IKEA Alpstuga Review: IKEA’s Most Capable Monitor Yet?

Welcome to the long-awaited IKEA Alpstuga review. And I do mean long-awaited, so before anything else, an apology for how long this article’s taken. Just finding an Alpstuga to test turned into a bit of a saga and I ended up wandering through IKEAs in four different countries hoping one would be sitting on a shelf, never finding a single unit. Even a shipping-forwarder that I tried using couldn’t receive the monitors because the nearby IKEAs in Florida were all out of stock. In the end, months went by with nothing to show for it before an IKEA in Los Angeles turned out to have them in stock, and I was able to walk out with three.

At this point, I’ve had them around my home for around two months, and honestly, I’d have wrapped this review up much sooner if the CO2 readings hadn’t kept dragging me back for more testing and side-by-side comparisons. Something strange kept showing up in the carbon dioxide data (more on that later, because it’s easily the most interesting part of this whole review), and each time I thought I’d made sense of it, more data threw up new questions. So here we are, several months on. Again, I’m sorry for the wait. I just wanted to understand everything I could about how this thing behaves before I published an in-depth review, even if that meant missing the window when the device was a very hot topic in the smart home community.

With that said, I’m surprised the Alpstuga exists at all. I was caught off guard when IKEA launched its first air quality monitors, the VINDRIKTNING and the VINDSTYRKA. A flat-pack furniture company moving into particulate sensing (even if from the smart home angle) wasn’t something I saw coming. Now I’m sitting here reviewing what you could call their second-generation effort, and I’m more surprised than I was the first time. I didn’t expect IKEA to stick around in this space, but it’s great that they have.

Anyway, I digress. Before this introduction turns into an article of its own, two things make the Alpstuga worth your attention. One is the price, which undercuts nearly everything else in its category. The other is the move to Sensirion’s newer sensor series, and that’s where most of my excitement lives (along with a few of my lingering questions). Let’s get started!

Contents

Accuracy & Sensor

The Alpstuga tracks four pollutants: CO2, PM2.5, temperature and relative humidity. To do all of this with one small box (and, I’m sure, to keep the cost low), IKEA has gone with one of Sensirion’s new SEN6x-series all-in-one sensors, specifically the SEN63C. This module can actually report PM1, PM4 and PM10 as well, but IKEA has chosen not to surface those channels, so you’ll only be able to see PM2.5 values. Considering how IKEA was been hesitant of using VOC readings (and rightly so) with the Vindstyrka, this doesn’t surprise me.

This new Sensirion series caught my attention because both the particle sensor and the CO2 sensor are brand new designs rather than quite familiar components (as was the case with the SEN4x and SEN5x series). Let’s start with the CO2 side, since it’s the bigger departure. Instead of the NDIR or photoacoustic sensing that most decent CO2 monitors rely on, the SEN63C uses a thermal-conductivity sensor, which works on a completely different principle.

Thermal-conductivity sensors read CO2 by measuring how the air around it carries heat, since carbon dioxide moves heat differently from the nitrogen and oxygen that make up most of what we breathe. As CO2 builds up, that heat transfer shifts, and the sensor works backwards from the change to estimate the concentration, which also means it leans heavily on an accurate temperature reading to separate the CO2 signal from everything else going on. This correction is handled on the module itself, since a temperature and humidity sensor also sits inside. I’ve spent time with thermal-conductivity CO2 sensors before (the Atmotube Pro 2 uses the same approach) and found they come with some considerable drawbacks, so I was curious to see how one behaves in a stationary, mains-powered monitor that isn’t being constantly exposed to varying conditions. But more on that soon.

The PM sensor is new too. It still works the way most low-cost particle sensors do, pulling air past a laser and reading the light that scatters off each particle, so a lot of the usual behaviour carries over. What’s changed is the hardware. Sensirion has shrunk the entire optical stage down into a single chip-scale MEMS component and wrapped the sampled air in a sleeve of cleaner airflow to keep dust off the optics (which is why the estimate a 10-year lifespan for this sensor). With these two new sensors onboard, I was quite curious to see how the Alpstuga would perform. Since PM is a lot more straightforward in this case, let’s start there.

