How to test the dimming linearity of LED drivers using an oscilloscope, AFG and multimeter?

Useful Information Sharing | How to use an oscilloscope, AFG and multimeter to test the dimming linearity of an LED driver?

With the development of LED lamp technology, digital dimming technology has made great progress in recent years compared to traditional analog dimming technology. Nowadays, there are thousands of dimming products to choose from in the lighting market. We need to consider these factors when choosing dimming drivers. Dimming smoothness, dimming depth, and whether there is perceptible flicker and ripple during the dimming process.

In order to achieve ultra-fine smoothness of dimming output, we must first understand the difference between each dimming level. The smaller the difference between each dimming level, the smoother the dimming. In this way, stepless dimming can be achieved throughout the dimming process. As shown in the figure below.

The case in this article is that a customer needs to quickly measure the linearity and smoothness of the output current during PWM dimming, and then use an integrating sphere to test the brightness linearity. The device under test is a highly integrated dual-mode BUCK with constant voltage or constant current output and a carrier two-bus chip with non-polarity access. At the same time, the customer is worried about whether this dimming chip with bus function will affect the LED dimming during bus communication. The switching frequency of the Buck part is as high as 1Mhz, which also brings considerable challenges to current measurement. This article chooses to use the TCP0030A probe, which can reach a bandwidth of 120Mhz, which is very suitable for measuring high-speed current signals.

Let's take these questions and start testing and verification together.

Purpose of the test:

1. Use the signal generated by AFG31252 for PWM dimming and use DMM6500 to scan the output current linearity.

2. Use a current probe to measure the inductor current to check whether the inductor is saturated during dimming so as to select a suitable inductor.

3. Observe whether the bus communication affects the dimming and causes the LED to flicker.

The test environment has been set up:

Test ideas

Let's test question 1 first:

Use AFG to generate PWM to adjust the duty cycle. Here, the multimeter is not connected in series to the loop. Instead, the DM6500 is used to measure the voltage of the current-sense resistor and calculate the average current of the light string. The following chart can be obtained:

From this point of view, the linearity is still very good within the full duty cycle dimming range. In each PWM cycle driven by the AFG, the SW can respond quickly and the current reaches the preset value without overcharging.

We then verified whether the LED current would be affected during bus communication. Thanks to the 120Mhz speed of TCP0030A, we can observe the very small current noise that may exist. This provides very clear measurement support for finding some inexplicable LED flickering, dimming and other abnormal phenomena.

In fact, it can be seen from the figure above that during the off cycle of the dimming PWM signal, the carrier waveform on the bus does not produce any spikes in the LED current. Let's take a look at the situation during the on cycle of EN.

It can be seen that when the bus communication waveform changes rapidly, the light string current still does not have any glitches or overshoots.

We then move the current probe to the inductor and observe the inductor current.

It can be seen that the current reaches a maximum of about 900mA during the on cycle. Thanks to the storage depth of TBS2000 up to 20M, the waveform is still very clear after zooming, which shows that the current selected inductor has not experienced saturation of the current tail. Of course, if you want to evaluate the availability of the inductor more carefully, you need to do more tests on the inductor current under various switching modes, as shown in the previous article. The rise time of this current is only 640ns! This is a constant current LED driver chip with a switching speed of 1Mhz. Low-speed current probes cannot measure it. For the measurement of this signal, this TCP0030A current probe with a speed of up to 120Mhz is almost the only choice.

The customer then used the serial port to transmit the MODBUS protocol on the second bus, and used the MCU to control the brightness of the light string according to the frequency given by the AFG, and quickly completed the design of a lighting solution.