Solution Finder Get Tech Support

LNK306 voltage ripple problem

Posted by: InqbateTadej on

Hi!

I am designing an AC/DC converter with LNK306 - buck-boost with optocoupler feedback. See the schematic and PIXls designer file. I am getting the attached output voltage waveform. It looks like that LNK makes 3 switching cycles, with voltage still rising after the last switch, causing LNK to skip a lot of cycles. Or is this auto-restart or over-current protection. If I increase the load I get similar waveform with increased sawtooth frequency. Can you point me where in the circuit should I look for an error?

I suspect two things:

- ferrite (feedback should be before the ferrite of the PI filter) - I'll try to replace ferrite with 0-ohm resistor

- optocoupler

The 180uF cap is Al org poly, with 30mohm ESR and 47uF is MLCC.

Thank you for your help,

Tadej

Files
design.pixls (17.5 KB)
schematic.png (18.54 KB)

Comments

Submitted by InqbateTadej on 07/13/2015

I've shorted ferrite and got slightly lower ripple, but I think that ripple should be even smaller.

I've also made measurements with light load and half load for different configurations. Please see the attached waveforms.

Submitted by PI-KSAN on 08/04/2015

At what Input voltage are you seeing this problem?

Regards,

PI-KSAN

 

Submitted by InqbateTadej on 08/04/2015

It was through the whole range from 50VDC to 250VAC.

I've solved the problem. Changed Zener voltage from 1.8V to 2.4V and removed R2. Now the ripple is about 80mV @ 0-200mA and 3.3V.

Looks like you have to choose the zener voltage based on your optocoupler CTR. I've moved design from protoboard to PCB and changed only opto, but I should have adjusted Zener voltage accordingly.

Kind regards,

Tadej

PS.: If anyone is interested I can post new scope screenshots.

Submitted by PI-KSAN on 08/06/2015

Thanks for posting us with the solution.