Over the past couple of weeks, I’ve been at the AirGradient factory in Chiang Mai testing a wide range of air quality monitors for a series of articles I’ll be publishing soon. If you’ve followed my work on BreatheSafeAir, you’ll know that I’ve been reviewing and comparing monitors for years — but testing so many devices side by side, in such a short span of time, highlighted some patterns far more clearly than usual.

I had monitors from a mix of companies on my desk, covering everything from basic home devices to products marketed for schools, hospitals, and other professional environments. Problems started to emerge almost immediately. On day one alone, a couple of devices couldn’t be set up because the apps or cloud services they depended on had already been shut down. A few others technically worked, but only after repeated resets, re-pairing attempts, and firmware retries. All of this consumed more time than the actual testing and highlighted just how fragile many of these ecosystems are.

Alongside these failures were a number of surprising hardware findings. Quite a few monitors that presented themselves as premium instruments turned out to use the same off-the-shelf sensor modules found in far cheaper devices. In some cases the modules were simply hidden in oversized plastic cartridges to make them appear more proprietary.

These sorts of issues and discoveries aren’t unique to air quality monitors, but the high percentage of devices affected made them stand out during testing. After discussing the most surprising challenges and patterns I encountered, I wanted to highlight a few of the major issues. Some are small annoyances, while others are far more serious, and all of them are worth addressing.

Please note that the images in this article do not necessarily show the devices or companies I am referring to. Rather, these are images collected during the testing process (many of the devices with issues never get to the testing stage).

‘Consumer’ Monitors ≈ Professional Monitors

A PurpleAir monitor with the Plantower PMS6003

Anyone who has spent time around air quality monitors already knows that many consumer devices and so-called professional devices share most of the same components. I was well aware of this before arriving in Chiang Mai. It is common knowledge in the industry that a fifty dollar monitor can use the same PM or CO₂ sensor as a three hundred dollar monitor, and that even devices costing thousands of dollars may use the same basic modules. Achim has actually already discussed this in more detail in this blog post.

Despite the enormous price differences between these monitors (which partially could also be due to different services provided), they all use the same type of Plantower particle sensors (PMS5003 or PMS6003):

Monitor	Price
AirGradient Open Air (kit)	$125
PurpeAir Zen	$299
Clarity Node	> $1000 p.a.

This overlap isn’t inherently a problem. Plantower modules (among others) are widely used, well-understood, and perfectly capable when integrated properly, which is why so many manufacturers rely on them. We at AirGradient also rely on the Plantower PMS5003 and PMS5003T!

What surprised me most was not the overlap, but how some companies try to reframe these standard components as something proprietary or more advanced.

The Illusion of Premium Quality

Building on this point, another thing that became clear very quickly is how common Premiumisation has become in this space. Many brands rely heavily on premium-looking designs to position their products as professional tools. It is easy to understand why these devices feel more trustworthy. A higher level of design often implies a higher level of performance - just think about a Porsche, MacBook or a Leica camera.

What makes this especially powerful in the air-quality space is that most people cannot verify the accuracy of their monitor. Unlike a car, laptop or camera, where performance is immediately obvious, air-quality readings depend entirely on trust. A device that looks premium feels more believable by default — even if the underlying components are identical to those in far cheaper monitors.

A Plantower PM Module

And this is where the problem begins. Instead of clearly showing the components they use, many brands hide the actual sensor modules inside proprietary cartridges or custom-shaped housings. On the outside, these cartridges look specialised and high-quality. On the inside, they often contain the same standard $10 off-the-shelf sensors used in much cheaper devices — just wrapped in a well-designed casing to make them appear bespoke.

I appreciate well-made products, and there is nothing wrong with paying a bit more for something that feels solid and looks clean on a desk. The issue arises when design is used to obscure what is actually inside, especially when those proprietary housings also come with significant markups. Some cartridges cost 10x more than the underlying sensor itself, even though they add no measurement benefit.

Replaceable Parts That Don’t Actually Exist

Because so many of these devices rely on proprietary cartridges or custom housings, they also require proprietary replacement parts. You can’t simply buy the underlying sensor module on its own, even when it’s a standard off-the-shelf component inside. In theory, this isn’t a problem, because the manufacturer provides those replacement parts. Right? In practice, however, many of the promised “serviceable” components are either unavailable, only sold through slow and opaque sales channels, or simply never offered at all.

During testing, this became obvious very quickly. Out of the eight monitors I evaluated, one of them didn’t offer replacement cartridges at all, despite advertising them as a core feature of the product. The manual described them, the website mentioned them, but there was no way to buy them. Not online, not through support, not anywhere.

