Sonoff Dongle Max - drowning in error messages

Configuration Zigbee
1

Any log interpreters out there.

Migrated from a smlight to a Sonoff dongle max today.

The smlight was ok but required me to re-pair 10 of the 50 devices daily with another 1 or 2 that were reporting but unresponsive.

Migrated successfully, but a couple later had devices falling offline and it is gaining on me, I cannot keep up with them, I get them working then 30 mins later they go offline again.

Looking the zigbee logs I am getting

at EmberOneWaitress.deliveryFailedFor (/app/node_modules/.pnpm/[email protected]/node_modules/zigbee-herdsman/src/adapter/ember/adapter/oneWaitress.ts:100:31)
at EmberAdapter.onMessageSent (/app/node_modules/.pnpm/[email protected]/node_modules/zigbee-herdsman/src/adapter/ember/adapter/emberAdapter.ts:458:42)
at Ezsp.emit (node:events:518:28)
at Ezsp.ezspMessageSentHandler (/app/node_modules/.pnpm/[email protected]/node_modules/zigbee-herdsman/src/adapter/ember/ezsp/ezsp.ts:5076:14)
at Ezsp.callbackDispatch (/app/node_modules/.pnpm/[email protected]/node_modules/zigbee-herdsman/src/adapter/ember/ezsp/ezsp.ts:734:26)
at Ezsp.onAshFrame (/app/node_modules/.pnpm/[email protected]/node_modules/zigbee-herdsman/src/adapter/ember/ezsp/ezsp.ts:387:22)
at UartAsh.emit (node:events:518:28)
at Immediate. (/app/node_modules/.pnpm/[email protected]/node_modules/zigbee-herdsman/src/adapter/ember/uart/ash.ts:1192:45)
at processImmediate (node:internal/timers:485:21))
[2025-10-15 17:13:06] error: z2m: Failed to configure ‘Porch/Plant/Poinsettia/Smart Watering’, attempt 1 (Error: Bind 0x282c02bfffeef62f/1 genPowerCfg from ‘0x54dce9fffe0998c6/1’ failed (Delivery failed for ‘12490’.)
at EmberOneWaitress.deliveryFailedFor (/app/node_modules/.pnpm/[email protected]/node_modules/zigbee-herdsman/src/adapter/ember/adapter/oneWaitress.ts:100:31)
at EmberAdapter.onMessageSent (/app/node_modules/.pnpm/[email protected]/node_modules/zigbee-herdsman/src/adapter/ember/adapter/emberAdapter.ts:458:42)
at Ezsp.emit (node:events:518:28)

Just getting pairing mode on takes between 10-15 button presses.

Any ideas anyone?2025-10-14T23:00:00Z

2

Did you clone the IEEE?

3

Nope :frowning:

I just followed the instructions :frowning: zstack to Ember instructions

I have now - will see if that helps - and thanks for coming back with that much appreciated

4

I am running with much reduced zigbee load, down to 50 25 mains powered 25 battery.

The mains powered ones are all (mostly) offline with the battery devices largely showing online and checking in.

I have been working on the issue of unreliable zigbee for two years.

Have confirmed the chosen zigbee channel 11is clear of WiFi and hue zigbee (channel 26).

With this the 6th zigbee coordinator I remain hopeful.

5

The devices, mainly mains powered remain paired for less than 30 mins, meaning that to keep this small zigbee network running I have to be constantly pairing 24 hours a day.

I’ll try wiping zigbee2MQTT again and building from the ground up again, I have had to do this many many times and getting quite good at it.

I just wish zigbee2MQTT wasn’t as flakey :frowning:

I know there are people out there who have several hundred devices running for decades without an issue, but unless their complete setup is up for sale with a guarantee it will work, it doesn’t really help.

6

This is just weird.

18 hours on.

All the Aurora wall sockets and standalone sockets/plugs are showing as offline, ie 90% of my mains powered devices.

They are currently showing last seen as 12 hours ago.

BUT, I can turn the sockets on off, and can set the brightness of the status leds!

Not bad for devices that have been offline for over 12 hours.

Also weird is attempting to put the bridge into Pairing mode can take up to 20 button presses while receiving Broadcast error ‘busy’ messages.

Also

IMG_8521
IMG_85212360×1640 239 KB

The Disable Join going into negative numbers is weird.

7

How are you doing now? Is your Zigbee network working properly?

8

I did a wipe and rebuild of zigbee2MQTT .

Gone back to a Sonoff Dongle P on USB, without using poe can’t locate the coordinator in the best location so have to settle.

I am seeing much lower LQIs

BUT

I woke up this morning and for the first time in 22 months ALL 67 devices showing online and responsive!!

But just noticed one Mains device offline - normally a very chatty device checking in every 30 seconds or as an Energy/powers reporting device does and I am seeing an Aqara blind controller not reporting for 5 hours same with a Third Reality plant sensor.

So zigbee2MQTT still unstable :frowning:

Funny thing is that `hue has been running for nearly 15 years with twice as many devices without an outage but zigbee2MQTT struggle to get through the day.

I know that zigbee channel 11 is free and clear but may move it to channel 20 now I have moved the two hue bridges into one on 25.

9

All the devices went offline/became unresponsive eventually.

So took the plunge and moved WiFi fro 6/11 to 1/6 and moved zigbee2MQTT from 11 to 20.

Had to re-pair everything again. A common requirement unfortunately with zigbee2MQTT.

Will now see if zigbee2MQTT after a 22 month battle will eventually stabilise.

A Hue bridge used channel 20 without issue for years without one issue and with moving WiFi away from it even further there should be no excuse.

Over the last 22 months I have wiped and rebuilt Home Assistant 4 times on three different platforms.

I have performed nearly 3,500 zigbee pairing, re-pairing over 5 devices a day.

