This article introduces the concepts of a DC power supply ripple tester and ripple coefficient. Four methods to suppress power supply ripple are proposed: RLC filtering, common-mode filtering, ferrite bead filtering, and combined filtering. Through experiments, the filtering effects of the four methods are obtained, and their advantages and disadvantages are summarized. By comparison, it is concluded that the combined filtering method can effectively reduce power supply ripple.
Introduction
The main function of a power supply is to provide electrical energy to electronic products, but while supplying power, it inevitably introduces ripple, noise, etc., which reduces the stability and reliability of the electronic system and even the entire product. Voltage ripple can greatly affect various circuits of the power supply, such as A/D conversion circuits, operational amplifier circuits, rectifier filter circuits, etc. In typical applications, the following hazards exist: generating unwanted harmonics that cause overvoltage or overcurrent and lead to accidents; increasing additional losses, reducing the efficiency and utilization of electrical equipment; causing abnormal operation, accelerating aging, and shortening lifespan; causing malfunction or failure of protective relays, automatic devices, computer systems, and other equipment; causing deviations in measurement and metering instruments; interfering with communication systems, reducing signal transmission quality, and even damaging communication equipment. Therefore, when designing electronic products, it is necessary to accurately measure ripple and suppress it within a certain range.
1 Power Supply Ripple and Ripple Coefficient
Strictly speaking, a regulated power supply includes four parts: power transformer, rectifier circuit, filter circuit, and voltage regulator circuit. Since DC-DC can also be considered a regulated power supply, the rectifier circuit, filter circuit, and voltage regulator circuit are regarded as the three essential parts of a regulated power supply [1]. The rectifier circuit uses unidirectional conducting devices to convert AC into pulsating DC, which is not smooth and contains a large amount of AC components. The filter circuit uses energy storage elements to convert pulsating DC into relatively flat DC. Due to the varying performance of filter circuits, although most AC components can be filtered out, they cannot be completely eliminated. The voltage regulator circuit after rectification and filtering uses the adjustment function of the circuit to stabilize the output voltage and minimize AC components. This AC component that accompanies the stable output voltage and cannot be completely filtered out is called ripple voltage.
To characterize the filtering performance of a DC regulated power supply, the concept of ripple coefficient is introduced [2-3]. The ripple coefficient ψ is defined as the percentage of the RMS value of ripple voltage Vr to the DC output voltage Vo, i.e.:
The ripple coefficient is an important indicator for evaluating the stable and pure output of a DC power supply. According to the above formula, to obtain the ripple coefficient, the ripple voltage needs to be measured.
2 Measurement of Power Supply Ripple
Accurately measuring power supply ripple generally requires two instruments: an electronic load and a digital storage oscilloscope (DSO). The electronic load facilitates current adjustment and is typically set to constant resistance (CR) mode; the digital storage oscilloscope can directly capture the entire ripple waveform, store, amplify, and read out the ripple value. Substituting the oscilloscope reading into the formula yields the ripple coefficient.
When measuring, the following two points must be noted (these two points are particularly important for the accuracy of the measurement results):
(1) The ground lead of the digital storage oscilloscope probe must be removed and replaced with the grounding spring pin from the probe accessory. This prevents ground loop coupling of EMI noise, which would make the measurement results inaccurate. As shown in Figure 1, the probe ground lead is too long and the loop area is too large, forming a receiving antenna; high-frequency noise or EMI noise can couple into the measured signal.
(2) The digital storage oscilloscope itself needs to be adjusted. The DSO must have good grounding to further filter out noise from the power source; use AC coupling on the DSO to block DC and make ripple testing more intuitive and accurate; general ripple testing requires the frequency to be limited to below 20MHz, so the DSO should enable the 20MHz bandwidth limit to isolate high-frequency noise.
3 Methods to Suppress Power Supply Ripple
To suppress the ripple of the regulated power supply output voltage, the following four methods are generally adopted: RLC filtering, common-mode filtering, ferrite bead filtering, and a combination of the three methods.
