Buck Converter Damage Traced to a Fixed-Output Part Variant
Description
Boards were coming back with damaged buck converters. The schematic was correct, the layout was sound, and the part number on the BOM was the right part — on paper.
The Fault
The AP63203 buck converter ships in more than one variant. One is adjustable and expects an external feedback divider to set the output. Another is fixed-output, with the divider already inside the package.
The design used an external feedback divider, which is correct — for the adjustable part. The variant actually fitted was the AP63203WU-7, the fixed-output version. With an internal divider and an external one both acting on the same node, the regulator was driven outside its intended operating point and damaged itself.
The suffix is the entire difference. Everything upstream of it was right.
Why It Matters
This is a BOM fault, not a design fault, and it is invisible in every artefact a normal review looks at. The schematic is correct. The layout is correct. The part number is almost correct.
It only surfaces by reading the ordering-information table against the circuit the part is sitting in — which is a habit, not a tool.
Deliverables
- Failure mechanism explained from the datasheet variant table
- Corrected part number for the topology in use
- BOM audit against variant suffixes across the rest of the design
Published without client, project name, or location under NDA. Silicon and findings are unchanged.