I’m having some problems with One to many routing, I have 4 ZB PRO ROUTERS and 1 ZB PRO COORD.
All the four routers need to send data non-stop to coord. at the same time, is it possible to do that without losing any data on the way?
I made a small application that recieve all the data, and actually i can recieve data up to two routers, but when i request the third one to send it’s data i start getting checksum errors.
I’m using baudrate of 115200
ZB is probably not the right technology - nor is mesh in general. ZB will at times not move ANY data for up to 5 seconds. You also need to use API node - I assume you are trying AT mode, and as you are seeing, the packets start fragmenting and getting intermixed.
My rule of thumb is don’t expect any 1 node to handle more than 10 of your packets per second. So that means 2 per second per router.
To not lose any data at all (& still use AT mode), you are probably need 4 routers, 4 coordinators, and use 4 different frequencies. But even then, someone using a microwave oven or a bluetooth headset might cause gaps in your data.
Actually I’m using API mode, does that make any difference? Sorry if i’m making a stupid question, i’m new with ZB
by the way thanks for the reply!
API mode lets you control the size & quality of your packets. For example, the ‘router’ will get a TX status response telling you that the ‘next hop’ accepted the packet - this is likely the coordinator, but might NOT be.
Will data get lost? Of course. ZB is much like UDP/IP, and even on wired Ethernet, UDP/IP allows things to get lost. Despite the TX status response, your router-attached CPU will never know that the external CPU on the coordinator accepted & processed the data. Also, if the coordinator becomes overloaded, meaning data is arriving faster than it can serially shift it out to the external CPU, then it will run out of buffer space and be forced to drop incoming RF packets.
In the end you’ll need to develop a simple windowing protocol (like TFTP) where the router-CPUs number packets sent and the coordinator-CPU acknowledges the packet numbers it has processed. This allows retransmission at an application level.