IKEA TRADFRI 5 button remote: no action event (E1524/E1810 Firmware 2.3.080, Adapter: RaspBee)

[rtitmuss]

What happened?

No action events are reported by z2m when pressing any button on the IKEA TRADFRI 5 button remote, when using the RaspBee coordinator and 2.3.080 firmware on the tradfri remote.

What did you expect to happen?

Action events are reported by z2m when pressing the buttons on the IKEA TRADFRI 5 button remote.

How to reproduce it (minimal and precise)

1. Remove the tradfri remote from z2m
2. Restart z2m
3. Enable permit join
4. Press the reset button on the remote 4 times
5. Wait for the remote to be interviewed
6. Disable permit join
7. Press the buttons on the remote

The remote is shown as bound to the Coordinator as shown in the screenshot below.

Note the remote has joined the network. Availability and battery status reports are working correctly. The remote can be bound to light bulbs. The only problem is no action events are emitted when the buttons are pressed.

I have tried multiple time unbinding and rebinding the Coordinator in the frontend. This does not fix the issue.

This is working correctly with a Philips Hue Remote and IKEA TRADFRI with firmware 1.2.223.

It does not appear to be related to #5350, as I don't see the unknown device with address error.

I have also tried binding using mqtt as described in #9439. The binding is successful but still no action events.

Zigbee2MQTT version

1.25.1

Adapter firmware version

0x26400500

Adapter

RaspBee

Debug log

Attached zigbee-herdsman logs from steps 3-7 above:

z2m.log

[MattWestb]

I think DE have not porting the last fixes from Rasp/CornBee II to first gen that is needed for getting IKEA remotes working OK.
Improve network robustness for various devices from Ikea and Osram.
Look on DE wiki https://github.com/dresden-elektronik/deconz-rest-plugin/wiki/Firmware-Changelog

[cadavor]

Hi,

I've got the same problem with one of my 4 E1810 remote that I updated 4 days ago...
No more "action" event is received.
I tried to change the battery, unpair/repair a lot of times.

The other 3 (not updated) work well.

I use a CC2531 adapter updated with lastest firmware (20211115)

[AlexKay88]

Did you disable legacy integration?

To disable the click event, set legacy: false for this device in configuration.yaml. Example:

devices:
  '0x12345678':
    friendly_name: my_device
    legacy: false

or via web gui:

[Shamane2000]

I have two of this ikea devices and also a conbee1 here. (integrated in HASSIO)
It makes no difference if i set it to true or false. No Messages from this tradfri device. :(
Any news regarding this?

[Shamane2000]

in addition to that, it seems to not working with more ikea devices. the same behavior appear with this shortcutbutton: https://www.zigbee2mqtt.io/devices/E1812.html#ikea-e1812
under "logs" in the zigbee2mqtt frontend is also nothing reportet, if i press any key (from tradfri buttons) An aquara switch ist working well.

[ministryofsillywalks]

Happening to me aswell using ezsp coordinator.
All my IKEA Remotes can be bound but send no more actions.
It pairs perfectly fine. LQI and Battery are being shown but no actions show up in the log.
Has anyone figured out whats going on?

[flechaig]

I have the Conbee stick (first version) and the same issue with IKEA remotes :(

[flechaig]

I just figured it out...
As state in zigbee2mqtt documentation, the pairing MUST take place in contact with the controller.
It may need a few tries to pair correctly, but it seems that it works well if the pairing is done as advised.

[amaximchuk]

I have the same issue with the IKEA 5-buttons remote. But with the Conbee II & Deconz works well.

[ahd71]

+1 on the same issue. Have a number if them and some few works, most doesn't anymore all are on the same firmware release

[lbschenkel]

Same thing here. I was migrating from deCONZ to Z2M and none of my remotes work (they are all with 2.3.014 firmware) but they all work in deCONZ. Using a Conbee (original model, not II). Debug logs show no activity when buttons are pressed.

