Tuya Smart Plug missing metering data (_TZ3000_cphmq0q7)

[muratcorlu]

Describe the bug

I bought 4 Tuya Zigbee smart plugs and easily connected to Phoscon (running with Conbee II on Rpi4). I use them on my HomeKit setup via Homebridge. One of them worked perfectly with metering functionality but 3 of them doesn't report metering data. Working one has _TZ3000_8nkb7mof vendor name and others has _TZ3000_cphmq0q7.

Steps to reproduce the behavior

1. Start pairing a smart plug with _TZ3000_cphmq0q7 vendor name.
2. Pair with with Phoscon GW
3. Restart HomeBridge
4. Check if you have consumption data on Eve Home app. (Or in HomeAssistant)

Expected behavior

I expect to see consumption data on Eve Home App or HomeAssistant.

Screenshots

Here is the list of devices in Phoscon App:

Detail screen of working one:

Detail screen of one of the non working devices (be aware this has version info):

Eve Home screenshot of working device:

Eve Home screenshot of one of the non-working devices:

Environment

• Host system: Raspberry Pi
• Running method: Raspbian / HomeBridge / homebridge-hue plugin / Eve Home App
• Firmware version: 266b0700
• deCONZ version: 2.10.04
• Device: ConBee II
• Do you use an USB extension cable: no
• Is there any other USB or serial devices connected to the host system? If so: P1 adapter, USB Bluetooth

deCONZ Logs

I use headless deconz. I currently couldn't figure out how to fetch debug logs. I'll update if I can.

Additional context

[NijsKei]

I have exactly the same problem. Why it always must be difficult. 2 are working well a 2 are working not. Only the switches are doing their job but no measering.

[Mimiix]

Unfortunately,this is the downside of having cheaper devices.

The support was added to the one that works. The others have new chips/manufacturer names. That needs to be whitelisted.

Can you provide screenshots from deconz as a device request?

[oywino]

I have exactly the same problem.
Setup:
Conbee Firmware: 266B0700
deConz Version: 2.11.02 beta / 4/30/2021

Device is registered in Phoscon as a Light. No measuring data is present. On/Off works fine.
Here are my deConz screen dumps:

On/Off Cluster:

Simple Metering Cluster:

REST-API:

{
    "etag": "db47fd907f3381714dc29b86101ac2f3",
    "hascolor": false,
    "lastannounced": null,
    "lastseen": "2021-05-19T18:38Z",
    "manufacturername": "_TZ3000_cphmq0q7",
    "modelid": "TS011F",
    "name": "On/Off Tuya 1",
    "state": {
        "alert": "none",
        "on": true,
        "reachable": true
    },
    "swversion": "67",
    "type": "On/Off plug-in unit",
    "uniqueid": "84:71:27:ff:fe:8d:96:1d-01"
}

When I manually press the button on the device ( from On --> Off ) I get this event:

{
    "20:50:43:832": {
        "attr": {
            "id": "18",
            "lastseen": "2021-05-19T18:50Z",
            "manufacturername": "_TZ3000_cphmq0q7",
            "modelid": "TS011F",
            "name": "On/Off Tuya 1",
            "swversion": "67",
            "type": "On/Off plug-in unit",
            "uniqueid": "84:71:27:ff:fe:8d:96:1d-01"
        },
        "e": "changed",
        "id": "18",
        "r": "lights",
        "t": "event",
        "uniqueid": "84:71:27:ff:fe:8d:96:1d-01"
    },
    "20:50:43:873": {
        "e": "changed",
        "id": "18",
        "r": "lights",
        "state": {
            "on": false,
            "reachable": true
        },
        "t": "event",
        "uniqueid": "84:71:27:ff:fe:8d:96:1d-01"
    }
}

[Smanar]

For me the code is blocking for the reason exposed by @muratcorlu, the device have

• No support attribute 0x4000
• Empty value for 0x0006

And the code don't like that, there is a hack for light device but not for sensor.
There is many way to help

Ugly one like

        if (node->nodeDescriptor().manufacturerCode() == VENDOR_SAMJIN)
        {
            swBuildIdAvailable = false; // empty string
            dateCodeAvailable = false; // unsupported attribute
        }
        else if (modelId == QLatin1String("HG06323")) // LIDL Remote Control
        {
            swBuildIdAvailable = false; // unsupported attribute
            dateCodeAvailable = false; // empty string
        }

But for me it s possible with using dateCode.isNul(), but in this situation we will have swversion with null value.

