Monitoring my home's air quality with AirGradient's DIY sensor
jeffgeerling.com
jeffgeerling.com
- Cooking is insanely bad for the air in your house, even with an outside-venting hood. Any sort of medium-high heat cooking will spike the PM2.5 to multiple times that of the haziest days here in NYC, and with the windows shut it will stay elevated for hours after.
- The 20" box fan + MERV furnace filter hack is super effective for cutting down on the PM. A few days this July the wildfire smoke from the west coast settled on NYC and we had 40-70 ug/m3 PM2.5 outside. The box fan filter took it down to around 5 in a matter of minutes. I didn't even bother taping the filter to the fan.
- My A/C system's ability to precisely hold the temperature within fractions of a degree of the thermostat setting is impressive.
- With two people working from home in an old, unventilated-but-kinda-drafty ~700 square foot apartment, the only thing that gets the CO2 levels into "drowsy" territory (my original point of curiosity when buying the sensor) is running the gas oven for a while.
I found bungee cord with J hooks works really well. As you correctly point out, you don't need to secure it at all, but it is annoying to tap the box fan and have the filter fall over.
https://news.ycombinator.com/item?id=27125060
tl;dr: seems like you want an NDIR sensor (https://en.wikipedia.org/wiki/Nondispersive_infrared_sensor).
Some options, with digikey @1 prices:
Senseair S8 (Sweden): $48
Sensirion SCD30 (Switzerland): $63
Amphenol Telaire T6713 (US): $96
If I were to buy an expensive black box iot solution, I'd at least verify that it had an NDIR sensor.
It's pretty tiny... I wonder if/when it will be possible to put a high quality co2 sensor in an IP68 iPhone...
With any IoT device I’m wary about longevity re app support, I can’t say how long Awair will support this device or whether they’ll provide access to it after they stop, but I’m less concerned for a device like this vs, say, a major appliance.
The problem with all this "intelligence" is the Internet is very good at showing you search results for what you want to see, so Googling "Should I be worried about fine particulates" is a search destined to turn your anxiety into a full-blown panic attack.
No idea what the answer is, and I am impressed with the level of effort that went into the Awair app, but it would be nice to have more reasonable context in all IoT devices. Yeah, I know.
I’ve also found that regularly glancing at the 5 minute resolution charts of just temp and humidity has helped me get a better sense of my own comfort range…for example I like to run my AC as little as possible in the summer, but I’ve learned that playing chicken with the heat is foolish beyond 81F as I only wind up sitting at my desk grumpy and unproductive.
It’s got an eink screen, and I haven’t had to recharge it since I got it 8 months ago. I really like it.
But the parent is not inaccurate, for they did not mention "each year". Even 10 minutes North of Ottawa, I saw -42C, real, not just windchill, for a week, about a decade ago.
https://en.wikipedia.org/wiki/Highest_temperature_recorded_o...
But when it's 65-75°F outside in the spring/fall... CO2 builds up in my apartment like crazy. I keep the windows open and my bathroom fan on to try to suck fresh air in... but even so can't manage to really get it below 650, and with just a window cracked it sticks around 850.
It's really frustrating.
[1]: https://en.wikipedia.org/wiki/Energy_recovery_ventilation
I could add the door cut back in, install return transfer grill, or look into adding a secondary heat recovery ventilator that acts independently of the primary HVAC system.
Or they buy new high efficiency windows and doors, and that 'tightens up' the house, but they don't also install an HRV (Heat Recovery Ventilator) to exchange outside air.
Having the foam crew do a blower door test for free if we go with them, so I guess I'll see. I live in climate zone 5a (cool, humid) so historically natural ventilation was used via induced pressure differentials generated by mechanical means (dryer, furnace, bathroom and kitchen vents). When we get new furnace and water heater that pull combustion air from outside, we're left with mostly manual means for generating negative pressure that drives infiltration. (But also wind, and pressure differentials caused by temperatures in winter)
In our case, we're definitely considering either an HRV or ERV to supply fresh outdoor air that we control (input and output CFM, MERV #). Our indoor air quality is poor enough to the point where we live with the tangible symptoms (Rhinitis), not just numbers on a sensor.
For those with interest in these topics, I suggest skimming a copy of one of the ASHRAE fundamentals handbooks[1]. It reads as a textbook but has incredible depth into the physical sciences that relate to building science, heating, cooling, and ventilation.
[1] https://archive.org/details/ashraehandbook2017hvacfundamenta...
And even if the installer air sealed your doors and windows, there are plenty more places for air to leak in.
You really need to run a blower door test to see if you need an HRV/ERV.
