[Device Support Request] IKEA inspelning power monitoring smart plug #3374
Comments
|
I've just bought this smart plug, and the values appear to be correct on my Home-Assistant instance.
|
|
I just bought this plug as well, and I have the same problem. @cristiangauma can you share your firmware version? |
|
I had the same problem, updated the firmware with ZHA OTA to 0x02040045 and this fixed the issue. It displays the correct voltage now |
|
@saversux is there a way to force an update? |
|
If you have Tradfri OTA activated it should display the update in the updates section. @parrel You need to set this in your configurations.yaml |
|
I have that but for some reason it says 0x02040034 is the latest version. |
I read on other threads about Z2MQTT that the new firmware fixes the voltage indeed, but still has the wrong Power. Can you confirm whether it fixes both, by testing a power reading you know for sure ? |
|
@gllmlbrt just tested it with a USB-C charger that has a display for the power output. Additional learning: it seems to push updates very frequent, much faster than Tuya TS011F.
|
|
Just got it upgrded to 0x02040045 and it all displays as expected. |
|
The OTA is only in Dirigera feed. Meta data in the signed file: |
|
Sorry, I forgot to confirm that yes, once I added the smart plug to homeassistant, a firmware upgrade notification appeared asking to upgrade to 0x02040045. The screenshot that I shared before was after upgrading. Before upgrading I didn't check the entities, as the first thing that I did is to upgrade to the latest firmware. |
|
THe Z2M community seems to be on to the topic of the bad power reading already, I also noticed that it's off only for loads >=1000W. Seems to be an auto-ranging feature of the plug, see my comment on Z2M thread. So it seems we need to check/update the ac_power_divisor on every update of the power-value, and reflect it in the shown value accordingly. @gllmlbrt maybe you wanna reopen this topic? Also I opened PR #3379 to remove the level controls from the configuration panel. |
|
Re opened for visibility until the power values are 100% confirmed to be OK. |
|
I have the same problem, I have updated the firwmare and from 22.7V I have gone to 227V, that part corrected (Voltage), but the part of W (Power) when it passes 999V makes an erroneous division. I know it is reported, but the interpretation error comes from the way the device works? or from the way zigpy manages the data? I mean, it doesn't follow a standard (or at least the same as other sockets). It's more for knowledge :) Thanks!! |
|
Same issues. Is there a way to update the firmware without Tradfri? I paired directly in ZHA. |
|
Configure OTA updates for ZHA then you can update all IKEA devices from ZHA. |
|
It seems like there is some kind of divisor that needs to be taken into account for measurements above 1000w Otherwise, the newest firmware does fix the incorrect voltage reading, but not the readings above 1000w. Seems like it would be as simple as multiplying the power rating by the multiplier that is supplied by the plug. |
|
Need adding binding and reporting to the quirk and it shall working OK. |
|
Am also having the 1000W issue. Temporarily worked around with a template. I assume we just need a quirk writing? |
|
Is there any progress in solving the power measurement issue in ZHA for |
|
Some news about this issue ? (1000W) |
|
Here is what I found to be a fix for the >1000W issue. Not sure if this is a permanent fix as I have only been using it for a couple of days. A firmware update will most likely revert this and you will need to perform it again. I have not tested this though. You will need:
This can be done from within developer tools as well but I stored it in an automation for easy access in the future. Just create a new automation and go into YAML-editing mode and paste the above. Then go back into the normal editing mode. You will need to change the IEEE to your specific plug, you can find yours under "Zigbee info" on the plug device page.
