I am trying to use a buck converter (https://www.diodes.com/assets/Datasheets/AP64501.pdf) to output a constant ~5.15V for a Raspberry Pi CM4. I have pretty much copied the circuit from the CM4IO board that Raspberry Pi makes:
I have one of these on my desk and everything regarding the 5V power supply functions as expected. Note that C61 and C59 on the CM4IO board are unpopulated, so the only difference between it and mine is I did not include R9 between the BST and SW pins, and did not include as many input capacitors (these are not included in the recommended application from the datasheet):
I realize now that I did not follow the layout guidelines in the datasheet as recommended, but I don't believe that the issues I'm seeing are a result of that. In particular, with no load, the voltage going to the feedback pin of the converter is not 0.8V. Instead it's around ~0.89V, though I do measure 5.15V at the output, which is my target. However, theoretically, I should measure 5.15*(2.2/(12+2.2)) = ~0.798V at the FB pin between the two resistors in the divider.
When I do add a small load, the output voltage as measured at the positive side of C20/C22/C25 falls to around 4.6V, and the voltage at the FB pin falls to around 0.71V. Which now theoretically matches what it should be as calculated from the voltage divider (4.6*(2.2/(12+2.2)) = ~0.713V). But voltage at the output capacitors should rise until the voltage at the FB pin reaches 0.8V. What is the reason that it is not doing so?
I have double checked that all of the solder joints are good and that all components are correct. Is my converter IC damaged, or am I wrong and this is somehow the expected behaviour?
I've looked at it with an oscilloscope after Frog's recommendation below, and it looks like it picks up to the targeted voltage every 100ms or so. However it falls back down to about ~3.4V within 2ms. It looks like this:
What could be the reason that it is not able to maintain the voltage?