[lbschenkel]

Follow up: this is what happens when pressing buttons:

2022-09-08T11:35:30.093Z zigbee-herdsman:deconz:frameParser DEVICE_STATE changed: 10101010
2022-09-08T11:35:30.093Z zigbee-herdsman:deconz:driver networkstate: 2 apsDataConfirm: 0 apsDataIndication: 1 configChanged: 0 apsRequestFreeSlots: 1
2022-09-08T11:35:30.097Z zigbee-herdsman:deconz:driver query aps data indication
2022-09-08T11:35:30.099Z zigbee-herdsman:deconz:driver DATA_INDICATION - sending read data request - SeqNr. 59
2022-09-08T11:35:30.125Z zigbee-herdsman:deconz:frameParser DATA_INDICATION RESPONSE - seqNr. 59 srcAddr: 0xeaf8 destAddr: 0x0 profile id: 0x104 cluster id: 0x5 lqi: 255
2022-09-08T11:35:30.125Z zigbee-herdsman:deconz:frameParser response payload: 5,124,17,31,7,1,1,13,0
2022-09-08T11:35:30.133Z zigbee-herdsman:controller:log Received 'zcl' data '{"frame":{"Header":{"frameControl":{"frameType":1,"manufacturerSpecific":true,"direction":0,"disableDefaultResponse":false,"reservedBits":0},"transactionSequenceNumber":31,"manufacturerCode":4476,"commandIdentifier":7},"Payload":{"value":257,"value2":13},"Command":{"ID":7,"parameters":[{"name":"value","type":33},{"name":"value2","type":33}],"name":"tradfriArrowSingle"}},"address":60152,"endpoint":1,"linkquality":255,"groupID":0,"wasBroadcast":true,"destinationEndpoint":53}'
2022-09-08T11:35:30.133Z zigbee-herdsman:controller:log 'zcl' data is from unknown device with address '60152', skipping...

2022-09-08T11:35:32.510Z zigbee-herdsman:deconz:frameParser DEVICE_STATE changed: 10101010
2022-09-08T11:35:32.510Z zigbee-herdsman:deconz:driver networkstate: 2 apsDataConfirm: 0 apsDataIndication: 1 configChanged: 0 apsRequestFreeSlots: 1
2022-09-08T11:35:32.513Z zigbee-herdsman:deconz:driver query aps data indication
2022-09-08T11:35:32.518Z zigbee-herdsman:deconz:driver DATA_INDICATION - sending read data request - SeqNr. 61
2022-09-08T11:35:32.542Z zigbee-herdsman:deconz:frameParser DATA_INDICATION RESPONSE - seqNr. 61 srcAddr: 0xeaf8 destAddr: 0x0 profile id: 0x104 cluster id: 0x6 lqi: 255
2022-09-08T11:35:32.542Z zigbee-herdsman:deconz:frameParser response payload: 1,32,2
2022-09-08T11:35:32.546Z zigbee-herdsman:controller:log Received 'zcl' data '{"frame":{"Header":{"frameControl":{"frameType":1,"manufacturerSpecific":false,"direction":0,"disableDefaultResponse":false,"reservedBits":0},"transactionSequenceNumber":32,"manufacturerCode":null,"commandIdentifier":2},"Payload":{},"Command":{"ID":2,"parameters":[],"name":"toggle"}},"address":60152,"endpoint":1,"linkquality":255,"groupID":0,"wasBroadcast":true,"destinationEndpoint":53}'
2022-09-08T11:35:32.547Z zigbee-herdsman:controller:log 'zcl' data is from unknown device with address '60152', skipping...

Wild guess here, but could the problem be zigbee-herdsman:controller:log 'zcl' data is from unknown device with address '60152', skipping...? so the "packets" (not sure what the right terminology is) are being discarded and not processed further?

Will try to take a look at the source code to find out more.

[lbschenkel]

@Koenkk : the interesting thing is that even if I rejoin the remote, and the NWK changes (as it should), all zcl data keeps being received with "address":60152. According to https://github.com/Koenkk/zigbee-herdsman/blob/v0.14.53/src/controller/controller.ts#L601, this should be the network address (NWK), which is definitely not what's reported by the web interface nor what's stored in database.db.

