ソリューション ファインダー 技術サポート

Auto restart under normal conditions on top256

Posted by: aman_geosys on

Hi,
I designed a flyback converter using top256 with reference from PI Expert for 24V, 2A output. When testing at 0.5A load, the power supply goes in auto-restart mode within a minute.
For testing I used a variac and tested on all voltages but behaviour is same. Earlier it would recover from the auto-restart but now the device just keeps restarting until I unload it.
All related calculations are verified in PI Expert but I don't understand what is causing the device to go under restart condition.
I've attached the schematic followed.
Please share your thoughts on what the problem may be.
Thank You

Files

Attachment サイズ
schematic24V.png 80.97 KB

コメント

Submitted by PI-csbabu on 01/03/2020

Hi Aman,

Does the unit regulate output voltage at no-load/min load? Monitor the control pin voltage and restart time duration. If control pin voltage initially stays above 5.8V and dropping below 4.8V after startup means there may open feedback loop.
Please share the design calculation sheet.

Submitted by aman_geosys on 01/03/2020

Hi,
Thanks for the reply.
Yes the unit regulates under no load/light load, also after some debugging I've got it to work with around 1A load. As I increase the load(1.2-1.5A), the device will stay in regulation for small time and then keep auto restarting till the load is reduced to 1-1.2A . I suspected it was a current overload fault so I reduced the resistance at X pin from 12.4K to 11K(which should be more than sufficient for 2A load). Result, supply now gives max 1.7A but still auto-restarts after a few seconds.
A thing I noticed is that the device doesn't count 16 times of auto-restart cycles, it turn on after I reduce the load.
I've attached the design sheet.

Attachment サイズ
TOPSwitch-HX_PIDesign24V.pdf 93.76 KB
Submitted by aman_geosys on 01/04/2020

With further testing I've gathered some observations.
I removed the current limiting resistor and shorted it to source pin to disable the overload protection.
Now although I get the 2A (and higher) current but the supply would still enter auto-restart after 30-60secs. Without any change the supply would turn on after 8-16 restart attempts, then regulate for 30secs and auto-restart and so on.
It cannot be over or under voltage as I've maintained it with Variac and I've monitored the waveform on drain pin too.

Submitted by aman_geosys on 01/05/2020

Hi,
Can anyone please share their inputs on possible solutions?

Submitted by PI-csbabu on 01/06/2020

Hi Aman,

You are correct, design is hitting the peak current limit and restarting. Remove the transformer and measure the magnetizing inductance. Try reducing the inductance to around 720 uH. Please measure drain current before and after restart at load and share waveform. In the schematic please change the capacitor C5 to higher rated voltage ( 10V to 16V).

Submitted by aman_geosys on 01/07/2020

Hi PI-csbabu,
Thanks for your suggestions.
I actually found out that the culprit was using the ultra-fast diode. As a diode with reverse voltage rating of >130V was required I first used UF5408 which gave the auto-restart problem, then I replaced it with 2 100V SR5A schottky diodes in series and the problem was solved.
I am not sure why this is the case. Are ultra-fast diodes not fast enough to be used with TOPswich family? or is it related to the transformer inductance (831uH as measured).
If ultra-fast are not suitable maybe PIExpert should suggest that a schottky diode be used instead of an ultra-fast.

Submitted by PI-csbabu on 01/07/2020

Hi Aman,

The diode selection is based on required DC blocking voltage/ current. Schottky diodes have lower blocking voltage compared to fast recovery diodes. Eventhough Schottky didoes has lower VF, with increase in voltage rating and temperature, leakage current is high. UF5408 is 1000V rated / VF is 1.7V, so the losses is high. I recommend to use MUR series diode where VF is 0.7. So for a given turns ratio, reflected voltage is low. Check with low Lmag value.