LYTSwitch-2 in smart lamp application

6 posts / 0 new


I have designed a smart lamp (RGB, warm white, cool white, dimming) with LYT2004. All seems to be fine but I get a flicker at only some dimming levels of the white light. 

Output voltage = 26 V

Current limit = 0.27 A

PWM = ~750 Hz

When looking at the output voltage using oscilloscope I get somewhat different looks for different brightness level ranges. In one of the ranges I get the flicker. It happens at around 20% duty cycle. The output voltage jumps up for a few PWM pulse and then gets back to where it needs to be. This seems to be happening at random. At 100% the output voltage line is nice and flat.

I have attached some plots. 

Any help, suggestions welcome. I am happy to provide more information if needed.



첨부 파일Size
20percent_dutycycle1.56 MB
50percent_duty.png2 MB

Anything, anyone???

Does it happen if it's a constant load and not PWM?The drain waveform looks like the device operates in continuous mode. Not sure if it's due to the PWM operation or the transformer was designed in continuous.

Please ensure that you operate in discontinuous mode. There could be other factors that might cause this but essentially the feedback has to sense the output voltage information through the bias windings.

Aside from checking that the design is discontinuous, adding a series resistance to the bias winding and secondary winding may help. Try adding between 2 to 4 ohms in series with the windings.



Sorry, I may have confused the matter a little. The drain in the previously attached pics, is the drain of the switch that turns the LEDs on and off. It is not the Drain on the LYT2004K device. 

There is no flicket at constant load. In fact, there is no flicker at most of the PWM duty cycle settings. The flicker only happens at certain PWM duty cycle range. Also, this range moves if I change the frequency of the PWM.

I have designed the circuit for KP well above 1.3, which, as I understand, ensures dicontinuous opration.

I will try the series resistances.

Thanks for your suggestions.


I tried some resistors as you suggested. This is what I found. Adding a resistor to the bias winding did not help. Adding a resistor to the secondary winding certainly helped. There is no fluctuations of the output voltage, which was there before. You mentioned that I should try 2-4Ohm. Is there a reason you did not reccomend going below 2 Ohms?

The adverse effect is that the output voltage is now a lot lower that what I desinged for, and the regulator is not really going into CC. Is there a procedure for selecting feedback resistors including a series resistance in the secondary? Or I should simply design for a higher voltage?


If <2 ohms works then it should be fine.  If the output voltage drops as a result, then please design for a higher voltage as the series resistor is not incorporated on PIXLS calculation.