Need to ask first the autorisation at others devs.

[oywino]

@Smanar, thanks for stepping in - but what you just said is "gibberish" to me. Could you "translate" it into layman's English? 😊

[Smanar]

Lol.
I have a code to test

sudo apt install deconz-dev
git clone --branch tuya_plug_1 https://github.com/Smanar/deconz-rest-plugin.git
cd deconz-rest-plugin
qmake && make -j2
sudo cp ../libde_rest_plugin.so /usr/share/deCONZ/plugins

This code (if it work) try to get the Swbuild, then the date code and finaly the Application version to fill the "swversion" field.

You will have in log with "info"

DBG_Printf(DBG_INFO, "[4.1] Get sw build id\n");
DBG_Printf(DBG_INFO, "[4.1] Get date code\n");
DBG_Printf(DBG_INFO, "[4.1] Get Application version\n");

And after the device will be included.

[oywino]

Sorry, I don't understand.
The issue in this thread is: Tuya Smart Plug missing metering data (_TZ3000_cphmq0q7)
That is my problem too.
Is your "code" going to solve this problem if your "test" is successful ?

[muratcorlu]

@oywino Smanar has a fix on his/her fork and asks for test it if it fixes our problem.

But I also couldn't understand this part:

This code (if it work) try to get the Swbuild, then the date code and finaly the Application version to fill the "swversion" field.

@Smanar What is swbuild?

[oywino]

swbuild is Software Build number, I guess.
I'd love to test whatever it is @Smanar want us to test, but I have no clue how to do this.
Is it possible to prove a step-by-step in human understandable terms ? 😊
sudo apt install deconz-dev
isn't something I can do on my system. I have deconz running in a docker container and it needs a ton of environment variables set correctly in order to run at all.

[Smanar]

Sorry.

To make short, the deconz code ATM, prevent sensor creation for device without "swbuild", it s a field visible in the Json.
By defaut it s the attribute 0x4000, visible on capture on first post, if this attribute is not available, the code take the datacode (0x0006) but for this device the value is empty, so the code stop here.
With my change it will use the application version.

The procedure is explained here https://github.com/dresden-elektronik/deconz-rest-plugin#install-deconz-development-package-optional-linux-only

But you need a linux OS with full OS, I can't explain how to do on docker.

[oywino]

Ok, I see.
But what did @Mimiix mean when he wrote that "It needs to be whitelisted" ?
I was hoping that meant that a quick-fix was imminent.... 😢

[Smanar]

And not sure my fix will work too ^^.
In all situation, we realy need a test, and for that we need you, we haven't the device.

[oywino]

I would need another ConBee dongle and a test system to do this. ANd I don't have that. My existing set up is a productions system that I cannot take down. It would be easier for me to send a device to you.
But it would be nice if @Mimiix would join us again and clarify his opinion.

[Smanar]

Don't worry, you are not alone with this bug.
No one here have the device, the problem, and a linux machine ?

[Mimiix]

@oywino where do you need my opinion on?

[oywino]

@oywino where do you need my opinion on?

Hi @Mimiix , you wrote (if I understood correctly) that all that was needed was to whitelist the device (?)
#4838 (comment)

I was hoping that this could turn out to be an easy fix ?

[Mimiix]

@oywino where do you need my opinion on?

Hi @Mimiix , you wrote (if I understood correctly) that all that was needed was to whitelist the device (?)
#4838 (comment)

I was hoping that this could turn out to be an easy fix ?

Well that's what I thought. But @Smanar says it isn't that easy, so I'm with him on this one. I'm not a developer so I can't change anything about this unfortunately.

[oywino]

Well that's what I thought. But @Smanar says it isn't that easy, so I'm with him on this one. I'm not a developer so I can't change anything about this unfortunately.

Ok, I see.
@Smanar - are you a developer (maintainer) ?
I wish I could be of any help....

[Smanar]

For the moment need someone with the device, with the issue and with a linux machine to test the code modification.

[oywino]

I guess a "linux machine" can be arranged by simply installing Linux on a laptop. Which Linux do you recommend, and what would be the preferred installation config?

