I’m working on a project trying to characterize the range (given a certain transmitter and environment) on the BLE 1 mbps PHY. I’m using a few nRF52840 dongles to do this as the nRF52840, according to the specifications, has a pretty good sensitivity (PSENS,IT,SP,1M,BLE -95 dBm according to Nordic Semiconductor Infocenter). However, the values I receive in my tests are far off from this sensitivity. Given that I transmit packages with payloads <= 37 bytes the Bit Error Rate (BER) should be <= of 0.1 % at the sensitivity limit -95 dBm, according to the specifications (nRF52840 and BLE). This should, assuming uniformly distributed bit errors (which is the worst case), give a packet error rate (PER) <= 5% at the receiver sensitivity limit of -95 dBm (unless I'm mistaken). However, in my testing, the RSSI limit for a PER < 5% is around -75 dBm (somewhere -70 to -80 dBm).
A 20 dBm difference from the hardware specification seems to be too much. What factor am I missing that could explain this difference?
Some specifics:
I’m transmitting advertisements from an Android smartphone with the nRF Connect app. I’m receiving the advertisements on three nRF52840 dongles running the nRF-sniffer firmware (v.4.1.0) each connected to a Raspberry Pi running the sniffer API (v.4.1.0) via a USB-cable. The receivers are moved along on a field where I presume that there is minimal interference from other devices.
The lowest RSSI that I’ve recorded is around -91 dBm (which makes sense given the RSSI calibration limits of nRF52840) for packets with errors and -87 dBm for any correctly received packet.
