Tuya Smart Plugs states drops to zero frequently

[Re4zOon]

Hi,

Dont know if anyone come accross a similar issue, but...
A picture tells better stories:

All those vertical lines are datapoints for "0" watt.
My question is: can I somehow average these datapoints, so I dont go mad?
It can be a new sensor for all I care, I just want a nice graph :)

More info:
I have several of these exact plugs, but only some of these do them.
There are no updates to them, everything is running the same FW, etc.
Model: TS011F_plug_1

I see the mention on the page about the polling, but its not recognized as _3 so I dont have that option.

[kzajac83]

The same issue for in my" TS011F_plug_1" and for my colleges using the same device in HA. In meantime update FW via OTA but nothing change.
Please note using ZHA have the same issue so problem probalby can by in socket or Zigbee concentrator but also try on two different device butissue still the same (however chip in both the same CC2652P). On the other hand other data in the same time like voltage is valid then communcation problaby is not a problem.

Because few hoursago change ZHA -> Z2M then my history is short but see the issue:

[Dash]

I too am seeing these drops to zero

I've only got one of these sockets at the moment, and is generally unbranded otherwise. I assume this is a firmware issue (OTA v.192), but no idea who to feed that back to.

Power and Current both show the same problem and the drop to 0 occurs at the same time.

[kzajac83]

Lastly buy some zigbee socket and after connect to Z2M is visible as TS011F but both sockets have others manufacturer codes:

So have two sockets registerd as TS011F and one report zero ramdomly (see two post above) and new one working fine:

Both OTA / FW version is 192.

Badly socket manufacter is _TZ3000_hdopuwv6 and working correctly _TZ3000_gjnozsaz

[Dash]

Interesting, mine is "_TZ3000_cehuw1lw", but no way of identifying that before buying.

[Re4zOon]

For me it doesn't really matter why it happens, its obvious its not an issue with Z2M.
However, as its a "common" problem, we need a workaround of sorts.
As I said, even a custom entity, that averages the last 10 datapoints or something is fine.

Since I made my original post I learned some about HA and how to use templates and such, so possibly I will solve this for myself and paste the config here.

[Dash]

This has been fixed with issue #16076. If you're getting the same thing on a different device, then it seems adding a comment to that issue is the way.

The current list of devices with exceptions is: '_TZ3000_gvn91tmx', '_TZ3000_amdymr7l', '_TZ3000_typdpbpg', '_TZ3000_hdopuwv6', '_TZ3000_b28wrpvx', '_TZ3000_1h2x4akh', '_TZ3000_cehuw1lw'

For these devices, if the value goes to 0, it'll hold off for 5 seconds before sending a zero value through.
https://github.com/Koenkk/zigbee-herdsman-converters/blob/75d28701a7b321ab95c4137a4ce753ff329a2178/src/devices/tuya.ts#L445C4-L445C4