Hi everyone,
I’m working on an SMPS using LNK3694 (900V internal MOSFET).
Design details:

• Topology: Flyback
• Outputs:
  • 5V @ 100 mA
  • 12V @ 50 mA
• Input: Universal AC (rectified mains)
• IC: LNK3694 (Power Integrations)

Problem:
The SMPS works perfectly under normal conditions. However, when the input line voltage goes above ~350V, the Line Overvoltage (OV) protection kicks in, and the device goes into auto-restart mode. Once the input voltage drops below ~350V again, the SMPS recovers and outputs return to normal.

Expectation:
From my understanding of the datasheet:

• The internal MOSFET is rated at 900V
• The device should survive high input voltage
• I expected the supply to continue operating, not shut down completely

I understand protection triggering is intentional, but in my case the supply stops delivering output, which is not acceptable for the application.

Questions:

1. Is this behavior expected for LNK3694 at high line input?
2. Is the Line OV threshold fixed, or influenced by external components (e.g. BP pin, transformer design, leakage inductance)?
3. Is there a recommended way to allow operation at higher line voltage without entering auto-restart?
4. Has anyone successfully designed LNK3694 to operate above this input level continuously?

This is my schematic.

 

input line voltage goes above ~350V

Why would you put 350V on a 240V input power supply? Are you trying to break something? Sorry that it did not break, better luck next time.

Is this behavior expected for LNK3694 at high line input?

Yes

Is the Line OV threshold fixed

Yes

Is there a recommended way to allow operation at higher line voltage

yes

I expected the supply to continue operating, not shut down completely

I expected you to blow up the IC and maybe the diodes. This function has saved you many times.
-----------------------------------------
The IC is not looking at the line voltage. It is looking at the voltage across the MOSFET. Probably the voltage pointed at by the left red arrow. Probably not the right arrow. The green line is the input voltage.

There are many things you can do. You did not show all of the schematic so I will guess what you are doing.
1a&b- The diode R & C is there to limit how high the voltage rings (see left red arrow) you can try reducing the value of R. You can put a high voltage Zener across the resistor to reduce the ring. The Zener must not turn on during the voltage of the right red arrow!
2- The ringing is caused by leakage inductance. You can remake the transformer with better primary to secondary coupling.
3- We do not know how much more input voltage you need??? Change the turns ratio. During the output part of the cycle the voltage above the green line, through the turn ratio, appears on the output. We need to reduce the voltage above the green line. (How?) If we added more turns to the primary the voltage would get larger. If we reduce the turns, it will take less primary voltage to get the same secondary turns. You could change the secondary if you want.

Hope this helps.
RonS

 

Showing us a partial schematic is useless indeed!!
My first advice is to read all of the specifications. There is probably a maximum input voltage limit. Beyond that limit specified performance is not promised.

 

There does exist a very reliable but not efficient device, offered as a "constant voltage transformer.
Probably you can achieve a similar result with an isolation transformer that the core saturates before the input gets that high.

 

There do exist commercially available over-voltage protection devices that would protect the supply from that 350 volts. The best ones fire a gas tube shunt that removes the spike by short circuiting. Then a series circuit breaker opens and the short is thus removed.
HOW are you getting that 350 volts, anyway?? That is not a common problem.