Bluetooth 5 range put to the test

The introduction of Bluetooth 5 added some key capabilities to Bluetooth wireless technology, notably increased throughput, range, and advertising extensions. Nordic Semiconductor’s nRF52840 SoC was one of the few devices on the market to offer all of Bluetooth 5 wireless’ features from the date of adoption.

Some competitive chips were slower to offer the long-range feature because it required a significant firmware upgrade. Specifically, the range boost comes from the use of Forward Error Correction (FEC) to detect and fix packet corruptions during data transfer.

This improves the Bit Error Rate (BER), effectively increasing the sensitivity of the receiver by enhancing the transmitter. Theoretically, for the Bluetooth 5 long-range feature, the increase in sensitivity should be around 12 dB (at a throughput of 125 kbps) compared to (non-FEC) 1-Mbps mode. A 12-dB enhancement promises a quadrupling of range.

Because the technique reduces the packet data payload, the downsides are reduced raw data rates, longer transmission times (reducing battery life), and a greater risk of interference from other 2.4GHz radio sources. But for some applications, for example smart-lighting or -agriculture, the potential of a quadrupling of range is well worth the trade-off.

But the promise of four times the range is one thing, actual range in practice is quite another. There are many factors that attenuate a Bluetooth 5 signal. In part this attenuation is due to interference from other 2.4GHz wireless technologies such as WiFi, ANT+, Thread, and Zigbee, but perhaps a bigger factor is the impact of obstacles between transmitter and receiver. Indoors these obstacles comprise people, walls, and furniture; and outdoors things like trees, buildings, and even the weather get in the way.

Boosting range

Rutronik is a German component distributor of Nordic’s wireless SoCs, including the nRF52 Series Bluetooth 5 chips. The company put the nRF52840 SoC’s long-range feature to the test using Nordic’s nRF52840 Development Kit (DK) in both indoor and outdoor applications. The results were compared with Nordic SoCs running Bluetooth 5 in ‘normal’ mode (i.e. without FEC).

Indoors, the results showed that a Bluetooth 5 SoC without FEC (raw throughput 1 Mbps) offered a range of 51m. With maximum FEC, that improved to 60m (at similar output power but a raw throughput reduced to 125 kbps); a range improvement of around 17.5%. Outdoors, for similar operational scenarios, range improved from 455 to 527m, an improvement of around 16%.

While this is a marked improvement in range, such improvement falls well short of what might be theoretically expected from the new technology. Rutronik noted that the indoor radio environment was particularly hostile. The tests were performed in an office with lots of walls, people, and furniture.

In addition, many other 2.4GHz devices such as WiFi routers and wireless mice and keyboards were present. Similarly, the company noted the outdoor results were likely to have been compromised by weather, ground reflection, and less-than-optimum positions of the respective transmitter and receiver antennas. The company further noted that a wider range of tests would need to be performed to establish Bluetooth 5 range performance across a range of typical usage scenarios.

Nordic has performed its own Bluetooth 5 outdoor experiments, which reportedly post-date Rutronik’s. The company notes that although a 12-dB increase in receiver sensitivity is theoretically possible, actual devices achieve around 8dB which equates to a range boost of 2.5 times.

Nordic performed its tests using an nRF52840 SoC running no FEC (raw data throughput of 1 Mbps) then with full FEC (125 kbps). For a normal Bluetooth 5 connection, the range for the former operational mode was 682 m and for the latter 1,300 m.

That’s an increase of around 91% or just under double the range. Nordic’s engineers also noted that the impact of external factors is so great that it is impossible to generalize what improvement in range Bluetooth 5 can bring. Tests in the target application are therefore advised.