A lighting distributor from Lagos visited the factory this week with a bag of failed components and a straightforward request: show me why these died, and show me how yours are different.
He had been importing solar products for years. Some batches worked. Others failed within months, and the returns were eating his margin and his reputation.
| Symptom | Usual root cause |
|---|---|
| Light dims after six months | Low-grade cells losing capacity, or undersized panel |
| Water inside the housing | Single gasket path, unsealed cable entry |
| Works at dusk, dead by midnight | Battery capacity quoted at 25°C, not derated for heat |
| Flicker or no start | Unpotted driver, moisture on the PCB |
| Sensor triggers constantly | PIR sensitivity set for indoor use |
None of these are exotic failures. They are the predictable result of shaving cost at the component level and skipping tests at the factory level.
We also spent time on what a healthy installation should look like after two years, because that is the point where cheap products and durable ones separate. A good lamp keeps its output within a predictable depreciation curve, shows no condensation behind the lens, and still reports a sensible state of charge on a grey morning.
He asked how to check a batch without a lab. The practical answer is a small sample run: leave five units on a test street for a month, log the switch-on and switch-off times, and open one housing afterwards to inspect the gasket and the battery pack. If the seal is dry and the cells are balanced, the batch is likely sound. If there is moisture or the pack is swollen, no certificate will save it.
He left with samples from two series and a plan to test them on a single street before replacing his stock. That is the right order of operations. A street test costs almost nothing and tells you more than any certificate.
If you are tired of return rates, send us the failure pattern. We will tell you which component is likely responsible and whether the fix is a different product or a different specification.