Sixty-six watt-hours, and the nights they cover
Six 18650s at 3,000 mAh and 3.7 V is 66.6 Wh; count 53 Wh usable, because the controller stops well before the cells are flat. Standby is 5 mA — 0.02 W — which is 0.44 Wh across a whole day of doing nothing. Thirty PIR events of twenty seconds and twenty minutes of live view add about 1 Wh between them, so a quiet site sits near 1.4 Wh a day and the pack would run for five weeks in the dark.
| Duty | Wh a day | Days with no sun |
|---|---|---|
| 30 triggers, 20 min of viewing | ≈1.4 | ≈35 |
| 80 triggers, 1 h of viewing each evening | ≈3.3 | ≈16 |
| One full 12 h night on the app | ≈24 | one night, then two or three days of sun |
That third row is the honest one. Working draw at night is 500–550 mA, call it 2 W, so a twelve-hour night of continuous streaming is 24 Wh: nearly half the pack for one evening of curiosity. The build survives it — which the 5 W version does not — but it survives it once, and then it wants two or three clear days before it is comfortable again.
- ≈66 Wh installed, ≈53 Wh usable (6 × 18650, 18,000 mAh)
- ≈1.4 Wh a day at light duty; ≈3.3 Wh with an hour of live view every evening
- ≈26 h of continuous night streaming on a full pack — two autumn nights, not three
- 8 W panel: ≈28 Wh nominal at 3.5 peak-sun hours, ≈20 Wh into the pack; ≈6 Wh in December at one peak-sun hour

The panel is not sized for the alarm
Eight watts against a 1.4 Wh day is a factor of fifteen, and it is worth saying plainly that nobody needs that margin for PIR duty — five watts would do. The extra array exists for two other days. The first is December at 52° north, where an 8 W panel flat on a bracket returns about 6 Wh and a 5 W one about 3.5, which is the difference between refilling the pack in a week and watching it sag. The second is the site where the client discovers live view, and starts opening the app every evening.
Radar, and the afternoon the wall is hotter than the man
A passive infrared sensor works on contrast: a body at 37 °C against a background at 20 °C is obvious, the same body against a wall that has been in the sun since noon is not. That is the failure mode this camera is built around, and it is why the factory pairs the PIR with a radar front end. Radar does not care about temperature; it cares about movement, and it hands the camera a trigger the moment something enters the cone, before the thermal signature develops. A vehicle at 50 km/h is inside 12 m for 0.86 s, which is not long for a PIR to settle and is ample for radar.
The cost is the other direction: radar cannot tell a branch from a burglar, and a gale in a treeline is a stream of triggers. So the two are ANDed — both have to agree — and the false-alarm rate ends up closer to the PIR's than to either sensor alone. What the listing does not publish is the radar module, its frequency or its share of that 5 mA standby, so if you are sizing a site on the power budget, ask us for the split rather than guessing.


What we would tell you before you buy it
If the camera has to see what happened before someone arrived, this is not it — no amount of panel fixes a sleeping sensor, and the AOV build is the one that records continuously. If you need a number plate at 100 m, this is not it either; that is the 100× dual-head. What this one does that neither does is survive a post nobody visits: six cells you can replace, a panel with margin, IP65, −30 °C, and two sensors that keep working in the heat. Buy it for the fifth year, not the first.