What is the best way to test the load response of a DC DC buck converter? I thought that a power resistor to fully load the circuit would be the best thing to do. Any thoughts on that?
Specs:
synchronous buck converter
200kHz
VIN: 18-36V
VOUT: 3.3-12V
IOUT: 2A
Pmax: 10W
Hope this is enough information to help me, otherwise give me some time to include more information.
Thanks in advance.
You can use Burn Test for prolonged hours ( say 10-12 Hours ) on your buck converter and for this, connect 2A load as per your specs (LAMP, FAN or Power Resistor etc).
Very first, you have to ensure your PCB layout, Heat Sink, filter capacitors are as per recommended design of buck.
Buck convertors have become even more common so I rigged up a boost convertor with a 60V N chan powerfet ,a 60 V shottky and a 100microhenry inductor.On my test setup I used a total of 10 10microfarad 50V maruta ceremic caps , 5 on the boost input and 5 bon the boost output .The gate drive for my boost convertor was derived from the boost convertor output .Now I connect the boost output to the buck input to the positive of my bench supply .The bench supply is set to the input volts of the buck ,It could be 24V 32V etc.My special buck convertor is configured to draw a variable current from 0,1 to about 3.2 amp .I can now test a buck convertor of almost any output volts which makes a good burn test . When I do my tests the bench supply current draw represents the total losses in my test jig and the buck convertor under test .Because I know the jig losses reasonably well I can calc efficiency while I am doing the burn .My bench supply is 3 amp which is fine .Its feasible with this test method to use a smaller supply than the Buck convertors rating or test buck convertors larger than the bench supplies rating .The reason for this is that only the losses are bieing supplied.If you build this boost convertor dummy load for buck convertor testing with the same technology and parts as your buck convertor under test then the losses in the jig will work out about the same as your buck convertor losses.I built my boost convertor using the S TRAP switching regime for testing S TRAP buck convertors .If you are doing hard switched fixed freq current mode which is more common then you would probably be betteroff to build your buck like that .In fact the more similiar the boost is to your buck under test the better.So much for the burn test now for load response I just get a N chan fet and a resistor usually at close to the bucks rating .I hook the N chan gate to the siggen with a 10 V square wave at 300Hz and see on the scope what the output volts does and hence if it has to be modded.In other words we dont want crazy overshoots or dips.
