On a USB-C to ATX supply, the feedback divider computes to exactly 12.000 V. The board’s power-good supervisor releases at 11.6 V, so at nominal the margin looks comfortable. The open question is the one the files don’t answer: which resistors are actually fitted.
Why it can’t be decided from the files
The reference tolerance alone puts the rail between 11.88 V and 12.12 V. Add the resistor tolerance and the picture splits:
- With 1 % resistors, the worst-case low is 11.66 V — it clears the supervisor by 60 mV.
- With 5 % resistors, the worst-case low is 10.82 V — 776 mV below the trip point, so power-good could never assert.
Same divider, same nominal voltage. One tolerance class boots; the other never comes up. The BOM doesn’t state which one is on the board.
What Probe did with that
Probe took the reference tolerance from the datasheet and the resistor values from the BOM, established the bound the data actually supports, and named the empty tolerance field rather than quietly assuming 1 %. The most dangerous number in a review is one that is exactly right — it invites you to stop looking.
The finding here wasn’t a defect. It was the name of a missing field, and the reason it mattered. The most useful thing the review produced that day was not a defect — it was a question the BOM had to answer before the board could be trusted.