Solar Street Lights from China: The Battery Label Is a Promise, Not a Number

Two years ago a client asked me to look at why his £6 mower’s solar light and a £70 one both died in the same month. Both were mounted on the same industrial estate fence, both in the same British weather. When we opened them the answer was boring: neither had a proper charge controller, and the cheap one had a battery with a capacity label that was off by roughly half the real value. The panels looked identical.

Solar street lights are one of the easiest categories in China to buy badly, because every part of the spec can be written to look good while the actual assembly is optimised for a landed price. What follows is what I check, roughly in the order it has bitten buyers I know.

Battery capacity is the first number that lies

Most of these lights ship with LiFePO4 cells in a plastic housing that sits in the pole or the lamp head. The capacity printed on the box is very often the nominal cell figure, not the usable figure after the pack is built and derated. Worse, the cells themselves can be recycled or B-grade stock, so the real capacity sags within months.

The practical checks are straightforward. Ask for the cell brand and model and confirm whether they are grade A new cells. Ask for a discharge test report at the pack level, not the cell level, at a defined temperature and discharge rate, with the start and end voltage recorded. Then, when the sample arrives, run your own capacity test with a programmable load before approving the production run. If the pack claims 30 Ah and delivers 21 Ah through a measured discharge, you have found a real problem early instead of a customer complaining in winter.

Winter is the whole point of the battery. A LiFePO4 pack loses usable capacity at low temperature, and a light specified for 8 hours of runtime in July may only manage 5 in January. Ask the factory for the runtime curve at the low end of the operating temperature range and see whether the controller compensates by dimming the LED, which is the honest solution, or simply switches off earlier.

The controller is doing more work than the panel

Most integrated solar street lights from China use a small charge controller board tucked into the lamp head. The good ones use an MPPT-style scheme; cheap ones use PWM and claim MPPT in the brochure. The difference shows up in the morning recharge, especially when it is overcast for three days.

Practical tell: ask for the controller’s charging efficiency and the MPPT voltage window, and check whether the board has any temperature protection. A controller that runs hot on a black housing in the sun is derating itself without telling you. Also ask for the low-voltage disconnect setting, and whether the light is programmed to dim during the night. A light that stays at full brightness until the battery cuts out will be the first one to fail.

For pilot orders, ask for the controller firmware version and whether you can change the lighting profile with a remote or app. If not, the factory’s default profile will run the light brighter than the pack can support, and you cannot fix it later.

LEDs, lumen output and the honest figure

What is marketed as the wattage is not the load. A “300W equivalent” integrated street light typically draws 30 to 60 watts of actual power. Ask for the measured input power and the delivered lumens at the rated power, and get a photometric test rather than a datasheet boast. A realistic efficacy for a decent integrated light is around 120 to 150 lumens per watt of actual consumption on a good design, and anything promising significantly above that at a low price is worth a second look.

Lumen maintenance matters too. Ask for an LM-80 report from the LED chip maker, and confirm the chip model. Substituting an unknown chip with a similar CRI and colour temperature is common in this category, and it will show up as colour shift across a batch rather than a hard failure.

Optics and how the light lands on the ground

Integrated lights use a lens array that spreads light along a road. The useful checks are the beam angle, the uniformity of illumination on the ground at the mounting height you are using, and whether there is upward light emission that annoys neighbours. Ask for an IES photometric file if the factory has one; if they do not, ask for a photo of the light pattern on a wall at two heights. Cheap lens arrays tend to concentrate a hot spot in the middle and leave the edges dark, which is exactly the complaint on council and car park installations.

Mounting height and lamp spacing should be specified from the photometric data, not guessed. A 6 m mounting height with a narrow beam gives you patches of light between poles.

Housing, IP rating and the thermal path

The housing is the heatsink. On integrated lights with a die-cast aluminium body, check that the LED board is thermally bonded to the housing, not screwed through a thin plate with paste smeared on. Ask for the thermal pad material and confirm the housing thickness at the LED mounting area. A light with a good chip and a poor thermal path still loses half its life in two years.

The IP rating claim needs a real basis. IP66 covers jets and dust, and it must be backed by an ingress test report on the actual housing design, including the cable gland and any sensor window. The most common failure point is the cable entry and the seal at the pole bracket. On the sample, check that the gland is a proper waterproof type with a compression seal, not a plastic bushing with silicone around the cable.

Certification, and what the label actually means

European buyers want CE marks with an LVD and EMC basis plus RoHS, and increasingly an EN 62471 photobiological safety assessment for the LED array. The CE mark itself is the manufacturer’s declaration; it is not evidence. Ask for the test reports with a lab name, report number and model designation matching your light, and check the lab accreditation. Reports for a different model of a similar product are extremely common and tell you nothing.

Battery transport is a separate matter. LiFePO4 packs shipped by sea still need UN 38.3 test documentation and a proper MSDS, and if cells are shipped separately or as spares you need to confirm the packaging is compliant for the transport mode. That paperwork is the factory’s job to provide with the shipment, not after it.

What a good order looks like

Specify the actual input wattage, the lumens measured by test, the lumen maintenance figure, the applied LED chip model, the pack-level capacity with a discharge test report, the low temperature runtime, the controller technology and lighting profile, the IP rating with the test report, and the certification documents per model. Then test one sample per model before the container ships, including a full night of runtime on a real battery.

The £6 mower light in my opening story is the extreme case, but the logic runs all the way up. In solar street lights, the printed number is a promise and the only thing that matters is what the pack delivers in February. Where model-specific data is concerned, use the manufacturer’s own published documentation as the source of truth, and verify the delivered product against it.