Non-Isolated LNK305 Buck Converter Looses Regulation Above 220VAC

Posted by: alexanderrk on 04/20/2012

I am working on a non-isolated buck converter design using the LNK305 controller IC, which is used to drive a brushed DC motor. Currently I am experiencing a loss of regulation when the input voltage exceeds 220VAC. My power supply requirements are: 17.5VDC, Imin = .120A, and Imax = .150A. The attached waveforms show the top green waveform, output voltage, has approximately 2.40V of ripple at 220VAC input voltage. The switching signal is blue and the inductor current is yellow. As the input voltage is increased to 264VAC the voltage ripple increases such that the supply no longer functions as intended. From figure 1(a) of AN-37, my power supply is configued as follows: CIN1 = 10uF, LIN=1mH, CIN2 = 10uF, CBP=0.1uF, RFB= 18.2k, RBIAS= 2k, CFB=10uF, L=1.0mH, Co=330uF, Css=1uF. I was able to change the output inductor value L to 3.3mH, which is much higher than the calculated value, and achieve a better regulation, power supply droops by approximately 1VDC. However, I would like to completely understand what is taking place prior to making a design change.

Files

TEK00014.png (329.12 KB)

Comments

Hi,
Looked at the plot. Seems like classic oscillation behavior. Looks like you have an stability issue (low phase margin) at high line. I assume you don't see this problem at low lines.
Please confirm this.

rgds

You are correct, I do not see this occuring below 220VAC. It seems once the supply passes this mark it enters another mode of operation. I also noted that the frequency of the ripple corresponds to the 60Hz line frequency. I evaluated differrent feedback capacitor values in the design as well as switching to the LNK306 controller, which both exhibited the same behavior as previously stated. I did some additional investigation in terms of PWB layout and found ground plane directly below the feedback circuitry of the power supply. I found that if you lifted the feedback resistor away from the PWB the circuit would function as intended with the correct inductor value. I removed the ground plane from the feedback circuitry area and it is now working as expected. I now have a 1.5mH output inductor,which is the value calculated by the PI expert software. If you could elaborate on this a little for me that would be appreciated, I would like to fully understand exactly what may be occuring.

Hi,
Seems like a layout problem. The feedback trace must be away from any switching node, or else coupling (through board parasitic) from noise source (especially at high line and full load)would cause the PSU to behave in a erratic manner. Best thing is to guard ring the Feedback trace.

rgds..

Thanks, placing a guard ring around this area does seem to correct the problem and now I have a decent understanding as to what was occuring with the ground plane in the area of the feedback traces.