LNK306D transient load response
I have been having some transient load response issues. I changed the feedback configuration to a Zener/Opto, went to CCM and increased the output bulk capacitor. This addressed all but my worst case transient load condition. In desperation I placed a small capacitor (.001uF) across the BP and FB pins. This fixed my problem but I'm not exactly sure why it works and I am concerned about causing some other problem. Can anyone give me some feedback ?
Comments
Thank you very much for your feedback. I am using a FOD817D which has a minimum CTR of 300% and I am also pre-biasing the Zener. The larger inductor has a saturation current of 1 amp. Can you tell me why adding the capacitor helped ? One thing I did notice while monitoring the drain current was that after adding the cap, the MOSFET seemed to switch on more often i.e. the time between switching cycles was shorter. Before adding the cap it seemed as though the time between switching cycles was longer and very constant. When the worst case transient load condition occurred, the output voltage drooped (about 20%). I am assuming that the longer time between switching cycles would occasionally keep the output capacitor from charging fast enough to prevent auto-restart. Changing the output capacitor to a larger value fixed the drooping problem. I have not looked at the inductor current yet but do you think there might be an issue with saturation ?
A 1A inductor should be sufficient for your application. The maximum current limit (when di/dt is 610mA/us) is 647mA. If you happen to have a small inductor you may have a higher peak current, but if that's the case you can always check Figure 13 on the datasheet to see your actual di/dt. Anyway I wouldn't expect an issue on this front.
The experience I personally had while doing this type of feedback is that without the capacitor there was a tendency for pulse grouping, which was solved with the capacitor. This extra capacitor would spread out the switching cycles, which meant that the IC was reacting quicker, but, on average, since the we are sending the same amount of power, with the same peak current, and the same inductance, we will have the exact same average switching frequency.
Thanks again for all your help ! I can finally put this design to bed !
Hi
A 1nF capacitor between FB and BP is not going to cause you any trouble. I don't know what type of optocoupler you are using, but a 300% CTR should help you as well.
A resistor (1K) in parallel with the optocoupler diode to pre-bias the zener will also help a bit.