What does receiver sensitivity (dBm) mean, and how does it determine range?
Receiver sensitivity is the weakest signal level the receiver can still decode at a defined error rate, and it is a negative number in dBm; more negative is better. Sub-GHz FSK receivers typically report −125 … −130 dBm; the LoRa SX1276 goes down to −148 dBm at its lowest data rate. A 6 dB improvement in sensitivity doubles the theoretical range.
Sensitivity is not a number on its own; it is always given with a data rate and an error rate condition. There can be a difference of more than 10 dB between the same chip's sensitivity at 300 kbps and at 1.2 kbps. So when comparing the sensitivity figures of two products, do not comment without checking at which data rate they were measured.
Why is slower always more sensitive? Noise power is proportional to bandwidth. Reducing the data rate narrows the receiver bandwidth, the narrowed band collects less noise and the threshold moves down. The cost is that the same command stays in the air longer; this also fills the duty cycle budget quickly.
Modulation is decisive too. FSK/GFSK, being independent of amplitude noise, gives a clear advantage over OOK. Spread spectrum methods such as LoRa can decode a signal even below the noise floor; at low rates a gain of roughly 4–6 dB over FSK is reported, and far more at the most extreme settings.
These values are typical chip data and cannot be attributed to a specific product; a device's real sensitivity should be sought in its own technical data.
Full coverage of the topic: 433 MHz or 2.4 GHz?