For all the PM2.5 testing I ran my three Alpstugas next to two AirGradient monitors. Now, I want to be clear about this: the AirGradients are factory-calibrated, but they are not reference instruments. Without an actual reference in the room, there’s no way to know the true PM2.5 concentration at any moment, so please don’t treat the AirGradient numbers as truth. Instead, what these comparisons can show is the shape of the trend and how closely the group of identical sensors agree with each other, which is really a question of precision and reproducibility rather than accuracy. The two AirGradient sensors are included as a point of reference, and they become a lot more important for the CO2 comparison.

The yellow, orange and red lines represent the Alpstugas and blue lines are AirGradient ONE monitors.

It’s worth noting that I ran these devices side-by-side for around two months, and I consistently saw the trends I will discuss here. However, I’ve shown just smaller snapshots as the whole data set is too much for a single image. Regardless, across a full day (shown above), every sensor follows the same trends, and the broad agreement between all five is good. When PM2.5 climbs, they all climb together, and when it settles, they settle together. Both big pollution events (cooking smoke) on the day I’ve charted show up clearly, at the same moments, with the same shape.

What stands out, though, is how much three identical sensors can disagree on the actual figure. At the morning peak, the three Alpstugas read somewhere between about 280 and 440 µg/m³, a spread of over 150 µg/m³ between units sitting side by side. Sensirion’s own spec for these modules is a precision figure of around ±5 µg/m³ below 100 µg/m³ and ±10% of the reading above that, so a 40% spread between units is significantly outside this range. While it’s hard to know exactly why this is, it could be due to the PM calibration process as Sensirion calibrates the sensor against a KCl solution. Real cooking smoke and household particulate (the spikes in my home) look nothing like that reference aerosol to an optical sensor, so once you’re measuring genuine indoor pollution the calibration is likely far less useful.

It’s easier to see the difference between sensors when we zoom in on the spike.

I believe this to be the case because the disagreement is systematic: the same unit reads lowest consistently and the same one reads highest, across both the busy period and the quiet overnight window (below). It’s also worth mentioning that all three Alpstugas sat below both AirGradients at the peak, where the AirGradient pair ran up to 575 and 660 µg/m³. However, I wouldn’t read that as the Alpstuga under-reporting, because the two companies calibrate their particle output against different references, so some of that gap is likely just differences in calibration techniques.

Above is a short quiet-period graph that I wanted to include to show the noise of these sensors from reading to reading (in this case, at low concentrations). As you can see, both types of device fluctuate quite a lot between readings, but this is expected and makes a case for why you shouldn’t rely on a single PM2.5 reading from any low cost monitor.

So where does that leave the Alpstuga on particles? Well, it reads the trend well, just to varying degrees from unit to unit. For watching how your air changes through the day, spotting cooking or smoke events, and deciding whether to open a window or run a purifier, it’s up to the task. What I wouldn’t do is fixate on any single reading or treat a peak number as the truth. Instead, look at the trend and lean on averages rather than instantaneous values. That’s true of most low-cost particle sensors, and the Alpstuga is no exception.

Moving on to temperature, and this is one area where I was quite surprised. For this and the humidity test, I ran the three Alpstugas against a lab-calibrated hygrometer over a full day. Please note the same caveat as before: a hygrometer isn’t a reference instrument, so “close to the hygrometer” tells me the Alpstuga is in the right ballpark and that the units agree, not necessarily what the true temperature was.