I have wiped and rebuilt zigbee2MQTT 8 times using 4 different coordinators.

The only thing that has kept me going is the knowledge that a stable zigbee2MQTT is possible because of the number of people who have told me as much.

I am looking forward to it being stable and enabling mature automations similar to the ones I have made using Eufy, Tado, Hue, Govee, Shelly, Zwave, Victron, Nanoleaf, MyEnergi, which have all been stable throughout this.

In fact because zigbee2MQTT has been so unreliable, I have replaced all the zigbee leak sensors, temp sensors, door/window sensors with reliable Eufy and Hue devices, who have all remained solid.

There are four things I use zigbee devices for,

Power/Energy usage reporting (all the sockets in the house are Aurora ones),

blind controllers (they control all the window coverings throughout the house),

Presence Detection where a simple motion sensor just won’t cut it

And switches/buttons for where human override is required.

If zigbee2MQTT gets stable then I can increase the switches/buttons more.

10

Nah that failed

at 16:58 every single zigbee2MQTT device died and all the mains powered devices are showing offline and all the rest are unresponsive.

This is a consistent issue for 22 months across Home Assistant running in a VM on Mac, Home Assistant Green, and now Home Assistant running on a massively overspeed micro PC.

It is consistent across multiple coordinators Sonoff P, SMLight snz06, Sonoff Donlge MAX

11

Could there be significant 2.4G electromagnetic interference in your home environment?

There are many factors that can cause Zigbee2MQTT instability, including host hardware, sub-devices (some devices frequently send garbage data leading to crashes), sub-device density, and network environment. I even found that the activation of the range hood at home can also interfere with the 2.4G network.

Below is the corresponding relationship between Wi-Fi channels and Zigbee channels. If you switch your Wi-Fi to channel 1 or 6, I suggest you change the Zigbee channel to 20 or above.

image
image617×316 65.7 KB

By the way, you can also use the Channel Energy Scan function of the Dongle Max to test the interference in your network environment.

image
image963×589 29.8 KB

12

Using a microwave oven is often very disruptive for anything using 2.4ghz so don’t put any routers or gateways near your kitchen. I’ve also found that channel 15 works good for zigbee.

My zigbee router is starting to have issues when I go above 68 devices so I’m going to upgrade to the dongle Max. I have a couple of power outlets that report multiple times a second the power usage, which I might replace with something that’s more reasonable.
FYI I use Zigbee2MQTT, and it works great (since moving to channel 15)

13

Thanks,

But not the issue, have been down that path :slight_smile:

14

Thanks, but not Microwave.
I have resolved using a usb sonoff p and an smlight Ethernet coordinators and finally got stable zigbee2MQTTs running, well for 2 weeks so far, the longest stable stretch for over 22 months :slight_smile:

15

Feeling your pain I basically have the same issue. On my 5th coordinator the new sonoff max and was hoping that would fix my 100+ device issues.

After a migration I was having issues with aquara some sockets, they would show offline after a while to the point all of them would go offline and I could not control them via home assistant. Repairing did not help nor did forcing pairing to the coordinator.

Full rebuild and a switch to Channel 20 and I am in a better place, however having an issue where ALL devices go ‘offline’ for a few seconds periodically. I am starting to think it’s a sonoff max issue as pretty much have eliminated everything else.

Generally the network is back up after but it does give a few quirks with the odd device becoming unresponsive and going offline. No idea the cause but it’s making me configure a switch back to smlight coordinator.

Well let you know if I ever achieve the dream :sleeping:

16

I ended up returning the Dongle Max.

Split my 60 devices between two coordinators, sonoff P usb and a Poe Ethernet smlight 06.

Been solid for several weeks.

I had been experiencing weird zigbee issues for 22 months, where I would lose 5 devices a day and every 10 days or so every device would be offline.

I rebuilt zigbee2MQTT many times and Home Assistant several times each very vanilla.

Splitting it out has stabilised it.

Allowing me to move forward to complex automations for the first time in nearly 2 years :slight_smile:

17

Super interesting, so you have 2 coordinators running on 2 separate channels. Do you split like upstairs and downstairs or similar? So issues with WiFi or interference?

Do you need to run 2 instances of z2m?

18

Only went to two coordinators after 22 months of trying everything else :slight_smile:

The house is built with heavy duty cement walls, all downstairs walls are the same and trash wireless signals :slight_smile: upstairs is a wood frame so much easier.

As the house is divided between front and back by a hallway with these heavy duty walls, dividing zigbee2MQTT into front and back made sense and is working.

The house Wi-Fi , with 5 access points uses 1 & 6 freeing up the higher zigbee channels.

I have recently consolidated the two hue zigbee hubs into the Hue Pro, freeing up another common zigbee2MQTT channel.

Right now, for the last 4 weeks it has been stable, so hopeful.

I am now very slowly building more devices back on, from 30 devices ultimately to 50 devices per zigbee2MQTT.

But only adding a couple per week to each.

I did do an audit, and removed excessively noisy devices that were generating too many messages. That may have contributed.

Odd though that the noisy devices I had to remove are also used around the house but aren’t so chatty.

At some point will spend some time to determine the difference.

1 Like
19

I’m feeling similar pain after thinking the Dongle Max would solve my problems. Was using a Dongle P and just migrated. LQI is awesome but logs tell me Max Concurrency is being met causing backlogs of messages from spammy power plug devices. For some reason max concurrency is set at 2 and the network is dying.

Trying to remove and/or fix spammy devices to see if it helps. Unfortunately I don’t really want to toss out my Tuya energy monitoring smart plugs so I’m seeing if I can create a custom quirk to reduce chatty messaging.
Gemini is my new best friend at analysing logs and providing solutions. Definitely load your logs and see what it tells you