[Smanar]

I think all Linux version are supported, I have see some issues with dev package not found for some X64 but I think now it s corrected.
There is that in doc

Ubuntu
Supported Ubuntu versions

    Ubuntu 16.04 LTS 64-Bit PC (AMD64)
    Ubuntu 18.04 LTS 64-Bit PC (AMD64)
    Ubuntu 20.04 LTS 64-Bit PC (AMD64)

Classic install, not "light" "custom", headless (the desktop can be usefull for debugging) or docker.

[oywino]

Ok, I need to find one more ConeBee II because I cannot disconnect the one I have in production.
I hope setting up two ConBee coordinators in my house isn't going to cause any problems (interference) with my existing network (?)

[Smanar]

No, no danger, but just to make test, idk what is managed by your zigbee network, but you can just unplug the conbee from your production machine, make the test, and replug it.
You just need 2 mn to check if the code is working (if you have the sensor or not), and 15 mn to test it.

[oywino]

Ok, do you have step-by-step on how to install and config deConz on Ubuntu 20.04 (inkluding links)?

[Smanar]

Yep, first update the OS then

https://phoscon.de/en/conbee2/install#ubuntu

Then after with command line

sudo apt install deconz-dev
git clone --branch tuya_plug_1 https://github.com/Smanar/deconz-rest-plugin.git
cd deconz-rest-plugin
qmake && make -j2
sudo cp ../libde_rest_plugin.so /usr/share/deCONZ/plugins

[oywino]

Should I do all this first:

[oywino]

BTW you haven't enabled DDF file yet ?
No. Not as far as I know.

[Smanar]

Have you tried to include it using the GUI ?

• delete the node
• set phoscon in permit join "add new light"
• read basic attribute in deconz (cluster 0x0000 button "read") or node descriptor (the menu appear on clic on the first circle in node title) if needed, to trigger an inclusion

If deconz don't display 5 endpoints, no reason to have them in the API ....

[oywino]

Have you tried to include it using the GUI ?

By "GUI" I suppose you mean Phoscon ?
And yes, I deleted the node in deConz
I used "Add new light" in Phoscon
And when the node appeared on deConz, (only a single node), I read the basic cluster.

It still comes up with 5 endpoints in both Phoscon and Home Assistant, but not in deConz.

[Smanar]

Nope the GUI mean on desktop (with the application), phoscon is a webapp.

It still comes up with 5 endpoints in both Phoscon and Home Assistant, but not in deConz.

Ok, it s realy a mystery, can you take some log with "info" and "info_l2" during the inclusion ?

[oywino]

What does "inclusion" mean?
Do mean pairing?

[Smanar]

Yep

[oywino]

I was uncertain where to post, so I did it here:
#5334

[karsie1994]

Hello all,

Today I tried adding the mentioned switch/sensor to Deconz and showup to my Home Assistant frontend. However only the switch/plug is displayed. No information is shared about power consumption. I can however see the power consumption in the cluster information in Deconz.

It also appears that there is a new subversion of this device because the suffix is different:
TZ3000_typdpbpg

Has anyone seen this type before and was able to get it to work?

[DennisvDijk]

Hello all,

Today I tried adding the mentioned switch/sensor to Deconz and showup to my Home Assistant frontend. However only the switch/plug is displayed. No information is shared about power consumption. I can however see the power consumption in the cluster information in Deconz.

It also appears that there is a new subversion of this device because the suffix is different:

TZ3000_typdpbpg

Has anyone seen this type before and was able to get it to work?

I follow.

[basdejong95]

Hello all,

Today I tried adding the mentioned switch/sensor to Deconz and showup to my Home Assistant frontend. However only the switch/plug is displayed. No information is shared about power consumption. I can however see the power consumption in the cluster information in Deconz.

It also appears that there is a new subversion of this device because the suffix is different: TZ3000_typdpbpg

Has anyone seen this type before and was able to get it to work?

Same issue here... It seems that only switch is supported.

[Mimiix]

Should be fixed with a DDF.

[Mimiix]

@Smanar Can you give some pointers on that?

[Smanar]

Sure, for all plugs, same method #5633 (comment)

If the value is updated only one time, when you read it in deconz, it's because of bind (will explain later)

[rzechu]

Same issue

[Smanar]

Hello can you try that #5633 (comment) ?

[bigcookie]

In general I noticed that the manufacturer ID even is different if I order 4 in one lot...:
TZ3000_typdpbpg
the bold part seems to change. I then duplicated the DDF, changed the manufacturer ID and it picked up the correct DDF. In some cases I needed to add the Zigbee devices AGAIN via Sensors. Basically I paired it twice.

