Misunderstandings and solutions of electronic loads in LED power supply testing

　　The CV mode of the electronic load is the basis of LED power supply testing. CV means constant voltage, but the load is only a current-pulling device and cannot provide constant voltage by itself. Therefore, the so-called CV is only to servo the change of the output current of the LED power supply through the voltage negative feedback circuit to balance the charge on the LED output capacitor and achieve the purpose of constant voltage. Therefore, there are two core factors that determine the accuracy of CV:

　　Load Bandwidth

　　The size of the LED power supply output capacitor

　　When the ripple frequency of the LED power supply output current is very high, if the load bandwidth is insufficient, the servo current change cannot be maintained, causing oscillation. When oscillation occurs, the load input voltage changes sharply, and the LED output capacitor is frequently charged and discharged with large currents. At this time, the detected current ripple will be much larger than the actual current ripple when the LED power supply is working in a steady state.

　　When the load bandwidth is insufficient, if the output capacitance of the LED power supply is large enough, the oscillation amplitude can be controlled within an acceptable range. Unfortunately, the price competition of LED power supplies is very fierce, and the output capacitance is generally insufficient. Therefore, the load bandwidth requirements for testing LED power supplies are very stringent.

　　Manufacturers will not directly indicate the bandwidth index of the load. You can only refer to another index: full-scale current rise time. Obviously, the smaller the full-scale current rise time, the higher the bandwidth of the load. The higher the load bandwidth, the lower the requirement for the output capacitance of the LED power supply. Generally speaking, a load with a full-scale current rise time of 10uS can meet the testing needs of most LED power supplies, but theoretically, any load in CV mode may oscillate. In this case, when the LED output capacitance remains unchanged, the higher the load bandwidth, the smaller the oscillation amplitude, and the higher the confidence of the test result. Therefore, when users use electronic loads for testing, they must pay close attention to the changes in the load input voltage ripple Vpp. Once it exceeds the range, the entire test result is no longer credible. This is very important and users must keep it in mind.

　　In CV mode, the voltage is constant, while the current ripple is usually very large. In order to improve the test efficiency, the load often refreshes the data at a high frequency, so the data jumps a lot. Many users use this to determine whether the load is suitable for LED testing. In fact, this is a very serious misunderstanding. The stability of the data is actually very easy to achieve. It only requires increasing the time measurement of the data filter. Very short low-end electronic loads have to filter on a large time scale because of their low measurement accuracy, but it turns out to be a blessing in disguise, making the data seem more stable. In fact, this is an illusion. To achieve accurate measurement, the fundamental method can only be to increase the sampling rate. If the sampling rate is not increased, the confidence of such measurement results is very low, which may cause serious quality accidents.

　　Based on the above analysis, LED power supply testing has strict requirements on the load, mainly including the following points:

　　The full-scale current rise time is the basis for ensuring accurate load carrying. The lower this value is, the better.

　　Data sampling rate is the basis for accurate measurement. The higher the value, the better.

　　Real-time display of Vpp is the basis for judging whether the measurement data is credible;

　　The filter speed adjustment function is a small means to obtain stable current data;

　　Finally, please note that there are some loads on the market that claim to be electronic loads specifically for LED power supply testing. In fact, they are general electronic loads in a different appearance. Moreover, they are usually modified electronic loads whose bandwidth and sampling rate do not meet the test requirements. They do not improve their own bandwidth, because bandwidth technology is the core technology of the load and is closely related to cost, so it is difficult to improve. They are often improved through three ways to make the current data more stable, but also more unreliable.

　　The simplest way is to increase the filtering intensity and force the data to be stable. Simply using this method can easily lead to misjudgment and cause quality accidents.

　　Adjust the voltage feedback loop and filter the voltage feedback signal strongly to reduce the current oscillation amplitude. This method goes the other way and further reduces the load bandwidth, turning the non-oscillation and large oscillation situations into smaller oscillations.

　　Increasing the capacitance inside the load can suppress the occurrence or amplitude of oscillation, but the measured current ripple will be seriously smaller than the actual ripple, but it is very helpful for testing the DC operating point. However, because the rated working voltage of the load is generally high, the price and size of the high-voltage capacitor are very serious problems, so it is difficult to increase it to the ideal state, and it is often used in combination with the second method. Another problem is that in this case, it often uses relatively cheap high-voltage electrolytic capacitors, which will bring many parasitic problems.

　　This article comprehensively introduces the principle of electronic load, especially focuses on the misunderstandings of electronic load in the process of LED measurement. In addition, this article also proposes some feasible solutions to obtain more stable current data. I hope you can gain something after reading this article.