IKEA just announced 21 new affordable Matter-over-Thread smart home devices (which may also support Zigbee too)

Bought the ALPSTUGA today (located in Australia) and looking forward to the new light bulbs and hopefully power plugs launching in the next few months.

Can someone please explain the pros and cons of using the new Dirigera Hub (ZigBee + Thread) versus the HA Connect ZBT-2? From what I’ve found online, it looks like the Hub can expose all IKEA ZigBee and any Matter over Thread device to HA just fine. Hub is about $10-$15 cheaper than the ZBT-2 and at least will connect to some older IKEA devices I have.

Any good links to blogs or videos would be helpful. Thanks!

1 Like

Thanks @danieldk for this great overview.

I bought (2) ALPSTUGA sensors some weeks ago and it has been a bumpy ride to say the least.

After lots of trial and error I eventually got them to register with my setup: Homekit - HASS with an ATV 4K as Border Router.

I am mainly interested in reading NOx. As others have reported I found the sensors to be very responsive; I can live with the signal noisiness, so I am quite happy here.

What I am struggling with is to read ‘reasonable’ levels in general. I have been comparing the ALPSTUGA sensors with other sensors I have: Netatmo, Velux. Sometimes I get close tracking (within +/- 100 ppm) between all of them, but often the ALPSTUGA sensors read much higher, e.g. 1200 vs 700 ppm; and often with 200 ppm between both ALPSTUGA.
This makes the ALPSTUGA sensors unusable for me and I am close to returning them.

In my research of a solution, I found that the ALPSTUGA sensor should be exposed to fresh/outside air for several hours get recalibrated to those around 400 ppm. I typically have been doing this either outside or in my garage. This seems to bring all sensors to be somewhat in sync at that level.
However, when move the sensors to their desired location (office, bedroom) the ALPSTUGA often read much higher as the other brands again. While they seem to settle down after a few days, they typically still continue to read much higher. So in short, I am hoping for some pointers on how I can get more realistic reading levels out of ALPSTUGA sensors.TIA!

1 Like

Yes, Espressif makes one for $10. The $10 device is a self-contained Thread Border Router but “some assembly required”. You need to configure and compile some code and flash it to the ESP-32.

https://openthread.io/guides/border-router/espressif-esp32

The other thing people do is get a Zigbee/Thread radio USB dongle to put on their HA server, then run the border router in a container on the HA server. I think this is overkill as the BR can run on an ESP32 chip. And thge $10 price is better.

They also sell an Ethernet daughter card that about double the price but is worth it

At least for CO2, I don’t think this is possible. The STCC4 sensor inside the SEN63C inside the ALPSTUGA measures thermal conductivity between a heater and a temperature sensor. The theory is that if the mix of gasses modulo CO2 stays exactly the same between the heater and the temperature sensor, then you know the concentration in CO2 based on changes in conductivity.

Of course, in reality there are other factors that influence this, slight changes in ambient gasses, humidity in the air, etc. all influence conductivity. The STCC4 compensates by this partially by taking ambient temperature and humidity into account (by feeding data from the SHTxx sensor to the STC44). But I think there are simply too many external factors to influence thermal conductivity.

The CO2 measurements of the ALPSTUGA should not be taken for anything else than trends.

For my applications the ALPSTUGA is also worthless. I work in a relatively small home office and I want to freshen the air more actively when the CO2 concentration reaches 1000 ppm at breathing height. The problem is that 420-1000 ppm is not a huge range, especially when the ALPSTUGA is sometimes off by 200 ppm even at lower concentrations (e.g. at 1000 ppm when other (better) NDIR sensors are still ~800 ppm). See an example here:

https://danieldk.eu/IKEA-ALPSTUGA#sensor-quality

I don’t know, I found having a ZBT-2 directly connected to my HA instance much easier/nicer to debug when things didn’t work well. Besides that having the Matter sever and border router on the same machine makes it less sensitive (I think) to mDNS or IPv6 issues due to router misconfiguration (IGMP snooping, etc.).

I also like supporting HA/Nabu Casa buy buying two ZBT-2 :+1: .

3 Likes