|
|
It shall working only if the power is going up or down it can being the "old" power factor is being used for the first report but i think we can living with it. Ping: @puddly and @TheJulianJES do you both have some thinking implanting reporting if this cluster / attribute for IKEA devices so we getting it working OK ? |
I was specifically looking for that during my testing with 5 different plugs, but I never had that issue. You are right that it could happen. |
|
Its only theoretic and have not seen it then sniffing IKEA Dirigera but if its happening its only for one reading if its happening. Bra jobbat !!! |
|
Can this fix be added to a quirk ? |
|
Not only is my voltage off, but amps*volts do not make watts: 220V system, but neither 3.3A times 229V nor 3.3A times 22.9V result in 527W. |
|
Same >999W issue here. F/W xxx45, newest per Jan 2025. |
That's because for AC, true power P = U * I * cos(phi), where cos(phi) only ever can be considered almost 1 for resistive loads like a space heater for example. Apparent power S = 229V*3.3A = 756 VA, which would result in a cos(phi) of 0.697, which depending on your load could be very plausible. If you wanna know more, check out wikipedia. Since we don't know cos(phi) from the plug (some other plugs like the Aqara Smart Plug report this attribute too), we cannot reliably cross-check the values by the plug between each other. We can be pretty sure the voltage should be way higher and the scalar is not up-to-date (even in the US with their lower voltage), and it's possible to say if the values are likely off, if cos(phi) < 0.1 at higher loads like 100W+, and also cos(phi) cannot be greater 1 ofc, but for the most part it's still assumptions, i.e. a very hacky solution that will still cause issues for ppl. So to fix this properly, we have to figure out how to set reporting intervals like @BMAN1995 already showed how to do via ZHA toolkit as a workaround, and implement the same solution in a proper quirk somehow. Unfortunately all this seems way easier in Z2M still, and while the issue also doesn't seem to be completly fixed there, I'm rly considering switching over the way better device support and expandability overall :/ |
|
@puddly and @TheJulianJES This is still one issue that no devs have looking on and its one of the most speed devices and its need attention for the devs then normal users is not having the knowledge implanting the reporting and how it shall being done in the best way (very likely IKEA specific for not messing with other devices). Thanks in advance. |
TheJulianJES
added
the
priority: medium
|
I think setting up the attributes for reporting should work: zigpy/zha#348. They're listed as reportable in the spec and since this is one of the few devices that actually changes them, this should be possible. I don't have a device to test with at the moment but I'm probably pick one up later this week. For reference, this is the config we'd be using: alias: IKEA INSPELNING Report
description: ""
triggers: []
conditions: []
actions:
- action: zha_toolkit.conf_report
metadata: {}
data:
ieee: 6c:fd:22:ff:fe:16:c9:cb
endpoint: 1
cluster: 2820
attribute: 1541
min_interval: 0 # This I think should be 0, not 2, because we want the device to always report changes
reportable_change: 1
max_interval: 900
mode: singleI'm not sure if the order in which attribute reporting is set up matters, as we ideally want the divisor and multiplier attributes to come in before the actual readings. |
|
By the way P of reporting order in your PR. |
|
@MattWestb @puddly I agree, the order of configuring reporting shouldn't rly matter, it's just values/attributes you're setting on the device (i.e. it doesn't keep track of any order changing them). @puddly I can try patching my HA / ZHA setup and having a look / provide logs and feedback. Actually I even got a second ZBDongle and IKEA inspelning plug, that I can throw into my HA testing container and provide access for you guys, if it'd help ya. I'd say min_interval of divisor/multiplier should be in-line / ident with what is set for the actual attributes? I could look into it / try to fetch it with ZHA toolkit during the week. |
|
The ZHA toolkit fix from @BMAN1995 seems to do the trick on my side with min_interval: 2 I tried with 2 Inspelning (dishwasher and dryer) next step will be on a washing machine and cooking oven |
Problem description
New IKEA inspelning power monitoring smart plug pairs fine and has all the entities in Hass. However power and voltage are reported divided by factor 10.
Solution description
An adjustment of the voltage and power (as a result energy calculation) via a quirk?
Screenshots/Video
Screenshots/Video
Device Signature
Device signature
Diagnostic information
Diagnostic information
The text was updated successfully, but these errors were encountered: