Support for more than 8 components with Zigbee nRF52 #3689
Replies: 1 comment 1 reply
-
|
Point 1 - it is possible to get access to source code if someone willing pay fee. It was couple of k$. And it would not be certified anymore.
This is something that I considered but to simplify implementation it was not done. Feel free to prepare PR :)
That would required changes in zha and z2m to make it user friendly. It is possible but first I would reuse endpoints.
@luar123 we discussed this before and consensus to not relay on that. |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
Describe your idea
Hey, I was a bit frustrated with the maximum 8 components supported per zigbee device on nrf52 so I dug a bit deeper.
Currently this limitation comes from two things:
Point 1 is a limit that cannot be changed. Point 2 comes from the fact that we use standard clusters definition (like 0x000F = "binary input cluster" for binary sensors and so on), so we could in theory mix components types in a cluster but we would still only be able to put 8 of each component type per device.
How about we create a custom external converter that takes advantage of the manufacturer range of clusters (0xFC00 -> 0xFFFF = 1024 values) where we would define that 0xFC00 -> 0xFC0F is for binary sensors, 0xFC10 -> 0xFC1F is for sensors (analog input) and so on?
This would require manufacturer (always "esphome" currently) and model (currently friendly name of device) to be constant.
Why is this exciting?
No more hard limits on nRF52 zigbee devices! 8*16 = 128 components per devices would be possible.
Additionally support for other component types would become easier to implement as we could move away from standard cluster IDs
Potential use cases
Low power device returning lots of data of the same type (LinkyTIC for example)
Anything else?
Is this a direction you would be interested in going?
I am currently trying to do a proof of concept using AI but I may not be able to provide robust enough code for final implementation.
Beta Was this translation helpful? Give feedback.
All reactions