v0.1.11

[Nerivec]

Initial release.

---

This discussion was created from the release v0.1.0.

https://github.com/Nerivec/zigbee-on-host/releases/

---

The basic ZigBee features are now available.
See testing section for more information on how to set this up.

Important

This is experimental. Expect crashes, and several non-working features.

Tip

Make sure to run with logging set as debug so you can share them in case you encounter any issue!

Known limitations (no need to create issues for these)

• Current status
• Devices known to be troubled (especially regarding pairing) may not work (properly) for now. Current implementation is void of workarounds.
• [Silabs EFR32] Direct pairing of Green Power devices to Coordinator requires a fix in RCP firmware (these firmware have the fix, note however that the impact on Thread RCP use with these has not been tested)

Tips

• If your adapter doesn't want to start (after a crash for example), try unplug/replug.

[chris-1243]

The EUI64 (IEEE address) in the firmware of the coordinator is ignored in this mode. A static one is used instead (set by Zigbee2MQTT), allowing you to change coordinators at will on the same network (although you may encounter device-related troubles when radio specs vary wildly).

Would you think there is the possibility to manually add an IEEE in configuration.yaml under serial ? It would ease, in my point of view, the move from NCP to your ZOH implementation. At the moment, may we edit zoh.save and add an IEEE value?

TI firmware are working (at least to create the network. I have not gone further for the time...). Let's wait for the latest-dev to be published

[Nerivec]

Yes, further down the road, we'll look into making it configurable (requires some edits on Z2M side).
https://github.com/Koenkk/zigbee-herdsman/blob/a5c4fe767b02b2a937f70a0459520e7e72ee92f5/src/adapter/zoh/adapter/zohAdapter.ts#L83

For now, I think it's best not to offer the possibility to move from NCP to ZoH, too many questions left to answer. It could do more harm than good, and best to focus on specific issues for now (rather than a whole network having issues after a swap).

[chris-1243]

So, I had a quick try... and this is an amazing start!

I edited my configuration.yaml and started a fresh network.
Result: no errors, the network was up and running.

I added an Ikea bulb, some errors were thrown but the light joined and was perfectly working. As I was curious enough, I paired a ZGP device (PTM215ZE). This was a success as well with some errors like the one I had before. The device, on the other hand, was paired as broadcast. This should be investigated as I selected the Ikea light as a sole means of pairing.

I also tried to take over a NCP network with 5 devices. I had a positive result as well.

In Z2M, LQI values are negatives. Small bug to correct.

I will provide a debug log later of this test.

(I used a TI chip, Sonoff Dongle-P. It was not the latest firmware version due to a small bug which broke the bootloader...)

[Nerivec]

some errors were thrown

This should be investigated

Can you send me the debug logs? (on Discord or open an issue here)

LQI values are negatives. Small bug to correct.

That's expected, as mentioned in the checklist, these values are currently RSSI in dBm (from Spinel metadata - I haven't investigated that part much yet).

[chrivers]

Interesting!

With regular z2m, I've had that exact same issue. I have some ikea bulbs, and some PTM215Z green power switches, and I've been trying to get them to pair to a single bulb, but they seem to be pairing to everything.

It's actually causing a bit of a broadcast storm on my network, whenever I use these switches...

So I'd argue that:

A) It's not ZoH-specific
B) It might not even be z2m-specific

At least point B) is possible, if the ikea firmware is buggy, no?

[chris-1243]

Which Ikea bulb and which coordinator chip?

I am using a CC2652P (TI chip) on my regular network and all my PTM are paired through a specific router and they are all working as unicast.

This is where things come interesting.

A) It's not ZoH-specific

From my experience, I would argue for a yes. It might be ZOH-specific

B) It might not even be z2m-specific

Yes and no. No, as some HUE bulbs are known to have buggy firmware.

Unfortunately yes as it could be related to the driver used in Z2M and the implementation of the zigbee green protocol.

Even more interesting, PTM215Z/ZE (GreenPower_2) are quite easy to pair through a specifi router.
PTM216Z (GreenPower_7) is quite close to a nightmare in my scenario as only a few bulbs are able to translate the frame sent and as broadcast (HUE bulbs whith those buggy firmware...).