Issue with 2018 SmartThings Water Leak Sensor (IM6001-WLP01) treated like 2015 version (F-WTR-UK-V2)

[airdrummingfool]

I have found a weird circumstance when pairing my IM6001-WLP01 SmartThings water leak sensor. I have verified that the device's firmware is fully up-to-date (version 0x11 / 17).

During pairing, it identifies itself as follows:

• Manufacturer: Samjin
• ModelID: water

When pairing is complete, Zigbee2MQTT treats it as a SmartThings F-WTR-UK-V2, which is a completely different device. When I short the sensors (to simulate a water leak event), zigbee2mqtt cannot interpret the resulting zigbee message, so the leak is not reported.

Interestingly, the F-WTR-UK-V2 is referred to as the "2018 model" by z2m, but it is actually a 2014 or 2015 model according to this SmartThings support article. I believe the device I have (the IM6001-WLP01) is the 2018 model. The zigbee2mqtt documentation website has the correct pictures for both the IM6001-WLP01 and the F-WTR-UK-V2.

I believe that one of two things has happened:

1. There are multiple different devices with the modelId water. This would be bad!
2. zigbee-herdsman-converter's devices.js has an error in it.

I do not know what to do if 1 has happened, and I'm hoping that Samsung wouldn't allow that kind of oversight. Thus, going forward, I will assume 2.

Here's what I've been able to gather:

• The code that associates modelId water with F-WTR-UK-V2 was added here: Add support for SmartThings Water Leak sensor (2018 model) #554. @Fabiancrg , if you still have this device, could you confirm both the model number printed on the device and the modelId as reported by the device (available in database.db)?
• I found a separate PR (add samsung 2018 water sensor #556) that adds code to handle the IM6001-WLP01, but with the (ostensibly incorrect) modelId IM6001-WLP01 instead of water. The issue that PR references shows that the modelId is water, but there is a commit in the PR that changed the modelId before it was merged. @johntdyer can you check your device's model number, and the modelId it reported to zigbee2mqtt? Do you remember why you changed the modelId before submitting the PR?
• The related issue, Samsung SmartThings Water Leak Sensor (2018)  zigbee2mqtt#663, was filed by @pfischbach. @pfischbach, is your SmartThings water leak sensor IM6001-WLP01 working properly with the newest version of zigbee2mqtt?
• @skandragon, you have 3 of the same devices I do (according to Unable to fully configure a Smartthings water sensor #713), can you verify that they paired as F-WTR-UK-V2 and check and see whether or not the water leak reporting works?
• Other versions:
  • STS-WTR-250 identifies itself as modelId moisturev4.
  • F-WTR-US-2: I can't find any references to its self-reported modelId, but this might be the US analogue for the F-WTR-UK-V2.

In order to make my device work, I would have to change the modelId of water for the existing devices.js entry for the F-WTR-UK-V2 (and replace the modelId of IM6001-WLP01 for the existing IM6001-WLP01 entry). I don't know what the correct modelId for the F-WTR-UK-V2 is, and I certainly don't want to break any existing device configurations.

[Koenkk]

Thanks for this throughout investigation, hopefully the people will respond so it can be solved.

[johntdyer]

@airdrummingfool ... Yea, I dont recall why I picked that modelID, I kinda pieced this together from other PR's and other models... It would not surprise me if I messed something up, Node is far from my strongest language

[johntdyer]

@airdrummingfool - Its possible I Cargo cult'ed it from here

[johntdyer]

@airdrummingfool IMHO you should go ahead w/ your suggestion since you seem to have a firmer grasp of the situation then I do at this moment

[airdrummingfool]

@johntdyer thanks for your responses. I've been testing some modifications to support the sensor I have on hand correctly, and I'll be submitting a PR soon.

Would it be possible for you to check database.db for the actual modelId reported by the sensor you have (or upload your entire database.db)? This will help me feel more confident that I won't be breaking compatibility for any existing sensors.

[achurak]

Hello,

I have exactly the same problem. I have several newer leak sensors (bought in US), but they are treated as F-WTR-UK-V2 for some reason (though it does say 2018 model). It also doesn't recognize the "contact" events:

zigbee2mqtt | zigbee2mqtt:debug 2019-12-18 12:56:09: Received Zigbee message from '0x286d970001016dcd', type 'commandStatusChangeNotification', cluster 'ssIasZone', data '{"zonestatus":33,"extendedstatus":0}' from endpoint 1 with groupID 0

zigbee2mqtt | zigbee2mqtt:debug 2019-12-18 12:56:09: No converter available for 'F-WTR-UK-V2' with cluster 'ssIasZone' and type 'commandStatusChangeNotification' and data '{"zonestatus":33,"extendedstatus":0}'

Here's the database entry for one of these sensors:

{"id":2,"type":"EndDevice","ieeeAddr":"0x286d970001016dcd","nwkAddr":36939,"manufId":4673,"manufName":"Samjin","powerSource":"Battery","modelId":"water","epList":[1],"endpoints":{"1":{"profId":260,"epId":1,"devId":1026,"inClusterList":[0,1,3,32,1026,1280,2821],"outClusterList":[3,25],"clusters":{"genBasic":{"attributes":{"modelId":"water","manufacturerName":"Samjin","powerSource":3,"zclVersion":2,"appVersion":17,"hwVersion":0,"swBuildId":""}}},"binds":[{"cluster":1026,"type":"endpoint","deviceIeeeAddress":"0x00124b00194a20da","endpointID":1},{"cluster":1,"type":"endpoint","deviceIeeeAddress":"0x00124b00194a20da","endpointID":1}]}},"appVersion":17,"hwVersion":0,"swBuildId":"","zclVersion":2,"interviewCompleted":false,"meta":{"configured":1},"lastSeen":1576691619459}

Let me know if you need any additional information.

[airdrummingfool]

@achurak thanks for the information. I had the same issue with contact events on my sensor, and have fixed it in my local branch. I'll be submitting a PR in the next couple days that includes the fix for both the contact sensor and the identification of the device.

[Koenkk]

Closing this as #833 has been merged, thank you very much @airdrummingfool for figuring this out.