How to measure power supply ripple with an oscilloscope

Why ripple and high-frequency noise fool engineers on the bench

Measuring ripple (the residual low-frequency AC voltage generated by power conversion) in switching power supplies seems straightforward at first glance. In practice, any electronics engineer or enthusiast has run into oscilloscope screens filled with ghostly spikes that seem to invalidate the entire circuit design. If you want a deeper understanding of how to diagnose and resolve stability issues, be sure to check out our previous article on voltage drops and transients in power supplies, which perfectly complements bench analysis.

The great secret behind a reliable measurement lies not just in the instrument itself, but in how you connect the probe to the printed circuit board. Switching power supplies work by turning transistors on and off thousands of times per second, generating intense magnetic fields and spurious electromagnetic noise that can mask the true behavior of the DC voltage.

The crucial role of AC coupling in noise visualization

When you plug an oscilloscope probe into a five-volt power rail, for example, the instrument must handle a very high DC voltage relative to the tiny ripple you want to see. In practice, this means that if the oscilloscope is set to DC coupling, the vertical scale must be zoomed out so far to fit the five volts on screen that your twenty-millivolt ripple vanishes into a single pixel.

By switching the channel to AC coupling, an internal capacitor blocks the entire DC component (the steady five-volt level) and lets only the alternating variations around that value pass through. It is precisely this rapid variation that we call ripple and switching noise, allowing you to adjust the vertical sensitivity to the millivolt range and spot the subtlest details of the waveform.

The long ground lead trap and the antenna effect

One of the most classic bench errors involves using the standard ground lead with an alligator clip that comes with almost every oscilloscope probe. On an electronics workbench, that long wire acts exactly like an amateur radio antenna, picking up electromagnetic noise from the environment, the power grid, and even nearby switching power supplies.

To eliminate this issue and ensure you are measuring the actual circuit noise rather than laboratory interference, remove the plastic cap and alligator clip from the probe. Use the ground spring method or wrap a short piece of bare wire around the metal barrel of the probe tip to make a direct, low-inductance connection to the nearest ground plane on the board.

Twenty-megahertz bandwidth limit to clean the spectrum

Modern power supplies operate at high switching frequencies, but surrounding circuits also suffer from parasitic resonances that create very high-frequency voltage spikes. While these spikes exist, trying to measure them with the full bandwidth of the oscilloscope (often exceeding one or two hundred megahertz) usually brings in more measurement noise than useful information.

Most modern digital storage oscilloscopes feature a software-selectable twenty-megahertz bandwidth limit filter in the channel menu. In practice, turning on this filter is essential to remove radio frequency (RF) interference and ultrafast transients that do not affect the practical operation of your low-frequency circuit, stabilizing the trace on the display. To complement this implementation in production environments, see our related technical analyses on How to calculate UPS sizing for modern servers and How to create predictive disk failure alerts using SMART.

Conclusion

Accurately measuring switching power supply ripple requires rigorous attention to the physical details of the test setup, from choosing AC coupling to eliminating the traditional ground wire. By mastering these advanced noise-isolation techniques, you prevent false diagnostics, ensure the stability of sensitive microcontrollers, and guarantee that your hardware designs operate reliably in any real-world environment.