18 s
keys: R replay · Space pause · H record mode · arrows step · URL: ?rec&dur=18&paused&loop

Ran it flat. Then counted what went back in.

FNB58 inline, wall brick, lights off, no host · 2026-09-11 12:36 to 2026-09-12 06:30 · three logged readings on a straight line
Went in at the port
0.00Wh
energy, which is what the pack stores
Elapsed
0:00
hours:minutes
Energy in against time, with the two answers drawn as bandsone cell 18.5 Wh · two cells 37 Wh · charger 74 to 92% efficient
PAST THE ONE-CELL BAND AT ABOUT NINE HOURS, STILL PULLING HALF AN AMP 50.7 Wh · TWO CELLS’ WORTH, EVEN AFTER THE BOARD’S OWN DRAW · 37 Wh PACK CONFIRMED FLAT TO FULL: 18 HOURS AT HALF AN AMP. PLAN ON OVERNIGHT
Logged: 0.41 Wh at 0:07, 15.26 Wh at 5:24, 50.70 Wh at 17:54; 0.49 A at the start, then a flat 0.58 A throughout. Unplugged, switched to wireless, battery check read full. A single 18.5 Wh cell finishes between 20 and 25 Wh at the port allowing 74 to 92% charger efficiency; two cells between 40 and 50, plus the board’s own draw, which at worst is the 145 mA awake figure for the whole run, 12.6 Wh, still leaving 38 Wh, over twice one cell.
The charger runs constant current to a cutoff with no top-off phase, which is why the trace never tapers · CK Tech Check