Solution Finder Get Tech Support

DER-289 - short circuit protection

Posted by: Robil on 02/26/2012

Hello everybody,

I have tried DER-289 circuit modified for 22V, 480mA (for wide range of amount of LEDs possibility) with EF20 core and LNK406 (spreadsheet attached).
Modification of circuit: R24=56,9K, R22=39,0K for required current, C7=22uF/50V for better bias voltage filtration, C10+C11=1uF/400V to minimize AC line ripple, C8=100nF/630V for better PF with less power (10,6W).
Everything worked very well but short circuit protection works only during startup (auto-restart mode).In case of short circuit during normal operation, transformer is emitting audio noise (it's normal - core saturation) and DC bias voltage doesn't go below 6,4V. To force auto-restart mode with R17=174K, it is needed max. 5.4V (in case of short circuit AC envelope at bias winding is about 3,0V but "hard" spikes charging bias bulk capacitor C7 to more than 6.4V).
Changing minimum current feedback resistor R17 to 220K didn't help.
I also tried to supply R17 from "independent" zener circuit as in DER-263 (R7,R10,VR5) but it didn't help.Unconnection of whole constant current opto circuit (daughter board) with R17 connected to V bias didn't help.It follows that the protection does not work even with reference design feedback circuit.

BTW: I think there is a mistake in daughter board layout in Q1 position does not match with schematic sheet - inverted collector with emitter.

Thank for help.



Attachment Size
Trafo_230V_22V_0.48A_LNK406.pxlph 39.65 KB


Submitted by PI-Jono on 02/29/2012

Hi Robil,
Very good observation, I think there is a leakage in Q3 during short circuit condition even you implement the circuit as DER-263.

Please refer to the attachment for the solution. Kindly verify if it solves your problem.

By the way, Q3 is used for tight CC down to 3V output, if your application is not very low, you can bypass Q3.


Submitted by Robil on 03/01/2012

In reply to by enkavak

Hi Jono,

Thanks for answer. I tried exactly the same circuit as in your attachment several days ago but the problem is not in leakage current through B-E of Q3 but in high peaks of bias winding during short circuit condition.Bias DC voltage doesn't drop below cca 8,8V and voltage on the zener diode below cca 6,0V - it is still too much for protection activation. These high peaks can't remove by any RC clamp circuit connected to the bias rectifier.
As I wrote in previous post - unconnection of whole constant current opto circuit (daughter board) with R17 (220K) connected to V bias didn't help (wihout any zener circuit). It follows that the protection does not work even with reference design feedback circuit. Please check attached plots from oscilloscope.

Once more thanks for help.


Submitted by PI-Jono on 03/01/2012

Hi Robil,
Very nice, you attached waveforms so that we can analyze it clearly. Normally during output short, the bias winding voltage will collapse since it is operating in flyback toplogy. Though it will auto-restart, it is not enough to raise the bias winding voltage.
Assuming the winding termination(flyback) is correct for the bias winding then you need to:
1. Improve the magnetic coupling for bias winding to the output winding. Wind the bias winding close as possible to your output winding, this will significantly reduce the unnecessary spikes in the AC bias winding.
2. Make sure your series resistor (R3) in the bias winding is not shorted. Start with 150ohms then increase its value if required.
3. Add a ferrite bead in series with R3 to totally block the unnecessary ringing.

Hope this solves the problem during short.