Apparently it may not perform noticeably better (other than in speed of filtration) than a 2010 Mazda 3 or, even more surprisingly, random Uber cars in Shanghai and Delhi. I'm unclear how accurate or widely applicable the results of this test are, however.
Not that it's still not expensive, but the cheapest model 3 is currently $46k, which they plan to reduce soon to $35k. Even the current model is effectively already $36k after tax incentives in some states, and the TCO is even lower if you factor in gas savings.
Slight tangent: Can anyone recommend a decent air quality sensor in the sub-$100 range? I've been curious lately about measuring PM2.5, VOCs, etc. but haven't found a good option.
The sensor in the article looks really nice (Twitter thread mentions https://www.amazon.com/Digital-Quality-Detector-Tester-Monit... as something similar. Twitter thread also links the exact product on Aliexpress, but I refuse to link Aliexpress here since that site is nearly malware).
But the problem with that sensor and all the other sensors I've found thus far (HPMA115S0, SDS011) is I can't be sure how accurate at all they are. I'm not looking for scientific grade stuff here, but if the sensor is off by a factor of two, for example, it's a bit useless isn't it? Especially if I have no idea how far off it is.
If I'm reading it correctly they show that the SDS011 sensor is completely useless for PM10 measurements. For PM2.5 it can be off by a factor of 2 under certain conditions. They were able to come up with a calibration formula that makes it much more accurate (on PM2.5), but I have no way of knowing if their calibration formula is "universal" for the SDS011 or if the constants would vary from unit to unit.
And all of those problems on a device that is marketed as being +-15% accurate. Being off by a factor of two is well beyond 15%...
I haven't dug too deeply yet. Are there better sensors? The HPMA115S0 is brand name at least, but again I have no way of knowing for sure what its real error bounds are. Is there some way to calibrate these things at home without buying an expensive, more accurate sensor (thus defeating the purpose of the cheap sensor)?
It seems a lot of hobby projects end up using these cheap sensor anyway, even if they could be completely wrong? Maybe I'm being too picky and an error of 100% isn't a huge deal?
Thanks for those links- this is something I'm interested in as well. I'm working on a RasPi Zero hat that has a CCS811 for eCO2 and VOC, a BME280 for humidity, temp, and barometric pressure, and a header for a SDS011.
I'm interested in indoor air quality, so just looking at PM2.5 trends over time would probably be enough for me. An error of 2x without calibration makes it useless for any sort of air quality index calculation, though. Maybe I'll include a header for the HPMA as well and compare the two.
Thanks! I started this project on a whim a few days ago, and haven't done a ton of research yet; just picked what looked like the most popular parts online.
There was a thread on HN around a week ago around just this topic. Well worth reading through it "Build a do-it-yourself home air purifier for about $25" for the comments on filters and sensors. [1].
I bought an SDS011 on aliexpress for 15 euros. At that price it's worth sort-of calibrating it using other sensors or even only city weather data as a comparison.
26 comments
[ 2.6 ms ] story [ 63.1 ms ] threadApparently it may not perform noticeably better (other than in speed of filtration) than a 2010 Mazda 3 or, even more surprisingly, random Uber cars in Shanghai and Delhi. I'm unclear how accurate or widely applicable the results of this test are, however.
The sensor in the article looks really nice (Twitter thread mentions https://www.amazon.com/Digital-Quality-Detector-Tester-Monit... as something similar. Twitter thread also links the exact product on Aliexpress, but I refuse to link Aliexpress here since that site is nearly malware).
But the problem with that sensor and all the other sensors I've found thus far (HPMA115S0, SDS011) is I can't be sure how accurate at all they are. I'm not looking for scientific grade stuff here, but if the sensor is off by a factor of two, for example, it's a bit useless isn't it? Especially if I have no idea how far off it is.
For example, I found this research paper: https://www.researchgate.net/publication/322628807_Developme...
If I'm reading it correctly they show that the SDS011 sensor is completely useless for PM10 measurements. For PM2.5 it can be off by a factor of 2 under certain conditions. They were able to come up with a calibration formula that makes it much more accurate (on PM2.5), but I have no way of knowing if their calibration formula is "universal" for the SDS011 or if the constants would vary from unit to unit.
And all of those problems on a device that is marketed as being +-15% accurate. Being off by a factor of two is well beyond 15%...
I haven't dug too deeply yet. Are there better sensors? The HPMA115S0 is brand name at least, but again I have no way of knowing for sure what its real error bounds are. Is there some way to calibrate these things at home without buying an expensive, more accurate sensor (thus defeating the purpose of the cheap sensor)?
It seems a lot of hobby projects end up using these cheap sensor anyway, even if they could be completely wrong? Maybe I'm being too picky and an error of 100% isn't a huge deal?
I'm interested in indoor air quality, so just looking at PM2.5 trends over time would probably be enough for me. An error of 2x without calibration makes it useless for any sort of air quality index calculation, though. Maybe I'll include a header for the HPMA as well and compare the two.
By the way, how accurate is the VOC measurement on the CCS811? I saw those sensors on Adafruit but haven't dug into them yet.
I'd definitely be up for putting some assembled spares on Tindie once it's complete. I'll post about it on https://hackaday.io/hylian or https://edwardsh.in once things get going.
[1] https://news.ycombinator.com/item?id=18445652
If they had named it something else, it probably would have been better received.
"Rushing To The ER With A Respiratory Emergency Mode"
"Just In Case You Visit Infamously Polluted Cities Mode"
"Post Volcanic Eruption Ash Filtration Mode"
I foresee it getting unofficially renamed:
California Wildfire Mode
And calling other people who help out paedos...