LNK454 - problem with flickering

11 posts / 0 new

Hello !

I have made LED driver which is based on LNK454D. I attached schematic, file with prepared calculation (from PIExpert) and Transformer's datasheet. I need to supply LED with forward voltage 12V and current 0.15A from 230Vac input voltage.

I have a problem with stable work of that PSU. When I connect 230Vac to the input of PSU, brightness of LED grow (about 1.5sec.), after that time the LED turn off for 1.5sec. After that 1.5sec turn off state, the PSU repeat hole cycle of hiccups.

I can't find the reason why my LED driver don't work stable (without hiccups).

I have measured  the voltage on BP pin, and there is a stable about 6V.

I have measured the waveform on FP pin and I haven't seen that level of signal exceed 290mV.

I have measured the current into the trafo's windings, and I have seen properly waveforms (for flyback topology).

I don't know where is mistake. Please help me to resolve my problem because I don't know if I need to change the driver.

With Best regards

Tomek S 

On the schematic, the drain and source pins are interchanged. Pls double check your connection of drain and source pins. 

Thx for your answere.

Yes I saw this mistake. I resolve this problem by cut a trace and provide a signals to dren and source by cable with diameter 1mm.

I think that it is good solution for repair schemat symbol mistake

Best regards,


So dou you mean it still doesnt work with the correct schematic?. Can you try reducing the output capacitor to 330 or 220uF? If this doesnt help, can you probe the voltage across C5 and attached the waveform.

Exactly it doesn't work with correct schematic. I soldered 2 pcs. of PSU and both doesn't work properly.

Tommorow I'll send you information about smaller capcitor and send waveforms.


Best regards



Reducing to 220uF and multiplying to 1000uF of capacitance didn't change the results.

I attached the waveforms from my PSU.

With best regards.




1. What load are you using, if not an led load pls use an LED load? 

2. Use 1k on R18 and stuffed R12 and D4.

3.The current sense R8 should be closed enough to the computed value on the PIXls which is 1.6R



Does the waveform you send the drain to source waveform or C5 waveform. If its across C5, it should be rectified and not switching.


Pls try the above suggestion and let me know of the results 


1. Load is a LED diode 12V


2. R18 and D4 was stuffed

3. I tried with 1.9 Ohm


I made all combinations with stuffing elements (my schemat is based on an50.pdf and other eval boards for -PL family)

and results is exactly the same. I didn't find the differences, the LED on the PSU output blink. 

The waveform if from C5. But this is not freqency of LNK switching. This is rectified AC 50Hz from wall. And when the level of xoltage drops, the LED is turn on, so I think that tis state should be all time. But I don't know why the PSU turn off after 750ms (properly) work?

Best regards




If youl already follow the PIXls values and the suggestions I gave you and problem still persist, I suggest you review carefully your layout, transformer,  and make sure no mistake in there. You can also try disabling the OVP first because maybe its injecting noise to the FB pin(whic is not normal if youre using the right zener and good layout). Have you tried changing the IC to a new one just in case the IC get damaged when the drain and source got reversed?

The problem have been resolved. Problem was in layout, FB signal wasn't delivered properly to IC.

Now everythink is ok. LED shines with constant brightness. So, schemat is properly.

Thanks very much for all replays.

With best regards

Tomasz Sobczuk 

Good to hear that.