Evaluation of nRF52382

[Linton_za]

Good day

Firstly, an overview.

I am currently busy with a technical proposal for a client. Initial details will result in a completed approx 60 000+ sensor network. My idea is to create small clusters with a target which will include a BT module/Lora/MCU. The target receives all the sensors data and Lora module relays this data to a gateway. I am currently evaluating different mesh network examples but am pretty excited to use bluetooth mesh.

The sensor data is simple, i will just be reading the temperature from the die on the chip. I have looked at using the Murata mbn52382 module. I do have a few questions please.

I am looking at creating smaller cluster networks of around 100-150 modules due to the modules being encapsulated with lost of metal in the vicinity, and the maximum amount of hops limit, the data would be small and unsegmented to limit latency issues, and would like to know the following

1. The logistics behind creating the cluster networks. From what i have read you use an application to provision/commission each module. Would there be a less time consuming way to accomplish this?

2. Different cluster network would be in close proximity so would there be significant signal interference even though the target addresses will be different?

Thank you in advance

[bjorn-spockeli]

Q1: You should be able to automate the provisioning process and use the remote provisioning feature to provision the entire cluster from one mesh node. Remember that the provisioning is a one-time operation. 

Q2: Are the nodes of the different cluster networks supposed to exchange data with each other or are they two segmented networks that communicate through gateway nodes? If its the first, then you could let each cluster be a subnet and limit the number of hops and TX output power so that the messages are contained within the cluster.  

Best regards

Bjørn

[Linton_za]

Good day

Thank you for the response

1) Would it be possible to post a link to the automated process?

2) To limit the amount of hops, the clusters will have their own specified target node with Lora, but will be in close proximity to each other. If the next cluster, within a few meters of another cluster, not create interference or intercept the previous clusters broadcast messages, or can i isolate them?

As i stated i am new to BT mesh, and have done about 4-5 weeks research. I have chosen this over Zigbee as i see it as more of a challenge and interesting. I know BT Mesh is still relatively new but I have to see if it is viable and feasible for what is needed to be accomplished. I am hoping my theory is correct in creating smaller cluster nodes.

As i understand it, i broadcast my temperature data with a target address in the packet structure, each BT module in the cluster will then compare the address and pass it on until it reaches the "home/target" address node, even if they are not commissioned as relay nodes?

Thank you again

[bjorn-spockeli]

Q1: We do not have any example showing this in a very large scale, but the light_switch_provisioner example will provision (up to 30 nodes) on the fly as they are discovered. Remember that provisioning is a one time operation.  

Q2: Yes you can create subnets within the mesh that will have separate network keys, i.e. they will not be able to interpret the payload of the other clusters messages or relayed. You will always get some interference with many RF devices in a small area, but by choosing the hop count and TX power you should be able to minimize the interference. 

A Bluetooth Mesh node will only relay a packet if it is configured as a relay node. You can have end  nodes that only listen for incoming packets, but do not relay.