Aqara Curtain Driver E1

FloatingBoater · February 18, 2025, 3:26pm

Hi,
No solution but another data point of a different coordinators making a difference with Aqara although with ZHA not Z2M:

I must copy the report to the ZHA GitHub as an issue report.

EDIT: Reported to ZHA GitHub

github.com/zigpy/zha-device-handlers

[BUG] Aqara+ ZHA+ ZBT-1 FAILS; Aqara+ ZHA+ Sonoff WORKS

opened 03:39PM - 18 Feb 25 UTC

FloatingBoater

possible bug

### Bug description Hi, **TL;DR**: ZHA + Aqara Zigbee seems to fail with Yell…ow/ SkyConnect/ ZBT-1 but work with a Sonoff 3.0 (TI CC2652) coordinator. **Anyone else seen this before I go into full logs / microscope / GitHub issue mode please?** I recently bought a few Aqara devices from the nice folks at [EST](https://shop.everythingsmart.io/), and have been going through troubleshooting endless Zigbee issues: * Aqara kit struggled to pair with a well populated Zigbee mesh. * Tried pairing 0.5m from the Zigbee coordinator. * Added IKEA 13A UK mains socket switches to give a very close routing device. * Factory reset -> Add Device -> fail, rinse, repeat. * After pairing, all of the entity vales go to `Unavailable` greyed out * Pairing works, for a second, device entity values populate, then to to `Unavailable` * Zigbee Device **RECONFIGURE** starts, but doesn't ever complete. * Adding to the mesh coverage, waiting, Factory reset - no difference. * I tried a custom "quirks" file for the Aqara E1 curtain driver - looked visually better, but still really `Unavailable` so removed it. GitHub seemed to have some Aqara quirk updates in the last few months, but merged and (presumably) released. I then repeated the tests on a pre-production HA install (spare kit, same software) with an old Sonoff Zigbee coordinator - **all my Aqara products worked first time!** The test system is not perfect (T1 Battery `Unknown`, Curtain Luminance `0` - it is dark though!), however controls work and the main sensors report (better than `Unavailable`). Now, I've looked about the forum and found ZHA issues, Z2M issues, specific coordinator issues, specific coordinator firmware updates, but not quite this combination of OK/ FAIL. ### Steps to reproduce ## System 1 - FAILS :x: :red_circle: * NC Yellow + RPiCM4 * SkyConnect / ZBT-1 Firmware: 6.10.3.0 build 297 * HAOS 14.2 / Core 2025.2.1 * ZHA "zha": "0.0.48" / "zhaquirks": "0.0.132" * All controls and sensors greyed out or `Unavailable` ## System 2 - WORKS :heavy_check_mark: :green_circle: * RPi3b (a minimal pre-production test rig, so fits memory OK) * Sonoff 3.0 (TI CC2652) Firmware: Z-Stack 20210708 * HAOS 14.2 / Core 2025.2.1 * ZHA "zha": "0.0.48" / "zhaquirks": "0.0.132" * AOK Note: Both systems were tested under 0.0.48m but are not running 0.0.49. ## Aqara kit * 2x Aqara Zigbee Curtain Driver E1 Rod version * 1x Aqara Zigbee Temperature and Humidity Sensor T1 (also includes pressure) ### Other details * UK 240V 50Hz, no known RF issues. * Stable, working HAOS install. * HACS limited to Spook, Dyson and a front-end Horizon card. * Lots of Z-Wave, Lots of Zigbee, decent WLAN, ESPhome, Tasmota, custom kit. * [EST](https://shop.everythingsmart.io/) deliver very quickly (when Scotland isn't in the middle of a 1:100 year storm!). * Engineer :mage:, years of HA/ openHAB/ MisterHouse, etc. * No hair (pulled out after too many 5-second button holds for factory reset on a precarious chair! :man_bald: :face_with_symbols_over_mouth:). * Don't want to move from ZHA to Z2M, especially now the coordinator hardware seems to be a possible root-cause (and the Yellow has been otherwise great). * No delays nor latency on either Zigbee mesh (so far...) so the expected interference isn't showing up. * The test System 2 only has the 3x Aqara devices on the mesh, and paired immediately from many meters away (unlike the production System 1). Yes, there's probably more mesh traffic, but also a load more mains routers. Thanks for any insight you may have! ### Expected behavior ZHA works (fairly) independently of coordinator hardware: * Yellow/ SkyConnect/ ZBT-1 * Sonoff 3.0 (TI CC2652) ### Screenshots/Video Detailed debug available on request - help to narrow down test scenarios would be appreciated. ### Device signature I decided not to include screeds of logs at this stage in case issue triage spotted a known issue, and you might be able to point at specific test scenarios. Would be pleased to capture specifically useful scenarios - * Pairing with ZBT-1? * Re-interview with ZBT-1? (can make entity states worse than initial pairing) * Is successful pairing with the Sonoff 3.0 (TI CC2652) useful? * Anything else? (software / microelectronics engineer) ### Diagnostic information <details><summary>Diagnostic information</summary> ```json [Paste the diagnostic information here] ``` </details> ### Logs <details><summary>Logs</summary> ```python [Paste the logs here] ``` </details> ### Additional information Note: Both systems were tested under 0.0.48m but are not running 0.0.49. Would be happy to re-test with debug logging under 0.0.49. The Yellow SilLabs is my production system, the Sonoff a test rig (although suspect you'll need SilLabs logs form the Yellow).