Written and maintained by CASRAI Editorial Board
Last updated
A spectrum analyzer is a test instrument that shows how the energy of an electrical signal is distributed across frequency. This guide is for lab staff, core-facility users and equipment buyers who need to understand how the three common architectures (swept, FFT and real-time) differ, what resolution bandwidth does to sweep time, and which specifications to check when comparing instruments. It covers radio-frequency (electrical) analyzers; optical and audio spectrum analyzers are separate instrument families and are not covered here.
What is a spectrum analyzer?
A spectrum analyzer takes a signal and plots it in the frequency domain, with frequency on one axis and signal level on the other, rather than level against time as an oscilloscope does. Its purpose is to show which frequencies are present in a signal and how strong each one is, so that closely spaced signals, noise and interference can be separated. Who uses it: electronics and RF engineers, instrument-development groups and test labs. For the underlying mathematics see what is signal processing; for a different kind of spectrum instrument see what is a mass spectrometer.
How a swept spectrum analyzer works
In the swept architecture, Rohde & Schwarz describes the signal path as input, then mixers, then a final intermediate-frequency (IF) stage. A local oscillator sweeps across the chosen frequency span, so the analyzer measures one frequency at a time. Near-Gaussian analog IF filters form the resolution bandwidth (RBW). A detector follows, and then a video filter, which is used to reduce the noise on the trace (Rohde & Schwarz application note 1EF80, K.-U. Sander, 2012). VIAVI describes the same principle as measuring each frequency sequentially and plotting the results, and notes that swept analyzers dominated the market for about 50 years.
RBW and sweep time: a worked example
The resolution bandwidth is the frequency span of the final filter applied to the input signal (Siglent application note on RBW and VBW, 2018). A smaller RBW gives finer resolution and separates signals that sit close together; Rohde & Schwarz states that RBW defines the ability to resolve closely spaced signals and is the main limitation of sweep speed.
Siglent gives the relationship: sweep time is proportional to the span and inversely proportional to the square of the RBW. Worked example from that rule: if you reduce the RBW by a factor of 100, the sweep time grows by a factor of 100 squared, which is 10,000. Siglent also notes that a smaller RBW lowers the displayed noise floor, but at the cost of a longer sweep. The video bandwidth (VBW) filter smooths the trace and decreases displayed noise.
Swept vs FFT vs real-time analyzers
| Type | How it measures | Strength | Limitation |
|---|---|---|---|
| Swept | Measures each frequency sequentially and plots the results (VIAVI) | Long-established design; controls include RBW, VBW, dwell time, averaging and max hold (VIAVI) | Can miss transient events; in the VIAVI example, with RBW typically 2% or less of span, there is a 98% chance of missing a transient |
| FFT | An analog-to-digital converter captures the spectral information simultaneously (VIAVI) | Whole span is captured at once | Sampling must meet Nyquist (at least twice the bandwidth), an anti-alias filter is needed, a window function such as Hann is used to avoid energy smearing, and processing gaps mean events outside the capture window are missed (VIAVI) |
| Real-time (RTSA) | Same concept as FFT, but capture and calculation happen simultaneously; an FPGA applies the Hann window and FFT (VIAVI) | No processing gaps (VIAVI) | Not addressed in the sources cited here |
How to choose: checklist from the sources
Only criteria stated in the Siglent and VIAVI notes are listed here.
- Look for an analyzer with many RBW settings, especially at the low end (Siglent).
- Check that RBW and VBW are adjustable (Siglent).
- Check that the analyzer’s phase noise is lower than the signal you intend to measure (Siglent).
- If you need to catch short, intermittent events, consider the swept architecture’s transient limitation and the FFT capture-window gaps described above before choosing (VIAVI).
Common mistakes
- Expecting a fast sweep at a very narrow RBW: because sweep time scales with the inverse square of RBW, a 100-times narrower RBW is 10,000 times slower (Siglent).
- Using a swept analyzer to hunt for brief transients, where the chance of missing the event can be high (VIAVI).
- Ignoring the Nyquist and anti-alias requirements when using an FFT-based instrument (VIAVI).
FAQ
What is a spectrum analyzer?
It is an instrument that displays signal level against frequency, showing which frequencies are present in a signal and how strong each is. Swept, FFT and real-time analyzers are the three architectures compared above.
What does RBW mean on a spectrum analyzer?
Resolution bandwidth is the frequency span of the final filter applied to the input signal; a smaller RBW resolves closer signals but lengthens the sweep (Siglent).
Sources (read 2026-11-07)
- Rohde & Schwarz, application note 1EF80, K.-U. Sander, July 2012: 1EF80 (PDF)
- VIAVI Solutions, application note “Choosing the Right Spectrum Analyzer”: application note (PDF)
- Siglent, application note on RBW and VBW, 2018: Siglent resource 38
Related reading: what is analytical chemistry, the Fourier transform in NumPy.
