Do you actually need a spectrum analyser?

Wi-Fi tools show you Wi-Fi. That sounds obvious until you spend two days troubleshooting a channel that is 80 percent utilised, where your own access points account for 15 percent of the traffic and no neighbouring SSID accounts for the rest.

Something is transmitting. Your Wi-Fi adapter cannot see it, because it is not Wi-Fi.

What only a spectrum analyser will find

• Microwave ovens — a broad, intermittent smear across the upper half of 2.4 GHz, correlated suspiciously well with lunchtime.
• Analogue video senders and older wireless cameras.
• Faulty fluorescent ballasts, arcing motors and failing power supplies.
• Wireless headsets, presenters, and DECT variants.
• Radar, before DFS reacts to it.
• Zigbee, Bluetooth, proprietary industrial telemetry, warehouse RF tags.
• A transmitter with a broken filter, splattering across channels it should not touch.

None of these appear in a packet capture. Several appear as "the Wi-Fi is inexplicably terrible in this one area".

When you can skip it

If you support a small number of conventional offices, you may go years without needing one. Some enterprise access points include spectrum monitoring — Cisco's CleanAir being the best example — and it catches the obvious cases.

Note that UniFi does not offer per-AP spectrum analysis. In a UniFi estate, a separate tool is the only option.

When you need one

Warehouses, factories, healthcare, and any site where you have exhausted the Wi-Fi-side explanations. Also consulting work, where "I do not know" is not a deliverable.

Choosing one

USB spectrum analysers in the mid-price range give real-time FFT, waterfall and duty-cycle views on a laptop. For most engineers this is the right tool.

Handheld and benchtop analysers cost considerably more and give better sensitivity and frequency range. Justified for RF-heavy environments.

The practical test: if you support more than a handful of sites, the tool pays for itself the first time it turns a two-week mystery into a twenty-minute answer.

Using one well

• Capture a baseline while everything is working. Interference is only obvious relative to normal.
• Look at duty cycle, not just amplitude. A weak signal transmitting constantly hurts more than a strong intermittent one.
• Walk the area rather than sitting in one place.
• Correlate with time of day. The strongest clue is usually "it happens at 12:15 every weekday", and that clue is free.