That must vary by region. I live in the PNW, and I have not seen a recently built house with a return in each room. One or more supply vents per room depending on size, one return for each floor seems pretty much standard.
But I wouldn't be the least bit surprised to find that there isn't really a specific reason, it's just "how we do things in this area." A surprising amount of building choices seem to be made that way.
We typically have returns in toilets, showers, etc. with one or more supply vents in each room; a large room may have several. Air flows in via supply vents, flows out under mandatory gaps in doors (if closed), through to the toilets and then outside. Moisture, CO2, and smells are constantly flushed in the right direction.
When the hood is on for cooking, air supply is typically enhanced so it doesn't mess up the overall pressure too much.
Incoming air is filtered, of course. The ventilation unit transfers heat from outgoing air to the incoming air to save energy and avoid problems with condensation.
In my home office, TVOC as measured by an Eve Room rarely goes above 100 (mostly happens when I pour an occasional whisky in the evening).
Old houses don't have ventilation machines at all, but rely on natural ventilation instead: warmer air flows upwards via chimneys and such due to gravity, pulling in fresh air from whatever gaps the structure has (accidental or purpose-built).
I guess the point is that it depends on when and where something was built. My most recent home is only a few years old and doesn't have a return per room and no code to say it should shouldn't.
https://homes.lbl.gov/ventilate-right/step-3-whole-building-...
During Christmas Day visits with only a half dozen people the room CO2 spiked to over 1,000ppm. Older house from the 1970s no HVAC. Opening a window (cross draft) brought it down to 300ppm. The room was stuffy before and great after so it shows how sensitive we humans are to air quality.
Hold on! Not all particulates are equally harmful - and by far!
A particulate sensor detects water vapor or particles of (non-burnt) cooking oil that are harmless to humans.
In the same way, it detects very dangerous (carcinogenic) compounds from combustion engines or burning coal or wood.
Cooking is in no way "insanely bad" (unless you are grilling your food black or cooking on an open wood/coal fire)
I have no idea what the particles are though -- metal? Paint? Carbon from burnt residue? Or just whatever the heating element itself is made of?
Separately, cooking with gas (oven+stove) also sends CO2 way way up -- in my kitchen, from 500 to 1300 or so within 20 minutes. And that's with my range hood on.
Not saying that's the case here, but several times I thought the range fan was exhausting the air, only to learn upon examination that it was just shooting it back out into the same room.
Our current place has a proper kitchen, but the cooker hood vents outside directly next to the central heating boiler's intake.
You can't exactly throw the toaster oven in the sink and scrub the inside surfaces with steel wool and Barkeeper's Friend under running water like you can with pots and pans. Nor do toaster ovens have a "self-cleaning" cycle where you try to just vaporize it all... they simply don't get hot enough.
"Natural gas and propane stoves can release carbon monoxide, formaldehyde and other harmful pollutants into the air, which can be toxic to people and pets."
https://ww2.arb.ca.gov/resources/documents/indoor-air-pollut...
"Gas stoves can generate unsafe levels of indoor air pollution"
https://www.vox.com/energy-and-environment/2020/5/7/21247602...
Also, the dangers of open flames are well known, unlike the ones of particulate.
>“We know that inhaling particles, regardless of what they’re made of, is detrimental to health. Is it equally bad as inhaling exhaust from vehicle emissions? That we don’t know that yet,” [1]
Regardless, my sensor, at least, doesn't register water vapor as particulate. I can run an ultrasonic humidifier with distilled water to no effect on the measured PM2.5, or with tap water which will spike the particulates from the dissolved minerals.
1) https://www.theguardian.com/environment/2019/feb/17/cooking-...
Now I really need to investigate more details - like does induction cooking (vs. gas) improve the quality? What about the type of food (ie, cooking vegetarian vs. grilling sausages)?
But to be fair, you may have a good point. Vehicle emissions and tire wear generates particulates consider to be ultrafine (0.1 max nanometer), which is even difficult for HEPA filters to deal with.
Even just with two people working from home, much of the space would quickly get up to 800 or 900 ppm. I’d usually open windows as I saw it creeping up, which would quickly reduce CO2 to under 600, and then it would slowly creep up over the day.
Living here in the west with hyper toxic AQI for the better part of two months now, I decided that the low CFM and leakage of a DIY box fan are not adequate to keep up. I went with a far more industrial solution: I went to an indoor grow shop and bought an activated carbon filter and high CFM inline fan. Cost $150 and is a bit noisy, but has been well worth it.
https://www.amazon.com/VIVOSUN-Inline-Controller-Ducting-Ven...
That’s fairly high - what is the size and cfm of the fan you bought?
