactly that voltage on exactly those pairs it works; if it expects standards-based power, or is not a powered device at all, the result can be damage.

Passive injectors are still supplied with some wireless and specialist equipment, often at a non-standard voltage. They should be kept with the device they belong to, labelled, and never used as a general-purpose injector.

Where there is any doubt, specify standards-compliant. The small price difference is trivial against replacing a damaged device.

What do splitters do and when are they needed?

Take power off a PoE link and present it as a separate low-voltage supply, for devices with a network port and a power socket but no PoE support.

The use case is a device in a position with no mains socket: a small display, a sensor, a thin client, an older camera. A PoE switch powers the link, and a splitter at the far end separates the data and the power into two connections the device understands.

The critical detail is output voltage. Splitters are made for common voltages, and many are selectable. Getting it wrong damages the device, so confirm the requirement including the connector type and polarity.

Check the current as well as the voltage - a device may need more than the splitter can deliver, particularly at start-up when inrush exceeds steady-state draw.

The advantage over a local power supply is real: one cable, remote power cycling from the switch, and the switch's UPS covering the device.

Do injectors affect data speed?

A good one does not; a poor one can prevent gigabit or multi-gigabit operation entirely.

An injector inserts itself into the cable path, so the signal passes through its connectors and internal circuitry. A well-made unit is transparent to the data.

Cheap injectors, and particularly older ones, sometimes support only 100 Mb data even while supplying power. A device links at 100 Mb, everything works, and nobody notices for months - a classic silent constraint.

High-power injectors using all four pairs must handle the signal on those pairs properly, and poor implementations degrade it.

So check the data rate specification explicitly rather than assuming, especially where the device or the link is multi-gigabit.

The injector also consumes part of the 100 metre budget through its patch leads and its own connectors, which matters on a run that is already long - and a marginal link with an injector in it is a recognisable source of intermittent faults.

Where should an injector be placed?

Near the switch, in a cabinet, on protected power - not near the device.

The injector needs mains power, so putting it near the powered device means finding a socket at that position, which usually defeats the point of using PoE at all.

Placing it at the switch end keeps it in a cabinet, on the same protected power as the switch, accessible, and where somebody will find it. The single cable then runs to the device.

That placement also means the whole cable run carries power, which is what consumes some of the length budget - so on long runs, allow for voltage drop and consider a higher-power injector to compensate.

Label it clearly with the device it feeds, at both the injector and the patch panel. An unlabelled injector behind a cabinet is exactly the sort of thing that turns a simple fault into an afternoon.

Multi-port injectors are tidier where several are needed, though at that point a PoE switch usually wins.

Can PoE power be run through a splitter and then further?

Not usefully. Each PoE segment is limited to 100 metres, and chaining introduces losses and failure points that make it a poor design.

The standards define power from a source to a device over a single 100 metre segment. A splitter terminates that segment; anything beyond it is a separate link needing its own arrangement.

Devices exist that regenerate PoE - PoE extenders that take power and data in and pass both on, adding another 100 metres. These work and are legitimate for a specific problem, but each stage loses power and adds a failure point, and most limit the number of stages allowed.

Where the distance is genuinely long, the better answers are fibre with a local PoE injector at the far end, or a local power supply with an Ethernet extender.

Chaining splitters and injectors to reach a distant device produces something that works on the day and fails unpredictably later, usually at whichever stage is least accessible.

What faults are common with injectors and splitters?

Power supply failure, wrong voltage, insufficient power, and the device being forgotten in the documentation.

The external power supply is the most common failure, and it presents as a device that has simply gone offline with no other indication.

Wrong voltage on a splitter damages the device, usually immediately, and is a commissioning error rather than a service fault.

Insufficient power shows as a device that works normally until it needs more - a camera engaging its heater or infrared illuminator, an access point bringing up another radio - and then reboots. It is seasonal or intermittent, which delays diagnosis considerably.

And the documentation problem: an undocumented injector behind a cabinet is invisible to whoever investigates the fault, and the link tests clean end to end because the fault is power rather than data.

Label everything, record in-line devices on the cabling records, and keep spare units with their supplies.