Detector Type
Understanding Detector Type
A swept spectrum analyzer tunes its local oscillator across the span, and the resulting IF signal passes through the resolution bandwidth filter. At each frequency point, the analyzer must condense the continuous IF output into a single number for display. The Detector Type determines how that condensation happens.
For a pure CW tone, every detector produces the same answer because the IF signal is constant. The differences emerge with modulated signals, pulsed waveforms, and broadband noise, where the IF amplitude fluctuates within each measurement interval. A peak detector captures the highest crest. A sample detector grabs one instant. An RMS detector computes true power. These differences can easily amount to 10 dB or more on a wideband signal like 5G NR or a radar pulse train.
Detector Comparison
| Detector | Algorithm | CW Signal | Noise Floor | Pulsed Signal | Primary Use |
|---|---|---|---|---|---|
| Positive Peak | Maximum value in bin | Correct | Reads high (~2.5 dB) | Reads peak amplitude | Spurious search, EMC pre-scan |
| Negative Peak | Minimum value in bin | Correct | Reads low | Reads noise between pulses | AM modulation depth |
| Sample | Single random sample per bin | Correct | Correct (statistically) | May miss short pulses | Noise measurements, random signals |
| RMS (Power Average) | Root mean square of IF | Correct | Correct | True average power | Channel power, ACLR, noise density |
| Quasi-Peak (QP) | Weighted charge/discharge | Correct | Between peak and avg | Weights repetition rate | EMC compliance (CISPR 16) |
| EMI Average | Linear voltage average | Correct | Reads low (~1.05 dB) | Linear mean amplitude | EMC compliance above 1 GHz |
PRMS = √( (1/N) × Σ |Vi|² )
Displayed Level (dBm):
L = 10 × log10(PRMS / 1 mW)
Peak-to-Average Difference for Noise:
Peak detector reads approximately 2.5 dB above true RMS for Gaussian noise.
Sample detector reads correct RMS if sufficient averaging is applied.
For a 5G NR signal with 12 dB PAPR: Peak detector reads ~12 dB above RMS detector.
Frequently Asked Questions
Why does changing the detector type change the displayed amplitude?
Because each detector processes the IF signal differently. A modulated or pulsed signal has amplitude that varies within each measurement bin. A positive peak detector captures the highest instantaneous value. An RMS detector computes the true power average. For a CW tone, all detectors agree. For a noise-like 5G NR signal, peak detection might read 10 to 12 dB higher than RMS because the peak captures the highest crest, not average power.
Which detector type should I use for EMC compliance testing?
CISPR 16 and FCC Part 15 mandate specific detectors. Pre-compliance scans use peak detection for speed: if peak passes, the device passes regardless. Final compliance for conducted emissions below 30 MHz requires the quasi-peak detector per CISPR 16-1-1. The QP detector has specific charge (1 ms) and discharge (160 ms for Band B) time constants that weight repetitive signals higher than random bursts.
What is the difference between video averaging and an RMS detector?
Video averaging averages logarithmic (dB) values across successive sweeps. This reduces trace noise but introduces about 2.5 dB systematic error when measuring noise, because averaging in log scale is not the same as averaging in linear scale. An RMS detector computes root-mean-square of the IF power in linear units before converting to dB, giving the mathematically correct power reading without the 2.5 dB log-averaging error.