60152 is 0xEAF8 and this remote's IEEE address is 0x..........eaf8... That can't be a coincidence, can it, especially if the address never changes on rejoins? But then the packet should have been logged with "address":"..." since the value should have been a string to trigger lookup by IEEE address instead of NWK.

[lbschenkel]

@Koenkk:

I did a packet capture and the actual source is 0x9FA9 which matches the NWK in the database:

Therefore it seems that there's a bug in the packet decoding logic in Z2M, because it is incorrectly parsing the source address as 0xEAF8. Probably around here? https://github.com/Koenkk/zigbee-herdsman/blob/b4cdf8ba6935ca20aa6967e1a0afed32f320e64b/src/adapter/deconz/driver/frameParser.ts#L241

[ahd71]

Some of my remotes work, some doesn't.same model and firmware. Haven't done any packet trace.

[lbschenkel]

The packet sniffing show that the remote is sending the right command, but it's being misinterpreted by the deconz driver in Z2M and it has a wrong source address. Since Z2M can't find the device it came from, it drops the packet. That's why nothing happens and nothing gets logged.

[ahd71]

Great Finding! Shouldnt be that hard to identify and fix then. Can I help in any way?

[github-actions]

This issue is stale because it has been open 30 days with no activity. Remove stale label or comment or this will be closed in 7 days

[ahd71]

But problem still exist. Actually the only severe problem I have in my large (200 devices) installation at the moment.
If I can help out (e.g with testing) or anything else, please let me know.

[ahd71]

The problem perhaps isn't in z2m but in a upstream system/driver. How do we identify that so it can be addressed to the right people?

[lbschenkel]

Please don't close this. I've done a lot of troubleshooting, debugging and submitted the problem at Koenkk/zigbee-herdsman#576 but I have heard no response from the maintainers so far.

[lbschenkel]

The problem perhaps is that in z2m but in a upstream system/driver. How do we identify that so it can be addressed to the right people?

I already did. All details are submitted at Koenkk/zigbee-herdsman#576

[ahd71]

Very good analysis! Shouldn't be too hard for someone written those code parts to adress it.

[github-actions]

This issue is stale because it has been open 30 days with no activity. Remove stale label or comment or this will be closed in 7 days

[pjcarly]

Happening here as well. Original Conbee stick, 5-button remote.

[github-actions]

This issue is stale because it has been open 30 days with no activity. Remove stale label or comment or this will be closed in 7 days

[pjcarly]

bump

[mozzbozz]

I have two of these Ikea Tradfri Remotes and use the Raspbee shield (1st version). The remotes were on really old firmware (1.2.214). Both showed only N/A on every button press. So I went and updated them without a problem to 2.3.080.

After pressing the reconfigure button (after the update; I've tried it before with no luck), they started doing something: Now the toggle button (the middle button) actually sends a signal. But the other buttons still don't do anything.

This problem definitely seems to be connected to the raspbee/conbee hardware.

[konstantint]

The button worked for ages, then at some point just stopped sending click actions for seemingly no reason. Firmware 2.3.014, I don't think OTA ever succeeded, although I tried.

It works with ZHA without any problems, though.

The coordinator is a Sonoff ZBDongle-E.

[konstantint]

... more news. I managed to force an OTA on the button (firware version became 24.4.5) and after a couple of "force remove / rejoin" iterations the "toggle" action started coming through.

[netweaver1970]

My 5button ikea switch was generating action events in Z2M gui but HA doesn't see them. Was this also your problem? My ultimate goal was to use ControllerX to emulate the normal behavior.

[ahd71]

In my case z2m didnt show it either (so not a HA issue). But with the latest firmware in Buttons and in the zzh stick and the z2m release everything works great!

[github-actions]

This issue is stale because it has been open 30 days with no activity. Remove stale label or comment or this will be closed in 7 days

[ALCD-clio75]

Any solution on this issue ?