Your "outdoor" lights survived the rain. Congratulations. Now it's the third autumn and the string does not. I've pulled enough dead strings apart to know the pattern: LEDs perfectly bright, connector half-filled with green rust, and a single strand of water sitting on the driver board where nobody thought to look. The label on the box said IP65. It was not lying. It just wasn't telling you the whole story.

IP65 is a test, not a promise

Two digits, two different jobs. The first digit is dust. Six means the enclosure blocks dust ingress, which matters more than people think. A dust-fine film of soot on a LED is a little heat blanket, and over a season it shortens component life. The second digit is water. Five means the unit was sprayed with water from any direction, roughly a 6.3mm nozzle at about 12 liters per minute, for ten minutes, in a warm, dry room, in 2011 (IEC 60529, and yes, it's an old standard for a reason).

Read that again. Warm room. Dry air. Ten minutes. No freeze-thaw. No weeks of damp. No UV already embrittling the silicone. The test proves the sealing exists. It does not prove the sealing survives winter.

Where water actually gets in

In our factory runs the failure pattern is so consistent it's almost embarrassing. The LED bead is nearly never the casualty. The silicone dome around each diode is the most overbuilt part of the product. Water gets in through three places, in order of frequency:

One: the connector. Every plug-and-socket is a designed gap. A properly made connector has a silicone O-ring or a tapered seal face, and when it works, water beads up and falls off. When it doesn't, the mating faces wick moisture in like a paper towel, and corrosion starts on the contact pins within a wet season. Check one connector on an old string under a lamp. Green or white powder on the pins means the seal failed. The string dies at the connector, not at the LED. That's where the dead section starts.

Two: the cable entry. Where the wire exits the housing. Cheap units poke the cable through and call it a day. Good units use a molded strain relief where the cable sits inside a taper, so water running down the wire can't climb into the body. Hold your thumb over the exit and squeeze the body. If you feel air move, you know where winter is going to move.

Three: the driver or battery box, from below. Condensation. This one nobody tests for. A sealed box that gets cold at night and warm during the day is a tiny pump that drinks moisture out of the air and deposits it inside. A driver left in a box in November is often covered in a fine film of water by December, even though the box "never let any water in." The fix in the spec sheet is small and unglamorous: a breathing vent with a hydrophobic membrane (Gore-Tex style, if you want to be fancy), or at minimum a desiccant pouch that the box can't live without. If your "outdoor" product has zero venting and zero desiccant, it's not waterproof, it's on a timer.

IP65 vs IP66 vs IP67: what the digit after the 5 is actually for

The second digit is the water spec, and the jump is physical:

  • IP64 — splashes from any direction. Fine for a covered balcony, a garage porch, anything under a real roof. Fine for candles that live under an eave.
  • IP65 — low-pressure jets from any direction. The floor for anything in open rain. Your roofline, your hedge, your string lights across a terrace.
  • IP66 — powerful jets. This is where water thrown at the unit from a distance by wind-driven rain gets classified.
  • IP67 — submersion to one meter for thirty minutes. This is what you want for a pond feature, a fountain, or a candle that lives next to an ice machine in summer. It is also where most "outdoor" products stop pretending.

Here is the part retailers skip. IP65 does not include standing water. A light that sits in a saucer of melting snow for two days is doing something the standard never required. If your installation has a low point where water pools, you need IP67, not a more expensive IP65.

The test you can do in 60 seconds

Before you hang a new string, run your finger around every connector, every junction box, every cable exit. Three things to feel for: gaps you can feel with your fingertip, silicone that's cracked or chalky (UV did it, and the seal is gone regardless of the label), and any connector that opens with a push instead of a twist. Twist-lock connectors seal; push connectors drain.

Then do the freezer test on the box itself. One week in a freezer, then one day at room temperature, done twice, and look inside. If you've got a film of moisture, that product's driver is on a two-season clock in a climate with wet autumns. We do this on every driver we ship, and the ones that fail it go back. It's the difference between a light that lasts three Christmases and one that dies in November and blames the batteries.

For the candle side of the shed

Flameless candles run on the same physics, with one extra complication. They have a power button, and buttons are where seals go to die. The button mechanism has to flex, which means it has a tiny gap by design. An IP65 flameless candle with a well-made button (molded silicone gasket, travel under a millimeter) will outlast one where the button is a thin plastic flap. Same math, different product.

And the same condensation problem applies harder, because a candle's driver box sits lower to the ground, in the damp. A grave candle left outside all winter is basically a condensation trap. That's the one we spec with a desiccant pouch and a vent, and it's the reason ours still flicker in year two while cheaper ones go dark by February.

What to actually buy

Open rain, no pooling: IP65, and check the connectors. That's the whole spec. Wind-driven rain or exposed eaves: look for IP66. Anything near standing water or a pond: IP67, and verify the second digit is 7, not a marketing "waterproof." Candles: IP64 minimum under a roof, IP65 for open air, and a driver box with a vent or a desiccant, no exceptions.

The IP code on the box is a test score from a warm room. The question worth asking is whether the product was built for the part of your property that has no roof. Solid connectors, molded cable entries, a box that breathes. That combination is the one that lets you replace a string every three or four years instead of every season. It is the difference between a spec and a product.

Worth a look

  • Lyxu Christmas Lights — every strand we make ships IP65 or better, with twist-lock connectors and molded cable entries. If your roofline gets open rain, start here.
  • Lyxu LED Candles — flameless candles with gasketed buttons and desiccant-packed driver boxes, so they keep flickering into their second year out in the open.

Frequently asked

Is IP65 really waterproof? It is rated for low-pressure jets from any direction, so open rain is fine. It is not rated for standing water or submersion. If water pools where your light sits, you want IP67.

Why did my "waterproof" light die in the rain? Usually the connector or the cable entry, not the LED. Water wicks into a failed seal, corrodes the contacts, and the dead section starts at the joint. Check the seals before you blame the diode.

What does the number after the 6 mean? It is the water test. 4 is splashes, 5 is low-pressure jets, 6 is powerful jets, 7 is one meter of submersion for 30 minutes. Higher covers more.

Do flameless candles need an IP rating too? Yes, and the button is the weak point. Look for a molded silicone gasket and a vented or desiccant-packed battery box, or they will die to condensation rather than rain.

How long should outdoor IP65 lights actually last? If the connectors, cable entries, and box are built to spec, three to four wet seasons is reasonable. If they fail in one, the seal failed, not the standard.