At noon in summer, the 5G CPE drifts from 800Mbps down to 200Mbps, and a reboot brings it back. Not carrier throttling — the device 'overheated and throttled itself'.
5G modules run far hotter than 4G
At high modulation (256QAM) and high bandwidth, a 5G module draws 8-12W — 2-3x a 4G part. Heat builds in a palm-sized shell; once it passes ~70-80°C the baseband auto-throttles to survive, and that's the 'speed drop' you see.
Three cooling designs, and where they differ
- Passive (heatsink + metal shell): fine for light load, fails in a sealed summer cabinet.
- Active fan: sustains high bandwidth but adds noise and fan-life risk.
- Orientation: upright with sink facing up beats flat; never shove it in a sealed plastic box.
Three hard rules for the field
- Keep ambient <= 40°C; ventilate/shade the cabinet, no direct sun.
- Pick a metal-shell model with a large heatsink; skip all-plastic mini boxes.
- High-bandwidth use (streaming backhaul, enterprise egress): choose active cooling with fan redundancy.
5G's speed ceiling is on the spec sheet; how long it actually holds is written in the thermal design.
FAQ
5G CPE power consumption and cooling?
5G modems draw more power—need metal casing/venting; industrially, check operating temp range.
Can 5G FWA replace fiber?
With good coverage it hits hundreds of Mbps—great where fiber is hard to run; but stability/latency still trail fiber.
What real-world speed does a 5G CPE deliver?
Depends on band (n1/n78/n41), signal, and carrier aggregation; real numbers fall short of specs.
Is lightweight 5G (RedCap) right for IoT?
RedCap cuts module cost and power—fits mid-speed IoT (surveillance/wearables), not high-bandwidth uses.