LNK306 High voltage issue
Hi,
I did a power supply design with the LNK306GN. The requirements was universal input voltages and 12V output @ 340mA. I use the formulas in the application notes (which I've used before) and used PI Expert Suite (design sheet, also used it for a previous design) to do the calculations and to verify my calculations. Everything checked out. I built the supply (proptype boards) and it worked fine at 220VAC input. The application switches a load switch, with coils that require say 200mA for 100ms, to switch from either the open or close posisition.
I then had to test it down to 50VAC, where it still worked fine (test - switching of the load switch and checking supply output voltages etc). Then I increased the input voltage to 250VAC, the supply worked fine. Then I increased the input voltage to 260VAC. When I switch the load switch, the supply resets.
I have tried every suggestion there is in the application notes etc. But nothing makes this supply work at 260VAC+ at 200mA. The supply is suppose to qork to 275VAC. And it was designed for a 340mA load. I dont know if the coils has an effect on the supply and whether I can compensate for it.
I have tried increasing the startup capacitor. I have tried increasing inductance, feedback capacitor. Nothing worked or even improved the supply. Please help me with some suggestions, I've at this supply for 3days now and this is last resort.
Specs
COnverter: Buck Converter
Input: 85VAC - 265VAC
Output: 12VDC @ 340mA (200mA for load switch and 140mA for other connected devices)
Rectification: Half-wave rectification.
Input capciticance 66 uF
Output Capacitiance 100uF
Startup Cpacitor 1uF
Feedback Capacitor 10uF
Inductor 1100 uH 1A
The output voltage is 11.8V and at lower voltages (60VAC - 259VAC) I've pushed the supply to 400mA (just to check). But at 260VAC+ it just does not want to deliver. Please help.
Kind regards,
Quintin.
Comments
Hi,
Thank you for the reply. I will check the switching waveform. I currently have a BYV27 with 25ns reverse recovery as catch diode, but I will check the waveform anyway. Q.
Hi,
I think I'm having the same problem now - excessive leading edge spike current.
Used output Schottky diode, tight PCB traces, layered transformer windings, sunbbers across diodes, etc. Even tried to add inductor/flter in series with the drain to limit the surge.
I've ran out of ideas to address this.
Could you share some of your thoughts on the solution to this problem?
Thank you.
Kind regards,
CSK
Hi,
I did get the problem fixed. The issue at the time was that I was supplied with the wrong diode. I should have used a BYV26C specifically the C because of its voltage rating of 600V and it's Trr = 25ns. I used a BYV27-200 (a lot slower) and a bit obvious but the reverse recovery of the BYV27-200 is 200V. Definitely not high enough. So I would suggest, unless you are using the correct diode, to check the diode's specification that you are using.
Using the correct diode, or even one with at least a reverse recovery voltage of 600V worked immediately. Hope that helps.
Regars,
Q.
Hi,
Thanks for your reply. I thought I've sent it out earlier but evidently not.
I'm using a Schottkt, so I don't have the trr problem. Infact, putting a snubber across this diode helped to cut back the ringing but not the spike.
Disconnecting the secondary, I traced the start up Ids and the same spike appeared, so it' not the diode recovery problem.
I'm now tracing this to the transformer, after isolating all the possible branches that could contribute to this spike.
Question:
a) is whether the transformer shielding and cancellation winding increases the winding capacitance. Logically, I think it does and that could be my problem after all. As of now, I'm "guessing" the winding capacitance is some 30pF, ringing at about 600kHz.
b) is there any other way to measure this capacitance, other than looking out for the winding resonance?
Cheers,
SK
It sounds like you may have leading edge drain current spike at turn-on. Check switching current waveform for excessive leading edge spike. You may need to use faster catch diode.