How to Accurately Measure Power Supply Ripple

Measuring power supply ripple itself is a skill. Figure 1 shows an example of improper use of an oscilloscope to measure power supply ripple. Several mistakes were made in this example. The first was using an oscilloscope probe with a long ground lead. The second was to allow the loop formed by the probe and the ground lead to be close to the power transformer and switching components. Finally, additional inductance was allowed to form between the oscilloscope probe and the output capacitor. The resulting problem is that the measured ripple waveform carries the picked-up high-frequency components.

There are many high-speed, high-voltage and current signal waveforms in the power supply that can easily couple into the probe, including magnetic field coupling from the power transformer, electric field coupling from the switch node, and common-mode current generated by the transformer interwinding capacitance.

Figure 1: Improper ripple measurement yields poor results

The results of ripple measurement can be improved by using the correct measurement technique. First, the bandwidth upper limit of ripple is usually specified to avoid picking up high-frequency noise that exceeds the bandwidth upper limit of ripple. The bandwidth upper limit of the oscilloscope used for measurement should be set appropriately. Second, the antenna formed by the long ground lead can be removed by removing the "cap" of the probe. As shown in Figure 2, we wrap a short wire around the probe ground lead and connect it to the power ground. This has the added benefit of shortening the probe length exposed to the high-intensity electromagnetic radiation near the power supply, thereby further reducing high-frequency pickup.

Finally, in an isolated power supply, the real common-mode current is generated by the current flowing in the probe ground lead, which causes a voltage drop between the power ground and the oscilloscope ground, which appears as ripple. To suppress this ripple, common-mode filtering needs to be carefully considered in the power supply design.

In addition, wrapping the oscilloscope lead around the core can reduce this current because it will form a common-mode inductance that does not affect the differential voltage measurement but can reduce the measurement error caused by the common-mode current. Figure 2 shows the ripple voltage measurement results of the same circuit using the improved measurement technique. It can be seen that the high frequency spikes have been almost eliminated.

Figure 2: Four simple improvements greatly improve measurement results

In fact, once the power supply is integrated into the system, the power supply ripple performance is even better. There is almost always some amount of inductance between the power supply and the rest of the system. The inductance may be formed by wire or etched lines on the printed circuit board, and there are always additional bypass capacitors near the chip as the power supply load, which form a low-pass filtering effect and further reduce the power supply ripple and/or high-frequency noise.

As an extreme example, a filter composed of a one-inch short wire with an inductance of 15nH and a bypass capacitor with a capacitance of 10μF has a cutoff frequency of 400kHz. This example means that high-frequency noise can be greatly reduced. The cutoff frequency of this filter is many times lower than the power supply ripple frequency, which can actually reduce the ripple. Smart engineers should try to take advantage of this during testing. (end)