Power Supply Ripple and Noise in Industrial Equipment: Causes, Impacts, and Solutions

Defining the Disruption: What Are Ripple and Noise?

Before diving into solutions, it is crucial to establish a clear technical distinction between ripple and noise.

Power Supply Ripple: The Predictable Residual
Ripple is the periodic AC component superimposed on the DC output. It is deterministic and synchronous with the power conversion process.

In Linear Supplies: Ripple appears at twice the mains frequency (100 Hz or 120 Hz), resulting from incomplete smoothing of the rectified sine wave.

In Switch-Mode Power Supplies (SMPS): Ripple is dominated by the switching frequency (typically 50 kHz to several MHz) and its harmonics. In a buck converter, the peak-to-peak ripple current in the inductor is defined by the duty cycle, switching frequency, and inductance value.

Because ripple is predictable, it can be quantified and addressed through deterministic methods: increasing inductance, adding low-ESR output capacitance, or adjusting switching frequency.

Power Supply Noise: The High-Frequency Anomaly
Noise is the non-periodic, high-frequency component superimposed on the DC output. Unlike the clean waveform of ripple, noise consists of narrow voltage spikes, ringing, and broadband energy.

Primary Sources of Noise:

Switching Transitions: The rapid rise/fall times (dv/dt and di/dt) of MOSFETs translate into voltage spikes via parasitic capacitance and inductance.

Diode Reverse Recovery: The abrupt cessation of current during diode turn-off generates significant voltage transients.

Parasitic Coupling: High-frequency energy couples from the switching node to the output through transformer inter-winding capacitance or PCB trace coupling.

Ground Bounce: Current commutation through a finite-impedance ground plane creates differential noise at the output terminals.

The Key Distinction: Ripple is a low-to-moderate frequency periodic disturbance; noise is a high-frequency, non-periodic disturbance caused by switching edges and parasitic elements. Noise amplitude is far more dependent on PCB layout and parasitic elements than on component values alone.

Measuring Ripple and Noise: Best Practices to Avoid Misreadings

The most common mistake in power supply evaluation is inaccurate measurement. A standard oscilloscope probe with a 10 cm ground lead forms a loop antenna that picks up radiated switching noise, resulting in readings that are orders of magnitude higher than reality.

Correct Measurement Configuration
Probing Methods:

Tip-and-Barrel: Use a probe tip with a short spring ground directly at the measurement point to minimize loop area.

Coaxial Connection: Solder a short coaxial cable directly across the output capacitor and terminate the scope at 50Ω for a controlled impedance path.

Differential Probe: Essential for floating outputs or high common-mode noise environments.

Bandwidth Limiting: The industry standard (per JEITA RC-9131B) specifies a 20 MHz bandwidth limit. This filter rejects high-frequency artifacts caused by probe resonance or radiated pickup. However, for high-speed communication systems, consider measuring with full bandwidth to capture all relevant content.

Measurement Setup: Always use AC coupling to remove the DC offset, enabling higher vertical resolution. Measure at the load pins, not just the supply terminals, to account for cable and trace impedance.

LEYU’s Commitment to Standardization: At LEYU, all ripple and noise measurements are conducted per industry standards using a 20 MHz bandwidth oscilloscope, with 12-inch twisted pair wires terminated with 0.1μF and 47μF capacitors. This ensures consistent, reliable, and comparable specifications across the LRS, ZDR5, and DC DC series.

Previous