Power Spectral Density
Understanding Power Spectral Density
PSD is the frequency-domain description of power distribution. It is essential for understanding noise performance, spectral efficiency, and interference. Every noise figure, phase noise, and spectral emission specification is ultimately a PSD measurement.
Key PSD Values
| Source | PSD |
|---|---|
| Thermal noise (290K) | -174 dBm/Hz |
| Good LNA output noise | -173 to -171 dBm/Hz |
| Typical receiver noise floor | -170 to -164 dBm/Hz |
| Spectrum analyzer noise floor | -155 to -140 dBm/Hz |
PSD and Total Power
P = PSD + 10*log(B)
Example: -174 dBm/Hz in 10 MHz BW:
P = -174 + 10*log(10e6) = -174 + 70 = -104 dBm
Frequently Asked Questions
What is power spectral density?
PSD describes power distribution per Hz across frequency. Thermal noise = -174 dBm/Hz. Total power = PSD + 10*log(bandwidth). PSD is the fundamental characterization of signals and noise in the frequency domain.
Why is -174 dBm/Hz important?
-174 dBm/Hz is the thermal noise floor at room temperature (290K). It is the absolute minimum noise that any receiver will encounter. System sensitivity = -174 + NF + 10*log(BW). This defines the ultimate receive sensitivity.
How do I measure PSD?
Use a spectrum analyzer with known resolution bandwidth (RBW). The displayed power divided by RBW gives the PSD: PSD (dBm/Hz) = measured power (dBm) - 10*log(RBW in Hz). Modern analyzers can display directly in dBm/Hz.