High ambient design
I have an application where the maximum ambient can be 105C. The requirements are 190V to 265V 50Hz, non-isolated 40V 1W. At the moment I am using an LNK306 but on occassions it shuts down presumably due to thermal protection. Is the LNK306 the best choice for this application? Are there parts with higher thermal shutdown temperatures?
Brian
Comments
Ok, done some measurements on actual performance.
Input power 2.53W (7.4mA @ 342V)
Output power 2.15W (47mA @ 45.7V)
These values do shift around a bit but it looks like 84% efficiency.
Original free wheel diode was ES1J but also tried a 30nsec diode with no significant improvement.
As for thermal margin, I reckon 105C is where it drops out. It will actually run for about an hour or two before shutting down.
Schematic is attached, only thing to mention is that I switch the output between two levels (~10V and ~45V) hence the two zeners and the FETS.
Efficiency numbers look good. Do you have a current probe? Can you monitor the drain pin current and post? Would you also send the PCB layout. I wonder if the area is too small for the ~300 mW dissipation in the device.
Cheers PI-Chekov
Hello Brian,
we do have other design that operate at such high ambient but they do require special consideration.
First to answer your question, no there's not a better part in terms of thermal shutdown level, all our products have a minimum thermal shutdown in the 135 C range.
However if your ambient is 105C and the part is reaching thermal shutdown (135C) processing only 1 W then there is room for improvement.
Firstly was there a specific reason why you selected the LNK306? For a 1 W output at 40 V (output current of 25 mA), the LNK302 or LNK304 would be able to deliver that output current. As a general rule oversizing by more than one family member device size reduces efficiency due to increased MOSFET capacitance based switching losses (especially true here because your design is high line input only).
It would be worth a quick review of your schematic. I'm assuming a buck topology so the free-wheeling diode has a big impact on switching losses.
Would you please post your schematic and also your measured efficiency, the efficiency will give a clue - if it's low then we can look at improving the losses and lowering temperature rise.
Finally check the thermal margin you have in your design by increasing the ambient above 105C very slowly. The idea is to allow the LinSwitch-TN part to stabilize at each temperature level. Start with 110C and then go in steps of 5C.
Cheers
PI-Chekov