Learning Library · Frequency domain · 4
Scope FFT vs. spectrum analyzer
Most oscilloscopes have an FFT button, so why buy a spectrum analyzer? Feed both the same signal and the answer shows up in the harmonics: a scope's own distortion sits right where you are looking, and its dynamic range depends on settings most people never touch. A scope does things an analyzer cannot, too.
One 14.2 MHz transmitter sample with a second harmonic at −50 dBc, a third at −62 dBc, and a spur at 23.1 MHz at −78 dBc. Top: an oscilloscope FFT at 1 GSa/s. Bottom: a spectrum analyzer. ADC distortion, effective bits, and the analyzer's noise and distortion are illustrative; check the datasheet of the instrument in front of you.
What sets each floor
On the scope, the noise floor of an FFT bin is the ADC's noise spread over the bins: the more points in the record, the lower each bin's share. That processing gain is 10 × log10(N ÷ 2) dB, 37 dB at 10k points and 57 dB at 1M. Using only half the screen height costs 6 dB, because the noise stays the same while the signal shrinks. The scope's own harmonic distortion does not average away; it lands on the transmitter's harmonics and adds to them.
On the analyzer, a narrow resolution bandwidth (RBW) filter in front of the detector sets the floor, 10 dB lower for every factor of ten narrower, and the input attenuator and preamplifier let you put the signal where the mixer is cleanest. That is how an analyzer reaches 80 to 100 dB or more between the carrier and what it can see. See reading a spectrum analyzer and ENOB.
When you need which
| Use the scope FFT when… | Use a spectrum analyzer when… |
|---|---|
| You need to see when something happens: key-down, PTT, an amplifier switching in, a burst | You need to prove a level: harmonics 43 or 60 dB down, spurs, a signal near the noise |
| The signal is audio or low-frequency, or you want several channels at once | You need narrow resolution, Hz to kHz, to separate close signals |
| A rough look at the spectrum is enough, and the scope is already on the bench | The span reaches past the scope's bandwidth: VHF and UHF harmonics, GHz |
| You want to trigger on an event and look at its spectrum | You want calibrated dBm, a tracking generator for filters and antennas, or EMI pre-compliance |
Try this
- On the 8-bit scope with 10k points, the 2nd harmonic reads −43.8 dBc. A transmitter that passes the 43 dB HF limit with 7 dB to spare looks borderline: the extra comes from the scope's ADC.
- Switch to 12-bit. The 2nd harmonic reads −49.7 dBc, close to the truth, and the 3rd −59.0 dBc, still 3 dB high.
- On the 8-bit scope the 23.1 MHz spur is under the floor at 10k points (−78 dBc) and at 1k (−68 dBc). 100k points lowers the floor to −88 dBc and brings it back.
- On the analyzer, both harmonics read within 0.1 dB of the truth. At 100 kHz RBW even the analyzer loses the spur; narrow the RBW to 1 kHz and the floor drops 20 dB.
The right tool for the question. 12-bit oscilloscopes for time and spectrum, analyzers for dynamic range.
Shop UNI-T spectrum analyzers Shop UNI-T oscilloscopesUNI-T Learning Library. Design. Debug. Deploy.™


