Output Ripple: Measurement and Design Implications

Every switching LED driver delivers an output that is not perfectly flat. Superimposed on the nominal 24V DC is a small periodic variation called ripple: a low-frequency component at 100Hz or 120Hz, inherited from the rectified line, plus a high-frequency component at the converter's switching frequency. Ripple is specified on datasheets, measured with oscilloscopes, and judged by what it does to the LED load. It is the electrical cousin of flicker, but it is measured in volts and amps, not in light.

Measuring output ripple correctly requires care. The standard method uses an oscilloscope connected across the output terminals through a short ground lead, with the scope's bandwidth limited, typically to 20MHz, so that the measurement reflects the ripple that matters rather than picking up radiated noise from the switching node. Ripple is expressed as a percentage: peak-to-peak ripple divided by the average output voltage, times 100. A 24V driver with 240mV peak-to-peak ripple has 1% ripple. The measurement must be taken at full load and at the dimming levels of interest, because ripple amplitude changes with load.

The two components of ripple have different causes. The low-frequency component comes from the power factor correction stage: the line delivers power in a sinusoidal envelope at twice line frequency, and the intermediate bus voltage dips and rises accordingly. A larger bulk capacitor reduces this component by storing more energy between line peaks. The high-frequency component comes from the switching converter's output filter: the inductor and output capacitor do not fully remove the switching pulses, leaving a sawtooth or spike residue at the switching frequency. Both components flow through the LED load as current variation.

The LED converts current variation directly into light variation. A 10% current ripple at 100Hz produces a light output that varies at the same frequency, and whether the eye perceives it depends on amplitude and frequency, the subject of IEEE 1789-2015. Even when not visible, current ripple stresses the LED: it raises peak junction temperature, adds a small thermal cycling component, and can contribute to phosphor degradation over decades of operation. Ripple also couples into the audible domain, because the periodic current can make the LED module and the output inductor vibrate at 100Hz.

Designers trade ripple against cost and size. Adding output capacitance reduces ripple but costs money and board space, and electrolytic capacitors have a finite life that limits the driver's service life. Some modern designs use very low capacitance and accept higher ripple, relying on the LED driver topology to limit the damage; others add active filtering. The datasheet's ripple figure, measured at full load, is the contract between the designer and the buyer. A driver with 1% or 2% ripple is typical for well-designed constant-voltage products.

The specification lesson is simple. Ripple matters most in applications with visible dimming, cameras, and sensitive loads, where even a few percent of 100Hz modulation can be detected. For general lighting, a ripple figure below 3% is usually safe, and the PV100's low-ripple output is achieved through the output filter stage rather than through external components. When comparing drivers, compare ripple measured at the same load and the same bandwidth; a number without measurement conditions is not comparable. Ripple is one of the few datasheet parameters where the difference between 1% and 5% is both invisible in a photo and visible on a scope.