>The EPA funded a study by Underwriters Laboratories (UL) which tested box fans with MERV-13 filters attached, finding them to not present a fire hazard even in extreme, modified conditions.
Then it links to, https://chemicalinsights.org/wp-content/uploads/DIY-Box-Fan-...
which says,
>Two-Sided Obstruction
>Figure 6 shows the time series of monitored temperatures under extreme condition where both sides of the fans were blocked. Despite the severe blockage of air flow, temperatures of all five box fans peaked within the first 20 minutes and then plateaued, remaining steady for the remaining 7 hours. Winding temperatures had the largest relative increase of 30 to 57 °C above ambient temperature, with peak absolute winding temperatures remaining at or below 80 °C. Peak temperatures on other fan parts did not exceed 50 °C.
> Temperatures at the power cord, guard, and switch, remained close to ambient room temperature (±6 °C) during all filter test scenarios as well as during the one-sided blocked tests. For the two-sided blocked tests, temperatures at the power cord, guard, and switch increased initially before plateauing to a steady temperature for the remainder of the 7-hour test duration. All fans remained operational for the 7-hour test duration and did not show any signs of potential hazard or damage.
Do different PM2.5 components affect the body differently, even for the same overall count?
Eg, are the PM2.5 particulates from burning toast comparable in bodily health impact to a similar PM2.5 count coming from a heavy Diesel engine’s exhaust (which I presume would have a different particulate makeup).
[1] https://ehp.niehs.nih.gov/doi/full/10.1289/ehp.0800185?url_v...
weeps in searing steaks with no hood at all :'(
I’ve found this varies quite a bit. Though, I don’t do a ton of high heat cooking, so maybe there’s no way to do that without pollution.
But at medium to low temperatures, you can get crazy variance in AQI output:
* Lack of enough oil/butter in the pan will send levels soaring
* Cooking one thing in the pan then another after is very bad. Eg I used to cook sweet potatoes then add meat after and not clean in between to save waiting. Boy was that bad for the air. I think it’s because tiny bits of food you may not even see stick to bottom of pan and burn
* Browning is when it gets worse. Any kind of browning on foods spikes pollution.
* Any food particle that gets on the burner will invariably lead to hazardous AQI levels. Like even a single fleck of meat left cooking will fill the house. You may not necessarily smell it as much as other things
* Turning heat off at end of cooking works wonders. I do this with bacon and it takes it from a hazard to not even spiking aqi. Finished product looks and tastes the same
What is the alternative?
I haven’t tested AQI with steak for example. And I can say that with bacon I have certainly managed to have it turn brown without pollution. But, once things are turning brown you are on thin ice and it is easy to go from “none” to “hazardous” in a flash.
And what is the boxfan hack? I live near very bad wild fires and I need a much better filtering then my current grow filter which doesn't seem to work that well.
Literally just a box fan w/ a hvac filter taped to it. There are a few articles/youtube videos where people claim it works a lot better than a fancy air filter+fan device that costs $100s.
If you try it, just be careful it doesn't burn the fan motor and/or cause a fire.
you just put a fan filter on the back, you can increase the amount of filters to not hurt the fan motor as much by making a triangle config or fan on top of 4 filters in a cube. fans are for air throughput not static pressure so real air purifiers have squirrel cage blowers
I used the sps30 from sparkfun, CAD'd up a panel mount, 3d printed it, built a wooden box for it + a pi, set up all the networking and hosting, and packaged my utils for debian.
Anyway, the AQ is fine, HEPA filters work when in-line with the A/C, and there's nothing much to monitor even when AQ is mod/bad outside. Boring, but nice to know, and fun to learn all those new skills.
Just don't burn incense.
I have some beeswax ones coming soon which are supposed to be better; my PM25 meter is standing by for their arrival.
Also they have an optional 24-hour timer so they'll remember to turn on the same time you turned them on yesterday.
https://www.potterybarn.com/products/flameless-outdoor-candl...
But I just like the 'ambiance' too much. There are always different levels of risk adverseness: if I was considering candles a pollution health hazard I should also stop riding motorcycles, vaping various things, and stop eating like I do...
Purple air sells a turn key indoor air quality device for $200, is very reliable, also comes in a (clear) 3d printed case, comes with power supply. Not as DIY as this guy's solution, but with limited time on my hands the $100 was worth the cost of admission. It also works with Purple Air's map which provides nationwide air quality, temp and humidity, etc which is really nice.
It's worked flawlessly for me now for about five months. The light turns from green to yellow when I open the window in my south facing condo, but turns green again when I close the window and turn on the air filter. I was surprised at how bad the air quality gets when cooking food on the stove.
