BLE sniffer vs ANT continuous scanning mode: nRF52832

[Oleg@LTG]

Right now we're using nRF52832 on our devices for various BLE Sniffer features.

We need to improve both capacity and throughput of the BLE Sniffer - ideally we need to receive 150-200 messages per second from different wireless sensors. Otherwise we start observing issues even with 30-40 BLE sensors around, the packet loss for BLE is pretty large ... 

We started thinking about TDMA-based wireless protocol instead of BLE, because it looks like the BLE Sniffer simply can't handle even 100 messages per second. But we need to handle even more.

After a small research we discovered the ANT protocol with its Continuous Scanning Mode, which is somewhat similar to the BLE Sniffer concepts. Plus, the ANT protocol is based on TDMA-protocol with much better capacity than BLE for scanning and sniffing.

So we're planning to start some performance testing to compare the BLE Sniffer vs the "ANT Sniffer". Our goal with ANT is about achieving 200 messages per second with acceptable packet loss numbers ...

Before we will jump into the implementation details and activities - any advices or useful references about the ANT Continuous Scanning Mode on nRF52? Has anybody tried using the scanning mode with hundreds of sensors?

Thanks in advance for your replies.

[Simonr]

Thank you for the clarification, Oleg. 

According to the ANT Continuous scanning mode application note (attached), the number of simultaneous transmissions that can be supported by a single scanning device is limited by the available bandwidth, which is determined by the master devices. Thus, a node configured in scanning mode can support the equivalent number of transmissions that can be supported by independent master/slave pairs in a given RF space. So first of all, this depends on the size of each of your messages, but I don't think as much as 200 messages per second seems feasible.

ANT_AN14 - Continuous Scanning Mode Rev1.2.pdf

Best regards,

Simon

[Oleg@LTG]

OK, thanks for your reply.

I have another question in this space - about BLE and using the Timeslot API in Nordic SDK for implementing TDMA-based protocols.

Again, my goal is about improving scalability and capacity of BLE to receive hundreds of messages per second. Using the time slots and TDMA might be an option here. Combining TDMA and FDMA is another possible option.

Question to you about the Timeslot API of Nordic SDK - can it be used to implement TDMA or FDMA data exchange protocol based on BLE?