What I noticed:
(+) power, voltage and current sensors seem to be working fine
(-) consumption is clearly off...

Has somebody else similar experiences? I have one plug on a system which normally consumes 11W straight - constantly. But the consumption reads +0,01kwh after about 6-8 hours. I would expect it to add ~10W each hour.
Did somebody experience similar odd measurements?

[muratcorlu]

Adding a file to ~/.local/share/dresden-elektronik/deCONZ/devices/tuya with name _TZ3000_cphmq0q7.json and with content given by @Smanar here (by replacing model id and vendor with _TZ3000_cphmq0q7 ) solved my issue.

PS: I didn't check the situation that @bigcookie mentioned above.

[bigcookie]

I found the issue... Actually you have tu multiply the consumption from TS011F plug with 10 to make it accurate.
I hooked up multiple meters in parallel - see attached screenshot.

Can somebody help me, how we can change the DDF to change the multiplication?

[bigcookie]

In DDF Editor (while connected via VNC), I changed the following under state/consumption:
Item.val = Attr.val
to
Item.val = 10*Attr.val

then "save" and "hot reload". Seems to work >happy< :-)
Accuracy is 1/10th of the Wifi sockets I do have, but I think fully acceptable. The TS011F seems to measure from 3W upwards (as already posted by others). So not good for low power measurements, but good for the rest imho.

[Smanar]

And pls share the DDF for others, can make PR yourself or ask for someone make it, but you can at least copy/paste the DDF on github.

[bigcookie]

@Smanar: Will try to make a PR (not a developer, but will manage ;-) )- but can you comment on the different manufacturer IDs?
_TZ3000_cphmq0q7
vs
_TZ3000_typdpbpg
vs
_TZ3000_STRING-I-DONT-KNOW

-> I have at least 2 more manufacturer identifiers in my pool of TS011F devices :-)

Before I will make a PR, please let me know, if we shall create a DDR for each manufacturer ID, or if in this case, the DDF application logic will be adapted to be "flexible" on the string part after "_TZ3000_"?
Depending on that my PR might look different.

Also I would like to get confirmation, if others see the same issue with consumption being reported. I would then make the changes accordingly.

Attached the two DDFs I modified - to be renamed to "*.json":
NAS-WR01B_2.txt
NAS-WR01B_3.txt
for two different manufacturer IDs for TS011F

[Smanar]

So first, don't worry, you can make error. (for me better having a broken code than no code at all, as it can't break others devices)

On my side I m trying to mix DDF, when they are absolutely clone, mean they use the same DDF (with same value convertion for power for exemple)
For exemple for this one https://github.com/dresden-elektronik/deconz-rest-plugin/blob/master/devices/tuya/_TZ3000_yd2e749y_temp_hum_sensor.json

You can that with

  "manufacturername": ["_TZ3000_yd2e749y", "_TZ3000_fllyghyj"],
  "modelid": ["TS0201", "TS0201"],

rules are

  "manufacturername": ["manufacturename1", "manufacturename2"],
  "modelid": ["modelid1", "modelid2"],

Else can make a new file.

If I m right they are all NAS WR10B, and use same code ? so can mix them in the same DDF.

Wana make the PR yourself ? need to put the file in the devices/tuya folder, Try to use same kind of file name than others.
Have checked the code, no DDF files for thoses devices yet.

[bigcookie]

Thanks. Will try to do the PR - give me some time :-). If I recall, I can even do that with the github web interface :-)... Otherwise will do that from my computer. I like the option to use the DDF for multiple models/manufacturers... Will handle it accordingly.

[bigcookie]

Easier than thought - I could create the file in the web GUI. PR is here:
#6131

[Smanar]

Nice have just it ^^, thx.

[rzechu]

Hello can you try that #5633 (comment) ?

I saw that but have no idea how to make it and stay persistent in HomeAssisant + Deconz as docker image.
I can access HA files but No idea how to access deconz docker image data.
Any ETA for this feature available inside common version?

Will try with
https://forum.phoscon.de/t/how-to-add-edit-a-ddf-on-home-assistant-using-text-editor/1839/2

[hubertgruber]

I currently have this issue for TZ3000_5f43h46b

[Smanar]

This device don't have DDF, so you need to create one or add your model to an existing one, generaly tuya are clones, so all devices are using the same DDF.