TS0505B bulbs become unresponsive and ZCL commands timeout on Ember coordinator, but work correctly with Tuya Gateway

[ilyakorolev2]

What happened?

What happened?

I am using three MOES ZB-TD6-RCW-GX53-MS (TS0505B) Zigbee bulbs with Zigbee2MQTT.

The bulbs frequently stop responding to ON/OFF commands and Zigbee2MQTT reports:

ZCL command ... timed out after 10000ms

Sometimes, after several minutes, the bulbs suddenly execute a backlog of previously sent commands.

The issue occurs even when controlling a single bulb directly (not a group).

What did you expect to happen?

The bulb should respond immediately to ON/OFF commands without timeouts.

How to reproduce it

1. Pair ZB-TD6-RCW-GX53-MS (TS0505B).
2. Send repeated ON/OFF commands.
3. After several commands, timeouts start appearing.
4. Commands may execute minutes later.

Zigbee2MQTT version

2.12.0

Zigbee-herdsman version

10.4.0

Zigbee-herdsman-converters version

26.63.0

Coordinator

EmberZNet

Coordinator firmware version

7.4.5.0

Additional information

• LQI is typically 170-180.
• The problem occurs with any of the three bulbs.
• The problem occurs with a single bulb and with groups.
• Home Assistant automations are not involved.
• The same bulb works perfectly through a Tuya Gateway. I tested more than 100 consecutive ON/OFF operations without any issue.
• Other Zigbee devices in the network work normally.

What did you expect to happen?

No response

How to reproduce it (minimal and precise)

No response

Zigbee2MQTT version

2.12.0 (unknown)

Adapter firmware version

{"build":0,"ezsp":13,"major":7,"minor":4,"patch":5,"revision":"7.4.5 [GA]","special":0,"type":170}

Adapter

EmberZNet

Setup

os: #1 SMP PREEMPT_DYNAMIC Wed May 6 08:19:51 UTC 2026 - 6.12.85-haos - x64
node: v22.22.2
ha: true

Device database.db entry

No response

Debug log

No response

Notes

Last health check

time: Fri Jun 19 2026 00:58:08 GMT+0300 (Москва, стандартное время)
os.load_average: 0.08, 0.08, 0.04
os.memory_percent: 31.0112
process.memory_percent: 4.5498
process.uptime_sec: 7204

[andrei-lazarov]

I have also seen this issue. I believe it's caused by Z2M polling all Telink bulbs. I'll see what I can do