LNK304 -> Feedback CAP and tan delta
Hi,
we are using many years the LNK304 in our designs. But now we have a new design and we take the same circuit but with other component size.
Our problem is the C3 in the feedback loop, we use years a 1µF/50V/X7R/Size 1206 without any problems. But now we change the cap to 1µF/35V/X5R/Size 0402 and the regulation did not work. -> output voltage is too much.
For better understanding for C3 look here: https://ac-dc.power.com/sites/default/files/product_document/data_sheet/linkswitch-tn_family_datasheet.pdf -> page 4.
We find out that the tan delta of the caps is different, with a small tan delta (0.01 and smaller) the chip works correct, with a tan delta bigger than 0.03 the regulation did not work correct. With increase of tan delta of C3 the output voltage increase.
Our minimum load is 5mA,so there is no special load with a resistor needed.
Here are some results with caps, messurement conditions of caps 1KHz:
1µF/50V/X7R/1206 -> 1,098µF / tan delta 0,01 -> output voltage 20,2V -> ok
1µF/50V/X7R/1206 -> 1,081µF / tan delta 0,009 -> output voltage 20,2V -> ok
1µF/50V/X7R/0805 -> 1,061µF / tan delta 0,015 -> output voltage 20,2V -> ok
1µF/50V/X5R/0402 -> 1,030µF / tan delta 0,036 -> output voltage 30V -> not ok
1µF/50V/X5R/0402 -> 1,045µF / tan delta 0,036 -> output voltage 30V -> not ok
1µF/35V/X5R/0402 -> 1,100µF / tan delta 0,065 -> output voltage 38V -> not ok
1µF/35V/X5R/0402 -> 1,158µF / tan delta 0,069 -> output voltage 40V -> not ok
Have anyone any answers for that to understand that?
What is the correct tan delta which is needed?
thanks for your help and kind regards
behrthom
Comments
Hi behrthom,
Thanks for choosing PI. Are you now okay? Will you still be needing our support on this so I can close the ticket if not?
Best Regards,
Lupin the Third
Hi behrthom,
The capacitance of ceramic capacitors decreases when the applied voltage increases. This behavior (% capacitance vs voltage) is worse for smaller sized ceramic capacitors. I suggest you check with the capacitor manufacturer for this specification.
I can solve the problem, it is not the tan delta, it is the DC Bias graph of the cap, at 20V the cap have only a value of 150nF, that is to small for the feed back loop. thanks a lot to codico FAE.
Kind regards
behrthom