IKEA Vindriktning Air Quality Sensor

I’ve also just reached a similar conclusion as searching for PM1006 on Google. It brings up PM1006K articles where it is just referenced as PM1006. So, it might be some naming thing and they should be the same. Let’s wait for their answer.

By the way, after browsing their website, if you ever come across one of their AM7000 device for a reasonable price, then please let me know. It looks really tempting…

I found one for EUR 149 - I guess that’s not reasonable? :money_mouth_face:

I thought of something closer to 49 Euro. :wink:

By the way, what is the TX message to the sensor from the IKEA device? Maybe that is the problem. The document says what it should be to query the measurements. Maybe that has some difference.

Read measures result of particles: 
 Send: 11 01 02 EC

Haha, I had not read your post yet until I came here to comment, but: I just sent it 11 01 02 EC and then I get a frame back that matches the PM1006K datasheet. PM2.5 value matches what I get from the PM1006-style command. PM1.0 and PM10 are zero.

The IKEA MCU sends 0x11, 0x02, 0x0B, 0x01, 0xE1, like the PM1006 datasheet says.

I did a guide on making it work with Tasmota and went a bit overboard with air quality sensoring - IKEA Vindriktning air quality sensor running Tasmota | Blakadder's Smarthome Shenanigans


Running into same issue, does someone have a solution? I used nodemcuv2

1 Like

I’m convinced that my sensor is damaged! I’ll get a new one to test it out

@Soloam, @shreyask - You can try to build it with RX/TX once @Habbie finished that PR to ESPHome. See above the post about it.

Otherwise, I cannot remember have I asked it. Are the LED color changes when you introduce some dust to the sensor?

If nothing works today for someone, the RX/TX PR won’t change that. RX/TX is for people who want to disconnect the sensor from the IKEA board.

I am assuming that the IKEA board is not sending TX so there is nothing to receive through RX. So might be not the sensor is faulty, but the IKEA board. That’s why I asked that has the LED color changed by introducing dust to the environment.

Right! However, using the RX/TX code might still involve disconnecting the sensor from the PCB then, because in a working setup, at least, TX is held idle by the IKEA PCB and we can’t TX over the test pads.

Cubic came back to me: ‘After confirmation with engineer, the DF7/8/11/12/13 are the Reserved byte.’ I guess these bytes are not for us.

1 Like

Note that there’s a light intensity sensor above the top LED on the front already. Maybe you do not need to add the TSL2561.

1 Like

Confirmed my sensor is faulty :frowning_face: , only shows green and nothing else. I tested using incense stick and placed the sensor near it, I have two sensors and one of them turned red and the other stayed green.

That’s a limitation of HA. HA won’t “update” a value, if it got the same value a second time, and only react to a change in values. So it will change to unknown if there’s no signal and keep saying “4 hours ago” if there’s the same value send for the last 4 hours.

Well, green might be the default color. So if nothing from the sensor is received, it will always be green.

So this could still mean that the TX-Part is not working correctly on the board.

Have you checked simple things like cold solder joins?

I’ve added some smoothing to my setup, since the values received are jumping up/down quite a bit - like 7, 7, 7, 8, 7, 7, 7, 8 and are extremly frequent.

So it makes sense to calculate a sliding average to catch not only the “last” value for the burst, but instead take all measurements into account:

  - platform: pm1006
      name: "PM2.5"
      accuracy_decimals: 2
        - sliding_window_moving_average:
            window_size: 50
            send_every: 10

Energy consumption

In case someone wants to add energy accounting for this device: A unmodified Vindriktning with an added Wemos D1 mini with ESPHome running on it is using the following amount of power:

~0.412 W while the fan is not running
~0.558 W while the fan is running

Measured with an Atorch UD24.

I ran a consumption measurement over 4 hours, and that’s the average consumption (at 5V):
0.5178 W


Indeed but using a more accurate sensor not placed next to LED lights might be beneficial

I disconnected the sensor last week. I can confirm that the VINDRIKTNING stays green if no data is received.