I went in expecting the device to consistently read hot. After all, it’s a small sealed box with a range of sensors and electronics all generating heat inside, and that usually leaks into the temperature reading. Instead the three units tracked slightly below the hygrometer most of the day, the gap sitting around half a degree overnight and opening to maybe a degree and a half at the warm afternoon peak. The largest disagreement I saw was only a couple of degrees, and better than I expected from something this size. I guess this is due the self-heating compensation Sensirion builds into these modules.

Of course, my testing only ran across a narrow band, roughly 25 to 31°C, so I can’t tell you how it behaves in a cold or (very) hot room, and accuracy will likely drift outside that. Within a normal indoor range, though, the temperature reading is quite usable.

Humidity was similar. The three Alpstugas consistently read a few percent below the hygrometer, usually two to five percentage points low, occasionally a touch more in the dryer afternoon stretches (when we run AC). At the humid peaks the gap narrowed and they nearly converged. So there’s a mild, fairly steady tendency to under-read.

Importantly, the trends all match. Every rise and fall showed up on the Alpstuga traces at the same time and with the same swing, from the high-60s overnight to the mid-30s in the dry morning and back. The three units stayed tightly grouped again, as they did on temperature. A steady offset of a few percent is quite normal in such monitors so I don’t see this as an issue. Furthermore, if you have a device to compare to and can identify the offset, it’s relatively easy to fix in Home Assistant.

That leads us to the parameter you are probably here for – CO2. To be honest, this was the reason this review took as long as it did. Everything else was relatively straightforward, but while I was logging CO2 against my other monitors, I kept finding things that pulled me back to the data for another week. Unlike for PM2.5, the AirGradient devices can be considered as a proper baseline here because they run NDIR sensors (the SenseAir S8/S88), which measures CO2 directly by how it absorbs infrared light. Once calibrated (and offset in this case), NDIR sensors provide close enough to a working baseline for me to trust. So for these tests I had something solid to compare against, and it quickly became apparent that the thermal-conductivity approach comes loaded with caveats.

Notice how closely NDIR sensors agree and how significant the Alpstuga noise is, even at low concentrations.

Before we continue, however, one important note: the CO2 reading on the display is not something to take at face value. The sensor updates roughly once a second, and the readings jump around a lot. This means there is a lot of noise in the readings. The SEN63C is specified at about ±100 ppm at its reference conditions, looser than the NDIR parts most decent monitors use (Sensirion’s own SCD30 is ±30 ppm and the SCD40 ±50). In that first week I nearly wrote the whole device off, because I’d glance over and two units would read 600 to 700 ppm while the third sat at 1100. A few hours later one would be showing 1500 while the other two read 1000. If you are purely looking at the number shown on the screen, you’ll just be seeing noise. You can see a couple of examples of this on the forum.

The graph above shows the average difference of the three Alpstuga CO2 readings vs two NDIR sensors acting as a baseline. As you can see, the sensors went from readings 200 ppm too high, to slowly becoming more accurate.

The first thing to know is that this sensor takes a long time to settle. IKEA and Sensirion both say to allow twelve hours for the CO2 reading to stabilise. In my testing, across several units and several attempts, the real figure was closer to a week. For those first several days the Alpstugas showed very different readings to a calibrated NDIR sensor. To check if this behaviour is normal, I ran another test and you can see the 12-hour differences below. This confirmed the behaviour. I even ran a third test because this seemed odd (that’s why this review took so long!) and I can confirm that, in every case, my three Alpstuga sensors over reported significantly for the first week and slowly became more accurate before underreporting. Even when the devices were initially powered on next to an open window I noted this behaviour.

So what is the explanation behind this? I assume it’s to do with ASC, as the sensor runs an automatic self-calibration that assumes it will see fresh outdoor air, about 400 ppm, at least once a week, and it re-anchors its baseline off that. It seems that the device doesn’t calibrate when it’s first powered on, and takes one week to perform a calibration. As you can see on the graph above, automatic calibrations occurred on June 23rd, June 30th, and July 7th. You can’t manually request a calibration, so you’re forced to wait for 7 days to get an initial calibration.

