I’ve been experimenting with high-power LEDs in open, non-commercial setups to better understand real-world thermal behavior outside finished products.

What stood out was how strongly non-electrical details affected stability: – mounting pressure – interface materials – real airflow paths versus assumed ones

Electrically everything stayed within ratings, but long-term thermal behavior varied more than expected.

For those who’ve worked with power-dense hardware: what thermal assumptions turned out to be wrong in practice?