Loto practical tips (1) Correctly and elegantly test power supply ripple
I find that the need to measure power supply ripple is becoming more and more common, not just for power supply engineers. Ripple is the core indicator of power supply, and power supply engineers naturally need to pay attention to ripple at all times. Hardware engineers in other fields will also have to deal with power supply ripple more and more. Because as our experience accumulates, the circuits we make and touch become more and more complex, and we are more and more aware of how important a clean power supply system is to the stability of the circuit. I myself have encountered inexplicable and illogical hardware problems many times, and finally found that it was caused by a dirty power supply system. So when we debug more complex circuits, especially mixed digital and analog circuits, we need to be highly vigilant about power supply ripple.
Power ripple is the most intuitive manifestation of power supply performance, and is the fluctuation of power supply output voltage. If it is a switching power supply, the output ripple is a regular swing, and the frequency of the swing is equal to the switching frequency. The ripple is formed because the current flowing through the output capacitor causes a voltage drop on the ESR of the capacitor. In the switching power supply, there is a constant pulsating current flowing through the capacitor, so its ripple frequency is equal to the switching frequency. The ripple of the power supply system superimposed on the noise will be transmitted to the power pin of the chip, affecting the performance of the chip.
Take the power system test of a motion control board I just made to show you the test process and method of power ripple intuitively, and take notes for yourself. As shown in the figure below, M is my motion control board, P is the switching power supply that powers M, and S is the USB oscilloscope OSC802 from Loto.
The power supply schematic and PCB layout of the motion control board M are shown below:
We need to test the power ripple at both ends of V and G to verify whether the power supply system meets the design requirements.
Of course, first we know that we need to use the AC coupling position of the oscilloscope to test the ripple size and remove the DC component that we don’t care about. The first time I used a longer ground wire of the oscilloscope probe for convenience, as shown in the figure below:
The ripple of the power supply itself may not be large. Due to the long ground wire, the loop path shown by the yellow line in the figure will absorb the radiation interference in this area and superimpose it on the actual ripple. In this case, my test results are as follows:
I measured about 80mv of ripple, and even superimposed more than 100mv of noise (using the 50mv/div setting of OSC802). Because I was worried that this loop path would be affected by noise interference, I improved the test method to minimize the loop area and reduce external interference to a minimum. I learned that the loop area is very small when measured using a grounding spring. However, I didn't have a grounding spring on hand, so I used a paper clip to straighten it and wound a simple grounding spring.
This allows a very small loop area to be used to measure the ripple between V and G.
The measurement results have indeed improved a lot. The ripple has become very clean and neat. A ripple of about 20mv was measured (using the 20mv/div position of OSC802), as shown in the following figure:
It seems that the ripple of the power supply system is relatively clean and stable at about 20mv, which basically meets our original design requirements. At the same time, we also saw that the test results were so different just by changing the length of the ground wire (actually changing the loop area). To be cautious, I borrowed a desktop oscilloscope and repeated the two operations. The virtual oscilloscope and the desktop digital oscilloscope reached the same conclusion, with 80mv and 20mv ripples respectively:
So it is certain that to measure power supply ripple, AC coupling is indeed required, and then a grounding spring is used to connect the ground to minimize the loop area. If there is no grounding spring, you can make one yourself. In addition, the USB oscilloscope OSC802 is equally good at measuring power supply ripple, and a 20mv ripple can be easily measured and clearly displayed.
I replayed the whole process and recorded it as a video and put it on Youku to facilitate communication between everyone:
https://v.youku.com/v_show/id_XNDA0NDcxMTAxNg==.html?spm=a2h0j.11185381.listitem_page1.5~A
QQ:173393190.
提升卡
变色卡
千斤顶
京公网安备 11010802033920号
