Splash, Frost and False Alarms: A Winter on the Garden's IP-Rated Sensors

The word 'weatherproof' does a lot of work

IP ratings describe test conditions, not a British winter. A sensor rated IP65 is protected against dust and water jets; an IP67 device can survive temporary immersion. Neither rating says much about condensation, UV degradation, freezing temperatures or the quality of a rubber gasket after six months in the sun.

At CavLab we put five outdoor smart sensors in a garden in the south-east of England from October to March. The kit covered the categories most buyers consider: a Zigbee temperature and humidity sensor rated IP65; a Wi-Fi leak sensor rated IP67; a PIR motion sensor rated IP65; a Wi-Fi garden camera rated IP66; and a second temperature sensor rated IP44 as a control. All were mounted as their instructions suggested, except where noted. The aim was not a laboratory ingress test. It was a field test of what happens when 'weatherproof' meets rain, wind, frost and a fox.

Where the water actually gets in

IP65 means dust tight and protected against water jets. IP67 adds temporary immersion. Neither rating covers condensation. Sealed enclosures breathe as temperature falls; moist air condenses on cold surfaces. The IP66 camera stayed dry inside its main housing, but water pooled in the mounting foot and froze, cracking the bracket. That is not an ingress failure in the rated enclosure, but it is a failure of the product as installed.

The IP65 PIR had a gasket that was pinched at one corner from the factory. After a storm, a bead of water sat on the PCB. The unit still worked, but it was on borrowed time. The IP44 temperature sensor, mounted under the eaves, was the only one with visible internal condensation. That is the counter-intuitive lesson: a sheltered position can be more humid than an exposed one, because it dries more slowly.

The IP67 leak sensor was the most convincing. It sat in a drain gully for the winter and showed no water ingress. It did, however, false-alarm once when condensation bridged its contacts. Ingress protection and false-alarm immunity are different engineering problems.

Batteries: the cold is the real test

Two sensors used AA alkaline cells, two used CR123A lithium cells and the camera was mains powered. At -4°C, the alkaline cells sagged. One motion sensor reported 20% battery and then died in January. The lithium units held above 2.8 V and kept working. Battery percentage is often a voltage guess, and cold makes it pessimistic. A dead sensor in February is frequently a battery chemistry problem, not a water problem.

That matters for buying decisions. If a sensor is outdoors through a British winter, lithium cells are usually the better choice, even at higher cost. If the manufacturer specifies alkaline, check whether the battery compartment is insulated from the cold. The camera's mains supply avoided the issue entirely, but it introduced a cable entry that needed a proper drip loop.

Connectivity through wet foliage

Wi-Fi RSSI at the camera fell from -58 dBm in dry October to -70 dBm after heavy rain. The hedge was holding water, and the 2.4 GHz path was measurably worse. The camera dropped two recordings during a named storm. The Zigbee sensor, mounted through double glazing, remained stable. That is not a universal result, but it is a reminder that outdoor connectivity is a seasonal variable, not a fixed number on a box.

False alerts were mostly a PIR problem. The motion sensor triggered 14 times in one windy week. The clips showed a fox, a cat and a moving shrub. A pet-immune setting helped, but the shrub still caused triggers until we narrowed the sensitivity zone. The camera's motion detection was more configurable and less prone to wind, but it needed a narrower activity zone to avoid recording every passing car.

What held up, and what did not

The IP67 leak sensor was the standout: no ingress, one false alarm and a clear lesson about condensation bridging contacts. The Zigbee temperature and humidity sensor was unremarkable in the best sense: stable readings, no dropouts and a battery that lasted the winter. The lithium-powered PIR survived, but only after we replaced the alkaline cell and reseated the gasket.

The camera's bracket was the weakest physical component. The enclosure was fine; the mounting hardware was not. That is a common pattern in outdoor kit: the rated part is tested, the accessory is not. The IP44 control sensor suffered condensation, which suggests that the rating was accurate but the placement was not.

Lessons for a British garden

Buy for the enclosure, not just the IP number. Look for UV-stabilised plastics, gaskets with a lip rather than a flat washer, and cable entries that can take a drip loop. Mount under cover where possible, but do not assume that shelter means dry. Avoid low spots where water pools and metal that conducts cold into the battery compartment. Use lithium cells for winter duty. Check seals each autumn and clean the lens or sensor window. Treat 'weatherproof' as a starting point, not a guarantee.

The verdict

Three of the five sensors ran the winter without intervention. One needed a battery change; one had a gasket fault. The garden is a harsher test than a lab and a more useful one. The best weatherproof sensor is not the one with the highest IP number. It is the one you can open, dry and reseal without breaking a clip.