Learning Library · Frequency domain · 3
Reading a spectrum analyzer
A spectrum analyzer plots power against frequency. Its vertical axis is in dBm, its horizontal axis spans a band you choose, and one setting more than any other decides what you can see: the resolution bandwidth, RBW.
Two signals, one at the centre frequency and one offset from it, over a noise floor. The trace redraws as the analyzer sweeps. Noise figure is set to a typical 20 dB for illustration.
dBm in one table
dBm is power relative to one milliwatt. Every 10 dB is a factor of ten in power, and every 20 dB a factor of ten in voltage. On a 50 Ω input:
| Level | Power | Voltage (RMS, 50 Ω) |
|---|---|---|
| +10 dBm | 10 mW | 707 mV |
| 0 dBm | 1 mW | 224 mV |
| −30 dBm | 1 µW | 7.07 mV |
| −60 dBm | 1 nW | 224 µV |
| −90 dBm | 1 pW | 7.07 µV |
What RBW does
The analyzer looks at the band through a narrow filter and steps or sweeps it across the span. The width of that filter is the RBW. It decides three things at once:
- Resolution: two signals closer together than about one RBW merge into one peak.
- Noise floor: a narrower filter lets in less noise. Every factor of ten narrower lowers the displayed noise floor by 10 dB, which reveals smaller signals.
- Speed: a narrow filter takes longer to settle, so sweep time grows with the square of 1 ÷ RBW. Ten times narrower is about a hundred times slower over the same span.
Try this
- Set the offset to 10 kHz and the RBW to 30 kHz. The two signals blur into one shape. Narrow the RBW to 3 kHz and two peaks appear.
- Set the offset to 500 kHz, the RBW to 100 kHz, and the second signal to 85 dB below the first. It is lost in the noise. Narrow the RBW to 1 kHz: the floor drops 20 dB and the signal stands clear of it.
- Watch the sweep time readout each time you narrow the RBW.
Two more settings matter for small signals. The reference level sets the top of the screen and the input attenuation: keep the strongest signal just below it so the analyzer's input is not overloaded. Video bandwidth (VBW) or trace averaging smooths the noise so a signal near the floor is easier to read, without changing the floor itself.
Find the signal under the noise. Spectrum and signal analyzers.
Shop UNI-T spectrum analyzersUNI-T Learning Library. Design. Debug. Deploy.™