When you are DIYing this sort of thing, I would think you would need to make at least a couple of them to see if they calibrate to the same measurement, assuming what you are after is greater precision and control (why else would you DIY it?). Even then, both devices could calibrate to the same very imprecise value.
To really validate this, you would need to put it in an controlled environment with known air pollution levels. Just placing it outdoors when the weather report says 200AQI wouldn't work since that's an average over a large area, not a point measurement.
I imagine the sensor manufacturers have done some calibration - the article mentions a self calibration period for some, but once everything is in the final package, I'd expect the behavior to change.
Without that, the readings from the device become more of a crapshoot, which is fine if there is no desire for greater precision (i.e a just for fun project)
Sensor manufacturers cannot calibrate for every possible packaging configuration.
As someone who has DIYed a lot of stuff related to indoor climate, I tend to save my effort for the mitigation work, and use well tested measurement tools to validate those. This is for the same reason that one doesn't usually DIY a multimeter before doing electrical work.
http://www.aqmd.gov/aq-spec tests and compares commercial AQM vs lab grade instruments. It isn't quite enough to compare a commercial unit with the same sensor, because you don't really know the mfr calibration (if any), but this will get you a much better idea of performance.
i personally prefer to have a commercial sensor that has an API, rather than DIY.
I know this is a very “non HN” thing to ask. But, as much as I’d love to learn and tinker with Adruino, sensors and their soldered connections, supplying power correctly etc, I am afraid all my spare time would be gone to it, and I will never get to the air quality part of the project. It will be yet another wide scoped project that never ends. Hence the ask for a non DIY yet Prometheus compatible solution..
https://www.aliexpress.com/item/1005003019471995.html?spm=a2...
Can be a bit tricky to get hold of the security tokens but at least doesn’t involve re inventing any wheels. One is very much at the mercy of firmware updates blocking diy access. General recommendation is not to update.
Plus US crowd is pretty paranoid on huawei stuff ofc
If the huawei thing is accessible in the local network, that is almost exactly what I want. I can connect it to a raspberry Pi on an isolated network and extract the data I need. Thanks for the tip.
It claims to not emit any Ozone, but I'd like to be able to meaure it, and the commercial sensors tend to be expensive.
[1] There are a few different kinds of MQ131 (for high and low concentrations); you'll probably want the low concentration one. I also removed the bare sensor from a breakout board I purchased so I could access it directly from my custom PCB.
[1] https://www.sainsmart.com/products/mq-131-gas-sensor-ozone-m...
[2] https://learn.kaiterra.com/en/resources/ozone-what-levels-ar...
If you haven't smelled it before - I would recommend going to a retail store that sells ozone-based air sanitizers. That should give you a whiff of the stuff for reference. You can also find it in [some] hospitals for air sanitization. I imagine once you smell it, you will realize you have smelled it before and maybe not realized it was ozone that you were smelling.
My air purifier is built into the furnace ductwork so it's already pretty diluted air by the time it comes out of a duct.
That's neat! Do you know if the ozone is produced before or after it reaches the blowing mechanism? I ask because if it is produced before, like right after the air intake, that's a really great setup for keeping your ductwork mold and mildew free. I've wanted to buy such a system but keep holding out for things to get cheaper... in the meantime, I use an ecoquest "FreshAir" model sitting on top of the fridge.
The unit is installed after the filter, but before the air is sucked into the blower.
[0] https://twitter.com/dcuthbert/status/1421822958604062726
It publishes its data via MQTT, and you can use any system that can subscribe to a MQTT queue to process, display, etc. it.
$5/sq inch and you get 3 copies of your board
I saw this post online, I’ve had the pm2.5 sensor sitting around for a long time…
https://benjamincongdon.me/blog/2021/08/08/DIY-Air-Quality-S...
Found my tvoc is constantly in .70-1.0 mg/m3 unless I open all windows and run fans. Then it drops to around .6 mg/m3. I now don’t worry at all about pm2.5 levels and just keep my windows open if the temps below 100. Buying the monitor showed me that my migraines and other issues (brain fog) were caused by the high voc. Opening windows even if it means breathing wildfire smoke makes the headaches disappear.
Anyone out there on the fence about buying a monitor, just do it. There are worse issues than pm2.5 and smoke, both of which you can detect with your nose.
Have you considered turning it around?
What I'm planning on doing is soldering a jumper to the board and having the temperature sensor hang out the side of the case a bit, towards the bottom.
With cheap probes and peer-review software (properly packaged in Linux distributions).
Used to be luftdaten.info (German), but seems more international now. They built a pretty accurate sensor based on an SDS011.
No soldering required.