What really confused me, however, was how the sensor would overreport, then slowly converge with the AirGradients, then drift apart again. While the automatic calibrations (which can’t be disabled) bring the devices in line for a while, they quickly begin to diverge again. After weeks of chasing this, the answer was something I probably should have seen coming from the words “thermal conductivity.” The sensor is non-selective, so it can’t tell CO2 apart from anything else that changes how the air carries heat, and Sensirion is clear that thermal conductivity is driven mostly by temperature, with humidity and pressure playing smaller parts. The temperature and humidity sensor built into the SEN63C exists to correct the CO2 sensor, but the correction looks to only shrinks the effect rather than erasing it. Below is a graph with one day of data from one of the Alpstugas showing CO2 difference (compared to the NDIR sensors), temperature and relative humidity.

You can see that the difference in CO2 concentration vs the NDIR baseline is greatly impacted by temperature and relative humidity changes. At around 25°C-26 °C, the Alpstuga provides almost identical readings to the NDIR sensors. As the temperature increases, so does the difference in the Alpstuga readings.

But why the agreement at 25°C? Well, the sensor’s data sheet defines the CO2 accuracy at a reference condition of 25°C, 50% RH and 1013 hPa, and the STCC4/SEN63C’s typical operating point is listed at 25°C. So the sensor appears to perform where it’s specified to perform, and drifts as the temperature moves away from that. Again, I analysed this across all three of my units over different time periods and this was a constant trend.

So where does that leave us with the CO2 accuracy of the Alpstuga? Well, further testing needs to be done at colder temperatures, but I didn’t want to delay this article further. Based on my findings, the sensor seems to perform quite well if you can maintain a temperature around 25°C. Anything higher or lower than this will cause the readings to deviate from the truth, and sometimes quite significantly. It might be possible to improve performance with a custom correction in Home Assistant, but that isn’t something that I felt was in the scope of this article. Regardless, if all you want to know is when you should open a window, this device is capable enough because it can monitor the general trends. For any case where you want better accuracy, you’ll need to look for an NDIR sensor.

Design

The Alpstuga is tiny. It measures 40 x 40 x 90 mm and weighs very little, with IKEA listing the packaged weight at around 90 grams, so the bare device is lighter still. When I pulled the first one out of the box, my immediate thought was that it looks like a little alarm clock, and that impression stayed with me (I even had visitors think it was a clock!). It sits comfortably in the palm of your hand and takes up less space on a shelf than a phone lying flat.

The front is one LED matrix display that brightens and dims on its own depending on the light in the room. It shows a single reading at a time, with a small label off to the left telling you what you’re looking at and a warning light sitting to the right. You move through the readings by pressing the large button on top, and the order goes CO₂, then PM2.5, temperature, humidity, time, and back around to CO₂. So in a roundabout way, yes, it’s also a clock. The text reads clearly from across a room and stays legible at night.

The LED is not colour-coded and rather has a binary on or off state. It’s a single orange warning light, and all it does is switch on when CO₂ climbs past 1000 ppm or PM2.5 goes above 15 µg/m³. The catch is that the CO₂ warning only shows while CO₂ or the clock is on the display, and the PM2.5 warning only shows while PM2.5 or the clock is up. So if you’ve left it sitting on the temperature screen, your CO₂ could be getting into headache territory and the light would tell you nothing.

The back of the device hold a USB-C port and the venting for the sensor’s intake and exhaust. That venting means you shouldn’t push the device flat against a wall. IKEA asks for at least 10 cm of clearance behind it, which is about the least you can get away with anyway, unless you buy a right-angle USB-C cable to stop the plug jutting out the back.

On the underside sit two small feet, regulatory information, the matter QR code, and four of the device’s five buttons.

• System button. Handles pairing and resetting. I got to know this one far too well during setup.
• Settings button. A short press swaps between 12- and 24-hour time, a long press flips Celsius and Fahrenheit.
• Set time button. Adjusts the hours and minutes, then confirms them.
• Plus button. Changes the hours and minutes.

