4 ms·

I went down the rabbit hole of air quality testing devices a couple years ago. The bottom line, AFAICT, it is not easy to accurately test air quality cheaply. W

I went down the rabbit hole of air quality testing devices a couple years ago. The bottom line, AFAICT, it is not easy to accurately test air quality cheaply. Whenever I would find a device that claimed to test air quality I would find reviews by people who were testing that device that stated either; It came miscalibrated from the factory. OR After a short period it would no longer be calibrated. Turns out keeping an air quality testing device calibrated properly is really really hard, and definitely not cheap. So maybe the computing side of this is not really all the difficult, but the physical measurement side of this is absolutely not easy. Eventually I gave up trying to find a device for my home because I realized I would have to really dig into the science hard, and I would have to spend more money than I was willing. And I was unwilling to do either.

jschwartzi 6y ago
I've not worked specifically with contaminant sensors, but I know oxygen sensors require very specific and very consistent flow rates across the element. The type I've worked with used a ceramic element which absorbs oxygen and changes resistance as a result. So increased or decreased oxygen results in increased or decreased resistance. Using a sensor like this to measure percent of oxygen in a gas requires you to eliminate a lot of confounding effects like pressure waves, flow changes, etc.. Basically you can cause the sensor to absorb more oxygen than is present in the gas by temporarily increasing flow. So if you're building an air quality sensor you have to pay very close attention to how the air moves through the sensor. That requires not just a calibration but also characterization of the flow through the system under environmental conditions. One other effect I could see is a contaminant sensor becoming fouled. Oxygen sensors can become fouled and report inaccurate values, which is why it's very important to control the inputs. This is one of the reasons an antifreeze leak in an engine block can cause your engine performance to go to crap. The oxygen sensor is part of how a stoichiometric ratio of fuel and air is maintained, by monitoring the combustion products for excess oxygen. So when it gets fouled either the ceramic element is damaged or the sensor stops receiving consistent flow and the resistance relationship ceases to be valid. In short I don't see how you could build a useful long-lasting sensor by just buying parts off Amazon or Digikey and throwing them together. You would have to treat it like a scientific instrument, meaning that you have to validate your assumptions about how the sensor performs and I doubt that most of the products on Amazon do that.

jschwartzi 6y ago

I've not worked specifically with contaminant sensors, but I know oxygen sensors require very specific and very consistent flow rates across the element. The type I've worked with used a ceramic element which absorbs oxygen and changes resistance as a result. So increased or decreased oxygen results in increased or decreased resistance. Using a sensor like this to measure percent of oxygen in a gas requires you to eliminate a lot of confounding effects like pressure waves, flow changes, etc.. Basically you can cause the sensor to absorb more oxygen than is present in the gas by temporarily increasing flow. So if you're building an air quality sensor you have to pay very close attention to how the air moves through the sensor. That requires not just a calibration but also characterization of the flow through the system under environmental conditions. One other effect I could see is a contaminant sensor becoming fouled. Oxygen sensors can become fouled and report inaccurate values, which is why it's very important to control the inputs. This is one of the reasons an antifreeze leak in an engine block can cause your engine performance to go to crap. The oxygen sensor is part of how a stoichiometric ratio of fuel and air is maintained, by monitoring the combustion products for excess oxygen. So when it gets fouled either the ceramic element is damaged or the sensor stops receiving consistent flow and the resistance relationship ceases to be valid. In short I don't see how you could build a useful long-lasting sensor by just buying parts off Amazon or Digikey and throwing them together. You would have to treat it like a scientific instrument, meaning that you have to validate your assumptions about how the sensor performs and I doubt that most of the products on Amazon do that.