Again: Problems with Buck design LNK306

I'm not an expert or anything with these Linkswitch devices and have an issue of my own but I noticed that your output capacitance is over 100uF. Using the LNK306 you'll probably want to add a capacitor across Rfb of 1uF to 10uF for softstart to keep the LNK306 from restarting. That might be causing the drop out issue you're seeing when loaded.

Hi, thanks for the response.

I put a 10uF capacitor in parallel with RFB, the same result. I noticed about the soft restar after I bught my componentes. 

Even I change the output capacitor to a tantalium  100uF.  I got the same result.

According with the PDF, below 680 uH are not recommended. But PI Expert recommend me a 470uH. Anyway I put a 1mH TH coil, the only difference was a noise.

My capacitor is a 220uF becuase accordinng with the pi excel, I need a 70mOhm capacitor. Is not easy to find it in digikey.

May be my requirement is impossible to achieve with this product. I need a non-isolated power. since  I am using it in a dimmer.

May be I am doing a big stupid, and the staff doenst  want to help me.

Anyway, I will wait unitl monday. if I cannot find a solution . i have to look for another brand, make a new PCB design, new tests and tell my boss I cannot start the production the next month.

thanks.

What about that RF2? Is that really 100 Ohms? Seems like that would be too high of a limit on the LNK306.

Yeah, I feel your pain. These Linkswitch devices can be deceptive in simplicity. Good luck.

I have two kind of RF2 resistor, I tried with both of them, the same result.  I measured with the tester to make sure they are 100 Ohm resistor.

By the way. I am watching you graphs. You have a spike about 40 Volts? 

I  the same situation I ahve, except my voltage never drop to zero.

what happens when you put a load to the system? 

Yeah, I was wondering why the 100 Ohms. Have you tried with RF2 shorted? You have the 8.2 Ohm already so not sure of the purpose of the 100 Ohm. Seems like it might starve the LNK306 of voltage or current.

Yes, I had a sample that showed some pretty good spikes, about 2x my desired output voltage, but that might have been a disconnected load issue. After tweaking around a bit and ended up with the cleaner looking output. I did a lot of messing around with the circuit to fix. I wanted to confirm nothing else was introducing an issue so I built my second board with just the supply and loaded it with simple LED. Got the same 24V output that pulses On and Off. Its much faster than a restart but not in the kHz so I'm really wondering what's happening.

Wel, the 100 ohm resistor flameproof, is the input stage I choose form the AN37. I guess there are 4 to choose.

May be I can try put a 8,2 ohm resistor instead of 100 ohm resistor.

By the way, Your output looks like there is a short circuit on the output. I mean, when there is an over current, the output drop to zero, right?

Hmm...Okay. What about the 8.2 Ohm resistor already in your line feed front end. Try doing a test without the 100 Ohm in there. Just make is a short and see how that affects your circuit performance. I'll review AN-37 and see where they mention the additional input resistance.

Yeah, turned out I was doing something dumb. On my scope I have to have the signals within the screen area or the math calculations won't calculate properly. So what looked like the shorting signal was actually just my scope calculating zeroes. Once I zoomed out and had all the signals in the screen area all was good.

Hi,

1. When measuring output ripple especially with a non isolated design, it is important to ensure that the ripple that you are measuring is real and not common mode noise. Easiest way to verify is to connect the oscilloscope probe to the same point where the GND wire of the oscilloscope is connected. since the probe and the GND wire connected to the same point is like GND to GND measurement and as such you should not be measuring anything (0 volt) If you still see any spikes and high frequency waveforms that suggests that your test setup is picking up noise and you will likely need to use isolated input supply to the unit under test to prevent common mode noise currents slowing in your test setup.

Another possible source of sharp spikes at the output is excess capacitance in the inductor. The inductor that you are using for this design, please confirm that it is designed for high voltage application.

2. You are using the standard LNK-TN schematic. In this schematic, it is asumed that the voltage drop across DFB and DFW is the same and hence with the schematic that you have used for your design, it is assumed that the voltage drop across CFB follows the output voltage. It is important that the routing used for the circuit follows our guidelines. It is assumed that there are no significant loop inductances contributed by the layout. If in doubt, I suggest you use one of our many example designs for your layout or even modify one of our designs by substituting your components and keeping the layout.

3. The feedback scheme that you have selected may not be suitable for a 3.3V design if you need precision output. To achieve precision output, it would be best to use isolated feedback using an optocoupler and LMV431 IC.

Regards

PI-Sarek

Hello.

Thank you so much for your response.

The inductor I am using:

http://www.bourns.com/docs/Product-Datasheets/SRR1005.pdf Is the 470uH inductor. According with the information, that inductor is suitable for power supplies.

I attached the layout of the Top botton of the PCB, in that side I put all componentes for the power supply.

I tried to make the PCB design similar to the PDF of LNK302-306. Since my components are big, and my space is very reduce, I tried  to fit them the best I can.

I am going to make the test of the commom mode. I am going to solder the optocoupler.

Right now I have the 817A. Sadly I dont have the LMV431. If I dont use this IC, what would be the tolerance of the output? The range of all my componente are 2.8V-3.6V. Do yuo think I still need that IC? Or only with the zener will be enough?

Thanks. I will be back with you after my test.

Thanks!

I have some problems designing with lm306 220AC to 3.3DC. Is there an example scheme for these voltages?

Hi Sajra,

Thank you for reaching out to us through this page. We are willing to help you with your concern but if you may, please kindly create new ticket for this one. If you think that this thread is relevant to your present inquiry, you may include this link in your new post.

We'll wait for your new ticket.

Thank you and Best Regards,

PI-Lupin the Third