For an ideal white-noise process used in communications engineering, how does the power spectral density (PSD) vary with frequency?

Introduction / Context:
White noise is a fundamental model in receiver analysis, representing thermal noise and other wideband disturbances. Understanding its power spectral density (PSD) is essential for calculating noise power in a given bandwidth and predicting signal-to-noise ratios.

Given Data / Assumptions:

• White noise is an idealization with uniform PSD across frequency.
• Real systems are bandlimited by front-end filters and components.
• Thermal noise approximates white within limited bandwidths.

Concept / Approach:

By definition, white noise has a flat PSD: S_n(f) = N₀/2 (single-sided conventions vary). Total noise power is PSD multiplied by the receiver bandwidth. While 1/f noise (flicker) and device corner frequencies exist, these are separate phenomena and not part of the ideal white-noise definition.

Step-by-Step Solution:

Verification / Alternative check:

Receiver noise figure measurements assume a flat thermal-noise density kT over the measurement bandwidth, consistent with white-noise modeling.

Why Other Options Are Wrong:

• Increasing/decreasing with frequency: describes colored noise, not white.
• 1/f: flicker noise, significant at very low frequencies but not white.
• Bandlimited to IF only: limitation of the receiver front end, not of white noise itself.

Common Pitfalls:

Confusing white noise (flat PSD) with pink or flicker noise that has frequency-dependent spectra.

Final Answer:

It is constant with frequency (flat spectrum)