Experiments are conducted to verify the filter circuits that suppress DC-DC power supply ripple. In the verification experiment, a 100W DC-DC power supply with 48V input and 5V output, model Meanwell SD-100C-5, is selected. The digital storage oscilloscope is a GWINSTEK GDS-1072B with a bandwidth of 70MHz, sampling rate of 1GSa/s, and maximum storage depth of 10M per channel. The electronic load is a GWINSTEK PEL-3021 with a voltage range of 1.5V~150V, current range of 0~35A, and power of 175W. Based on this, the maximum current in the circuit is 20A. Figure 3 shows the connection diagram for power supply ripple testing.
To make the effect of suppressing power supply ripple more intuitive and obvious, the filter circuit of the SD-100C-5 is first short-circuited, and the ripple of its output voltage is measured. From Figure 4, the power supply ripple is approximately 85.6mVpp, with an RMS value of 48.2mVrms.
3.1 RLC Filtering Method
The RLC filtering method does not necessarily use all resistors, inductors, and capacitors. Based on theoretical calculations and empirical values, a capacitor-only filtering method or an inductor-capacitor combined filtering method can be adopted, both achieving relatively ideal filtering effects. To briefly illustrate the RLC filter design, this article uses the capacitor-only filtering method as an example to demonstrate its effectiveness. A capacitor filter consisting of a 0.1μF and a 47μF capacitor in parallel is connected to the output of the SD-100C-5 power supply. The power supply is turned on, and the output current of the electronic load is adjusted to 20A. The output voltage ripple is measured with a digital storage oscilloscope, and the ripple magnitude is shown in Figure 5. The power supply ripple is approximately 18.4mVpp, with an RMS value of 10.2mVrms. Compared with no filter circuit, the power supply output ripple is reduced.
If the LC combined filtering method is used, with a 150μF capacitor connected in parallel at the power supply output and a 6μH inductor connected in series on each of the positive and negative output terminals, similar results can be obtained. Compared with the capacitor-only filtering method, there is improvement, but the difference is not significant.
3.2 Common-Mode Filtering Method
A common-mode choke and Y capacitors are added to the output of the SD-100C-5 power supply for decoupling to attenuate common-mode ripple. The parasitic inductance in the choke forms an LC filter to filter out differential-mode ripple.
The power supply is turned on, and the output current of the electronic load is adjusted to 20A. The output voltage ripple is measured with a digital storage oscilloscope, and the ripple magnitude is shown in Figure 7. The power supply ripple is approximately 13.2mVpp, with an RMS value of 9.14mVrms. Compared with the RLC filtering method, the output ripple is reduced by 5.2mVpp.
A common-mode choke is a coil wound on a closed magnetic ring with opposite directions and the same number of turns. When common-mode current flows through the coil, it generates a magnetic field in the same direction, thereby increasing the inductive reactance of the coil, showing a stronger damping effect, thus suppressing common-mode current and achieving the purpose of filtering out ripple [4]. Although the above comparison shows that the common-mode filtering method is more effective than the RLC filtering method, it is worth noting that the common-mode choke is several times larger in volume than capacitors or inductors. If the effective area of the power supply circuit board is considered, especially when the power density requirement of the switching power supply is high, the common-mode choke cannot be used.
3.3 Ferrite Bead Filtering Method
A ferrite bead is directly threaded onto the power cable at the output of the SD-100C-5 power supply, as shown in Figure 8. To better illustrate the filtering effect, a TDK snap-on ferrite bead is selected. The power supply is turned on, and the output current of the electronic load is adjusted to 20A. The output voltage ripple is measured with a digital storage oscilloscope, and the ripple magnitude is shown in Figure 9. The power supply ripple is approximately 12.4mVpp, with an RMS value of 10.1mVrms. It is very close to the common-mode filtering method, but the ferrite bead is simple and can be customized according to the size requirements of the power supply. Therefore, it is now widely used in printed circuit boards, data cables, and power cables for filtering.