Another device technically had replacement components, but they were nowhere to be found in the official store. Instead, the only option was a vague sales form and a “contact us for pricing” message. That might work for institutional procurement, but it’s completely impractical for individual or smaller scale users who just want to maintain their device.

A third monitor promised a whole ecosystem of expandable cartridges — future gas modules, upgraded sensors, and additional measurement capabilities. None of them exist. The product has been sold for years, the documentation still mentions upcoming modules, but no such add-ons have ever been released. The device was marketed as modular, yet the promised modularity never materialised.

These situations all lead to the same problem: when parts are unavailable, incomplete, or locked behind opaque channels, the advertised “serviceable” or “upgradeable” design is meaningless. And once a proprietary cartridge can’t be purchased, the entire lifespan of the device depends solely on the manufacturer continuing to support it. That leads directly into the next issue.

When Lock-In Turns Into E-Waste

A few of the monitors I tested were only a few years old yet already unusable. The cloud services were offline, the apps were no longer supported, and many devices lacked local interfaces or screens, meaning that without the company’s infrastructure, these monitors simply could not function.

Perhaps the most worrying part wasn’t the shutdowns themselves, but how tightly some devices were designed around the assumption that the manufacturer would always be there. A monitor with no screen, no web interface, and no physical controls becomes useless the moment its companion app (and often servers) stop running. Even simple actions like viewing data history, changing Wi-Fi settings, or exporting measurements are impossible once the supporting software disappears.

“I’ve used some monitors that have suddenly gone offline for days at a time as the company suffers server outages or other connectivity-related issues. These cases highlight the fragility of the monitors and made me realise that, at any moment, the monitors could stop working altogether.”

Ethan Brooke

A different kind of abandonment came from brands that marketed long-term expandability. As I mentioned earlier, some of the monitors I tested were sold with the promise of future add-ons (such as extra sensors). Years later, many of these additions never materialised, leaving buyers with products that never reached the capabilities they were advertised to have.

What ties all of this together is how quickly a working device can become stranded. Even if the hardware is perfectly functional, a monitor that depends entirely on a cloud service, an app, or proprietary or upcoming modules becomes fragile by design. Once any piece of that ecosystem breaks the whole product collapses.

This fragility is especially worrying in the air quality space, where many companies are small, young, or have unclear long-term business prospects. During testing, I saw several examples of once-promising products that were effectively abandoned in just a few years, leaving owners with hardware that had no path forward.

I’ve heard the saying ‘don’t purchase a product based on promises, but rather on what’s available at a given moment’ many times, and while I think it’s a good rule to live by, it doesn’t make these false promises any less painful.

Unsubstantiated Accuracy Claims

Two of the tested monitors deviating significantly from our reference device

Another persistent issue - and the last one that I will discuss today, but by no means the last one that I encountered - in this space is how many brands claim high levels of accuracy without providing any evidence. It is common to see specifications like “±5 µg/m³ accuracy” for PM2.5, even when the underlying sensor has a published specification that is far less precise. This appears to be especially common with proprietary cartridges that appear bespoke on the outside but contain standard modules inside.

What makes this even more concerning is the near-total absence of supporting data. Very few companies publish co-location results, reference comparisons, long-term drift analyses, or even basic calibration notes. Some mention “factory calibration,” but almost none provide the data that shows how well that calibration performs in real-world conditions. Even when accuracy reports are requested directly, many manufacturers are unable to provide anything beyond the marketing sheet.

This lack of transparency becomes especially visible once you begin comparing devices. In some of our tests, monitors simply did not match the accuracy implied by their specifications. There could be many reasons for this — perhaps the manufacturer calibrated with a different aerosol, under different environmental conditions, or against an unspecified reference instrument. It’s impossible to know. And that is ultimately the problem: users are asked to trust numbers that appear authoritative but are rarely substantiated, explained, or repeatable.

Conclusion

Testing air quality monitors

Testing dozens of monitors has exposed to me something that often stays hidden: how much of this industry runs on opacity. When devices conceal their components, rely on cloud services to function, or advertise accuracy no one can verify, trust breaks down and the people relying on these devices are left with little to go on.

“This experience highlighted, even more, why I chose to work with AirGradient. The commitment to open hardware, transparent calibration, and avoiding vendor lock-in has always been a core focus and it’s the practical answer to the problems I repeatedly run into when testing other many other devices.”

Ethan Brooke

Air quality affects our health in ways we can’t see, which means the tools we use to measure it have to be honest, open, and resilient. That’s the direction I believe the industry needs to move toward, and this experience only strengthened my conviction that openness is the way forward.