Finally, there is one point worth remembering: The box doesn’t come with a cable or a power adapter, so you’ll be supplying your own USB-C cable and plug, or paying IKEA a few dollars for theirs.

Connectivity

Before anything else, the Alpstuga is a Matter over Thread device, meaning It doesn’t use Wi-Fi at all. Thread is a separate low-power mesh protocol, so unless you already have a Thread border router in your home, you can’t access the smart side of this device whatsoever. The good news is that border routers are more common than you’d think, and you may already own one without realising it as Apple Home, Alexa, Google Home and Homey all have certain devices in their ranges that can act as a Thread border router.

IKEA also sells its own hub, the DIRIGERA, which since a firmware update now doubles as a Thread border router and Matter controller, so if you’d rather stay inside IKEA’s ecosystem you can run the Alpstuga straight through the IKEA Home Smart app. At around $109 for the DIRIGERA it’s a pricey way in, and worth stressing it’s optional, since any of the border routers above will do the same job. In my case I used an Apple TV as the border router and paired the Alpstugas into Home Assistant through it. One nice side effect of the Alpstuga being mains-powered (it runs off USB-C rather than a battery) is that it becomes a Thread router itself, which means it helps carry traffic for other Thread devices and strengthens the mesh for anything nearby.

The real beauty of Matter is that you can connect a single device to several ecosystems at once, so the same Alpstuga could live in Apple Home and Google Home together if you wanted. I skipped all of that. Since I was mostly running comparisons and logging data, I stuck to Home Assistant on its own, because it’s far and away the best of the smart home platforms for handling and analysing data.

Now, the pairing itself. This almost made me give up on the devices, to be honest. I had to reset each device several times before it would connect, and I spent a solid 30 minutes sat right next to the Apple TV wrestling with them before I got there. At one point I had a browser tab open to buy a Thread antenna, convinced something was wrong with my setup and I needed better reception. Then I paired a few other Matter devices without any trouble, which told me my setup was fine and the Alpstugas were just being fussy. If you hit the same wall, resetting is straightforward once you know how (You hold the system button on the device for around 10 seconds until the LED stops blinking, and it restarts and reopens its 15-minute pairing window).

Oddly enough, after all the grief getting them online, they’ve been flawless. I’ve run them for two months now and not once have they dropped off or given me any trouble. They even performed well after we went through a power cut. When the electricity came back, half my Wi-Fi network was a mess and I was manually reconnecting a pile of smart home devices one by one. The Alpstugas and my other Thread devices? They all recovered and rejoined on their own shortly after the power returned.

There is one catch that same power cut exposed, though. The Thread connection came back by itself, but the settings on the device did not. The clock and the temperature unit both reset, so anything you’ve set on the unit itself you’ll be setting again after an outage. I also suspect, though I can’t prove it, that losing power kicks off the CO2 sensor’s long settling process from scratch.

So I’ve ended up quite happy with the connectivity overall. The devices were a hassle to set up, but they’ve been steadier since than my Wi-Fi monitors, which is not what I expected when I was cursing at them during the setup process.

In terms of what actually comes through, the Alpstuga sends an air quality status (a simple label like “good” or “poor”), CO2, PM2.5, humidity and temperature into Home Assistant. It doesn’t expose any settings, and there’s no calibration control for the CO2 sensor. What you get in Home Assistant is basic. But the whole point of Home Assistant is that you can shape it around whatever you actually want to do with the data, so you can really integrate the device however you want.

One last thing worth saying: you don’t have to connect the device at all. Plug the Alpstuga in with no hub, no Thread and no app, and it runs as a standalone monitor showing its readings on the screen. For someone who just wants a number to glance at, this might be enough. I wouldn’t lean on it, though, and the reason is the CO2. The readings on the screen bounce around a lot and aren’t reliable moment to moment, and the value of this device really only comes out once the data is logged and averaged. So it’s nice that the option exists, but I’d treat connectivity as nearly essential.

