LNK304 failure.

Posted by: sridar on 11/26/2008

Dear team, We are using flyback convertor with LNK304 and facing lot of failures in the field but could not reveal the occurance of failure, drain source pin has totaly shorted and primarry inrush current limiting resistor being short in the pcbs. It would be nice if any one provide some guide line to reveal the failure mode. Note: No events were noticed during testing by exceeding the vds rating of the device.

Comments

Hello sridar,

From start we should ignore over-current failures, the circuit is well protected. This leaves us with over-voltage failures. Without access to your schematic, I will assume for now you have the right clamp circuit to cut any leakage related voltage spikes.
There are two main mechanisms for over-voltage events big enough to destroy the switch. One comes from AC grid surge events (thunderbolts, grid failures, etc). A possible way to identify and fix this problem is to add a TVS (MOV) selected for your input voltage range. If from now on the field failures are showing you the fuse-resistor destroyed or no failures at all, you know the problem came from bad grid quality.
The second over-voltage event could be generated by ESD. A quick fix is to use triple insulated wire for the secondary winding of your transformer with flying leads in order to move the solder point (the exposed conductive point) away from the transformer (or any conductive part: copper, magnetic core, etc), at least 15mm. The triple isolation is capable to withstand ESD events as high as 12-15kV. Adding a spark-gap PCB circuit 10mm apart will provide a safe discharge path from secondary side to earth (trough AC input).
You can check also this application note:
http://www.powerint.com/sites/default/files/product-docs/an20.pdf
Cheers,
PI_Crusher

Thanks mr Crusher, We do have a MOV in the circuit which is located in very first in the circuit and we have done the testing SURGE testing, the module got passed and no failures occurs while on testings.even we have gone upto 8KV. Herewith am attaching the schematic of the power supply pls review the same provide your valuable inputs. Awating for your reply. Thanks Sridar

Dear Mr.Crusher, Is there any possibility that parasitic capacitance between LED Anode Pin and Photo Transistor collector pin of Opto (feedback ckt)causes this failure? To simulate this I just added 0.01 uF 1000V across these 2 pins and I could see input fuse resistor(burnt) and LNK304 failure. Please comment on this. Thank you.