And Tindie has a slew of ready-to-order air quality stations too.
The new ones use the Honeywell sensor. It has a tiny fan on it (inaudible in a quiet room from 5 inches away); mine ran well for a couple of years until I knocked it off my desk. After that, the fan started rattling, so I ordered another Honeywell sensor ($40) and replaced it (no soldering required, just a screwdriver). That sensor's datasheet doesn't have any cleaning instructions, so I don't think it's a big issue.
Plus, my living in Canada means getting all the different components piecemeal, often from different suppliers/sellers is a giant PITA. This is further complicated by my self-imposed restriction to not buy from Amazon. And this is before the headache of shipping costs (often $15 as the lowest option, even for small parts)
I'd love a ~$200-300CAD pre-assembled or kit option where I can own the data, access it in a standard fashion, and optionally choose to crowd-source it.
Those sensors guesstimate CO2 concentration from air temperature and changes in ambient alcohol/acetone/hydrocarbon levels, which drifts into positive infinity and is simply do not correlate well with actual CO2 level.
Those sensors are dirt cheap, and can still be used as a rule of thumb to encourage regular ventilation, but are nothing more than that. If you want to actually see CO2 levels, avoid those.
It has 4 sensors: temp, humidity, CO2 and VoC. Helped me discover a problem with my attached garage leaking fumes into the house.
But soldering has a very low cost-of-entry, and even using solderable breadboards over solderless can make your project a lot easier to use (no more worrying about lose connections or capacitance issues)
This seems like it could either compete OR complement an Airthings Wave Plus for like 35% of the cost. Since they both offer different and overlapping data.
I wonder what you'd need to do to get Radon and VOCs too in this solution (essentially match and exceed the Airthings Wave Plus)?
* Radon * Carbon dioxide * Carbon monoxide * Nitrogen Dioxide * Sulfur Dioxide * Ozone * PM2.5, PM10 (any others?) * VOCs * Temperature * Humidity * Pressure
It is an opportunity for someone to make a "suite" sensor so you could scatter them around the house for good average measurements.
Way overpriced vs AliExpress but oh well.
Definitely going to pick one up when they are available. I might save a few bucks diong the DIY route, but at the end of the day, I would just buy 3-4 more shares of a vanguard thing, and have to tinker/maintain a grafana dashboard, where this thing is probably getting new apps/updates/features monthly.
Their app REALLY wants you to turn on tracking all the time, with a flashing unkillable banner at the top of the screen.
1: https://www.aliexpress.com/item/32529101036.html?spm=a2g0o.s...
I've also updated the link on my site.
https://partofthething.com/thoughts/weather-and-air-quality-...
EDIT:
> Radon is the second leading cause of lung cancer in the United States. Nationally, the EPA estimates that about 21,000 people die each year from radon-related lung cancer. Only smoking causes more lung cancer deaths.
You gotta be kidding me, that's wild.
Since air holds more water at higher temps (and a lot more, 90F air can hold almost 3x as much as 60F air), just comparing humidities without considering temp doesn't say much.
https://www.weather.gov/lsx/monthTab
It was in the upper 80s/lower 90s throughout July and August, when this data was likely collected. And relative humidity hit 80-90% every morning, averaging about 65% as it warmed up. It's sticky in summer. I live in Michigan, and even with a cool breeze from our magnificent freshwater lake, half an hour inland it's similarly hot and humid. I kept our house at 78 when we're home, 82 when it's just the dog, and 76 when company comes over. I think the air conditioner keeps humidity in the 40% range, though I'm reading this article because I lack a hygrometer/air quality sensor and don't really have hard numbers for that, only that humidity is significantly less indoors than out.
Also, he mentions that the orientation of the case causes convection off the ESP8266 to increase the air temperature by a couple degrees (Farenheit).
If you're one of the Silicon Valley HN readers looking to move to the wide open spaces and relaxed culture of the Midwest, do be aware that climate is a big deal; coastal California's perpetual mild climate is an aberration.
March spring hits and the pollen is crazy until June and by then it's too hot to really be outside without tree cover from 9am-8pm until around now.
My area is one of the worst pollen places in the country. You owe it to yourself before moving to look at a pollen map as well.
Likely because of this.
Understand, the average high is in the mid to upper 90's and trying to get a home a/c to cover a 20+ temperature difference is difficult and significantly drives up the monthly power bill.
When I lived in the desert, 82° was comfortable for me. Generally, the humidity was below 10% (NWS reading, as most cheap consumer sensors are not precise below 10%). But even in monsoon season, when the humidity would spike, it was still fine.
I guess you get used to it. My wife never did, so when she came home from work she'd turn the temperature down to 78°.