5 Switch Problems Killing Your CCTV Projects — And How to Fix Them

Your customer's cameras freeze. Cameras go offline. The whole site goes down and nobody can say why. In most of these cases the cameras are fine — the switch underneath them is the problem. Here are the five faults we see most often on surveillance sites, what each one looks like from the customer's side, and the specification that prevents it.

Problem 1 — Bandwidth bottleneck

Cameras look fine individually and stutter together. Frames drop when several people watch at once, and playback crawls. The cause is usually an uplink sized by port count instead of by bitrate: an eight-camera cabinet on a 100 Mbps path, or a trunk shared with office traffic.

Add up camera bitrates (a 4MP H.265 camera is typically 4–6 Mbps) and compare with the uplink, not with the port count.
Keep gigabit uplinks on every cabinet, and move to 2.5G or 10G when several cabinets converge on one core port.
Separate camera traffic from office traffic so a backup job cannot starve the recorders.

Problem 2 — PoE budget shortfall

Cameras reboot at dusk, or two or three drop off the network every evening. That is the signature of a PoE budget that was sized for daytime draw: as IR turns on, the total climbs past what the switch can deliver and the weakest ports get cut.

Compare total PoE budget, not "PoE on all ports"; a 16-port switch can share as little as 120 W.
Use night draw for every camera, then add 25–30% headroom.
Keep runs under 100 m and avoid thin CCA cable; voltage drop wastes budget before the camera sees it.

Our PoE switch range is specified by total budget and per-port class, so the number on the datasheet is the number the site gets.

Problem 3 — Network loops

Someone patches a wall port into another wall port, broadcast traffic multiplies, and the entire site stops — cameras, POS, door access, everything. An unmanaged switch has no defence against this.

Use managed switches with STP/RSTP and storm control on any site with more than one cabinet.
Enable loop detection on access ports used by staff or cleaners.
Label and document the patch panel; most loops come from undocumented ports.

Problem 4 — No remote management

One camera freezes on a Friday night and the only fix is a two-hour drive to power-cycle a port. Without management, every minor fault becomes a paid callout, and the customer remembers the invoice, not the cause.

Per-port PoE control lets you reboot a single camera from a desk.
Port statistics and error counters turn "the cameras are slow" into a measurable diagnosis.
SNMP or syslog lets the site report itself before the customer calls.

Managed and managed PoE models are listed under L2/3 managed switches and managed PoE switches.

Problem 5 — Wrong cable, connectors and installation

The switch is right and the site still fails, because the physical layer was treated as an afterthought: CCA patch leads, RJ45 plugs crimped on site, cameras at 130 m, or a cabinet with no surge protection and no earth reference.

Use Cat5e minimum, Cat6 for new runs; never copper-clad aluminium for PoE.
Buy factory-made patch leads, and test every run before the cameras go up.
Plan the 100 m limit before the cable is pulled; beyond it, use fibre or an extender.
Earth and surge-protect outdoor cabinets, especially where cameras sit on masts.

The five problems in one table

Symptom on site	Real cause	Switch specification that fixes it
Stutter and dropped frames when several users watch	Uplink sized by ports, not bitrate	Gigabit uplink minimum, 2.5G/10G at the core
Cameras reboot at dusk or drop off at night	PoE budget below night draw	Total budget with 25–30% headroom, per-port class checked
Whole site down after a patch change	Loop, no protection	Managed switch with STP/RSTP and storm control
Repeated callouts for single frozen cameras	No remote control	Per-port PoE control, SNMP, remote UI
Intermittent link, PoE faults at distance	Thin or CCA cable, unplanned long runs	Cat6 or fibre, tested runs, converters beyond 100 m

What we recommend

Specify the switch last, after the camera list, the cabinet layout and the cable plan — then check four numbers: total PoE budget, uplink bandwidth, management features, and spare ports. If all four are covered, the five problems above mostly stop happening.

Send the camera list and cabinet plan and we will return a port / PoE / uplink table with the matching models — usually within one working day. Get a quote.

FAQ

Why do my cameras reboot every evening?

Almost always a PoE budget shortfall: IR switches on at dusk, total draw exceeds what the switch can deliver, and the weakest ports are dropped. Check the total budget against night draw.

What causes a whole CCTV network to go down at once?

A network loop — two ports patched together — plus the absence of STP. Broadcast traffic multiplies until the switch is saturated. Managed switches with storm control contain it.

Is a gigabit switch enough for eight cameras?

Yes. Eight 4MP cameras at 6 Mbps produce under 50 Mbps. The usual failure is not the speed but an undersized PoE budget.

Do I need managed switches on a small site?

Not always. Managed becomes worthwhile when there are multiple cabinets, remote sites, shared office traffic, or when each callout is expensive.

Detailed technical specifications are provided in English.

← Learning