[Matter] Connecting Google home app

[taoyinzhou]

Hi all，

I am developing the door lock using matter. The SDK uses ncs2.1.0, nrf/samples/matter/lock, and I want to test the connection to google home app on Android phone.

Information of test equipment:

52840dk (End Device)

google nest hub (Border Router)

pixel 4xl + Google Home App (Matter controller)

I first tried the example of door lock, but it failed to bind successfully in the app. Then I tried nrf/samples/matter/light_bulb, and the test result showed that it could be bound by app, and it could also control the light on/off through app.

So why are the results different between the light_bulb sample and the lock sample? Has anyone else tried it?

Attach the failed print log.

2.1.0-lock-google.txt

[AHaug]

Hi,

Just to verify, did you do the sample testing with out of the box samples without modifications?

And did you see the non-connecting behavior the Light switch sample as well? If you've not tested this yet, could you try this sample with your configuration? The switch and the lock are both openThread Sleepy End Devices, so they might share the same behavior, while the bulb is acting as a router. 

And if you have a log for the functioning light bulb as well as the light switch after you've tried that, could you supply them as well?

Kind regards,
Andreas

[taoyinzhou]

Only the DISCRIMINATOR in chip_project_config.h has been changed.

I tried the Light switch sample as you suggested, but failed to connect with the app. The error displayed on the app was the same.

Attach a log of light bulb and the light switch.

2.1.0-light-google-success.txt

2.1.0-switch-google-fail.txt

[AHaug]

Hi, 

Thank you for testing the other switch sample as well and adding the logs.

Have you created an intration in Google Developer Center (to specify device's VID and PID)? Creating such integration is mandatory.

Link: https://developers.home.google.com/matter/integration/create. 

Currently we're looking into why the light bulb (VID/PID 0xFFF1/0x8005) works without such integration, and  why the Lock (0xFFF1/0x8006) does not, but we believe this integration might be what causes the issue

Kind regards,
Andreas 

[taoyinzhou]

Thank you very much for your help.

After I added the PID and VID in Google Developer Center as you suggested,my device could connect to the google home app. 

I have a question. The device type of matter has door lock(0x000A), but there is no lock type. Why is pid divided into door lock(0x8003) and lock(0x8006)? Why is the pid in the ncs example lock(0x8006) and not door lock(0x8003)?

[AHaug]

Hi,

Glad to hear you got it to work!

taoyinzhou said:

The device type of matter has door lock(0x000A), but there is no lock type. Why is pid divided into door lock(0x8003) and lock(0x8006)? Why is the pid in the ncs example lock(0x8006) and not door lock(0x8003)?

PID is purely given by the manufacturer. It does not indicate any device types (only ZCL and Data Model does). 0x8006 or 0x8003 are just example PIDs for SDK examples.

Kind regards,
Andreas

[Dan]

Hi Taoyin - Appreciate you sharing your experience here. I am curious whether you are using the "vanilla" NRF Connect as installed by Nordic's NRF Connect for Desktop, or whether you had to also install the patch that Google suggests on this page: https://developers.home.google.com/matter/vendors/nordic-semiconductor

"Download and apply the google-matter-preview patch, which contains fixes specific to the interop_testing_te9"

Asking because I am still struggling to get basic matter commissioning working through the Google Home App (although I am able to get it to work fine using the Android Chiptool App with an EERO6 as an OTBR). As an aside, I am using the NRF5340DK.

Thanks! Dan.

[AHaug]

Hi Dan,

In case the case creator does not reply, I would like to recommend you to create your own ticket so it will be assigned to, and answered by, one of the application engineers working with answering cases here at Devzone. Please feel free to refer to this case if you think it is relevant for your case.

Kind regards,
Andreas

[Dan]

Thanks Andreas - already done here.