Pricing & Competition

This is probably the easiest part of the review, because on price the Alpstuga doesn’t really have competition. IKEA operates at a scale that lets it buy Sensirion’s sensors for a fraction of what smaller-scale companies pay and it passes enough of that on that the nothing can really compete. At $34.99 you are getting particulate sensing, CO2, temperature and humidity, a display, and Matter connectivity in one box. If you buy the components at retail prices, a single decent standalone CO2 sensor module costs about as much as the entire Alpstuga does assembled. For under $35, you’d normally be choosing between an accurate CO2 monitor or a particulate monitor. Getting both, plus temperature and humidity, plus a smart-home hook, at this price wasn’t really on the table until now.

Here’s how it stacks up against the a few low-cost options, including IKEA’s own two earlier monitors:

Device	Price (USD)	Measures	CO2 sensor	Connectivity	Display
IKEA Alpstuga	$35	PM2.5, CO2, temp, RH	Thermal conductivity	Matter over Thread	Yes
IKEA Vindriktning	$15	PM2.5 only	None	None	Traffic-light LEDs only
IKEA Vindstyrka	$50	PM2.5, tVOC, temp, RH	None	Zigbee (via DIRIGERA)	Yes
SwitchBot Meter Pro CO2	$70	CO2, temp, RH	NDIR	Bluetooth, Matter via hub	Yes
Qingping Air Monitor Lite	$80	PM2.5, PM10, CO2, temp, RH	NDIR	Wi-Fi, HomeKit	Yes

When you read down the table, the point really makes itself. The cheaper Vindriktning does one thing and does it with no numbers and no connectivity. The Vindstyrka, the monitor that made IKEA’s name in this space, is a decent particulate sensor but doesn’t measure CO2 and costs more. The moment you want proper NDIR sensors you’re into Qingping Air Monitor Lite territory at roughly double the money. In that sense the Alpstuga carries the Vindstyrka’s legacy forward, taking the “surprisingly capable IKEA sensor for the money” idea and adding the one thing its predecessor always lacked.

None of this erases the flaws. and I’ve spent most of this review picking at exactly those things. But price is where all of that gets weighed up, and at $34.99 the Alpstuga is close to unbeatable even carrying every one of its downsides. It isn’t the most accurate monitor you can buy, but it might just be the most monitor you can buy for the money.

Conclusion

So, should you buy one? If you want both particulate and CO2 sensing in one device at this price, I’m not aware of anything else that does it. That alone makes the Alpstuga unusual. Add that it has decent connectivity and drops into more or less any smart home system through Matter, and you’ve got a monitor that punches well above what $30 normally buys.

The catch, as it has been throughout this review, is the CO2. If all you want is to spot trends and know when a room needs airing out, it does the job. Push it any further than that and you’re giving up a lot of accuracy, and the readings are noisy enough that you have to average them before they mean much. This is not the device to buy if an accurate CO2 number is the thing you care about most. For that you want a real NDIR sensor, and you’ll pay a fair bit more for it.

However, everything else holds up well enough, as long as you know the caveats going in. Particulate tracking is solid for trends, temperature is better than I expected, and humidity is usable once you know it tends to read a little low. None of these are best-in-class for accuracy, but none of them need to be for what this device is actually for.

Taken as a whole, and I’ll be upfront that the price is doing a lot of the lifting here, it’s a good monitor. It’s the cheapest way I know of to put an air quality sensor in every room of your house, and once the data is logging into a smart home system it’s more than capable of driving automations and giving you a genuine sense of what your air is doing. You aren’t getting NDIR-grade CO2 and you shouldn’t pretend you are. But for the money, as a way to blanket a home in air quality data, nothing else really comes close.

IKEA Alpstuga FAQ