Public Lab Research note

DIY: Wi-Fi Streaming Weatherproof PM2.5 Monitor (Dylos-based)

by kevinvivergy | December 04, 2015 23:10 04 Dec 23:10 | #12481 | #12481

Air pollution, specifically particulate matter 2.5, can vary significantly over space and time. Although the EPA's AirNow system is good in the sense that it has highly accurate monitors and releases the data hourly to the public, I was dissatisfied with its coverage in my area, the greater Boston area. There are 5 monitors that provide data to 4.7 million people... not very granular! Furthermore, this data is pretty impersonal- these monitors are not located in publicly visible places, and simply reporting an air quality value for a massive area is not going to make it feel very real to people after considering that air pollution is not in a high enough concentration to be visible here 95% of days.

So, I wanted to bring air pollution data down to a very human level so the average Boston resident could understand its relevance to their lives. My requirements for the monitor were:

  1. A technology that was proven accurate in independent testing for a reasonable price
  2. Live streaming to the web via Wi-fi for ease of sharing the data
  3. Weatherproof down to 0 degrees Fahrenheit, able to withstand the Massachusetts winter
  4. Power and internet demands that were easily available in the US
  5. Minimal maintenance for ease of use by end-user
  6. 100% uptime, no interruptions in service by device.

The following instructions show how I built V1 for approximately $500. I settled on creating a system based around the following 3 technologies:

  1. Dylos DC 1100 Pro with PC interface (link)
  2. CanaKit Raspberry Pi w/ Wi-Fi (link)
  3. Primex P1000 weatherproof case (link)

Other things you will need: -Reflectix Insulation (link) -Velcro adhesives (link) -Zipties -Incandescent rope lights for heating(link) -Trendnet Serial to USB cable(link) -3 Way Wall Outlet Plug Adapter (link) -Scissors -Box cutter

And it looks like this:


Here are 4 key resources I consulted to make this: 1. Drexel senior design project, outdoor Dylos system (link) 2. WorldAQI Build Your Own Wi-Fi Enabled Dylos Monitor (link) 3. Outdoor Cat Enclosure, heated and insulated for cheap (link) 4. Accuracy testing for Dylos monitor (link1)(link2)(link3)

Live experiment streaming currently at

The heating method is needed for proper functionality of the Dylos. In order for the Dylos to read accurately, it needs to be at a temp above 32 degrees Fahrenheit, I aim for 40 degrees to be safe.

Part 1: Connecting Raspberry Pi to Dylos/ Wi-Fi Enabled

  1. Set up your Raspberry Pi with Raspbian OS, as provided
  2. Connect your Raspberry Pi to the Dylos with the Serial to USB cable.
  3. Download the PHP scripts from the WorldAQI project (link) to the Pi and follow included instructions. You are also going to want to set it up so that they run on startup of the Pi, instructions at that link as well. Furthermore, you may want to stream the results to your personal server if you want to manipulate the data yourself. Check line 82 of /dylos.php to change to destination of data.

Finished picture:


Part 2: Setting Up the Enclosure

Start by lining your Primex P1000 with insulation, attached with Velcro.

Bare enclosure


Plan for velcro on Primex


And Reflectix Insulation ReflectixBottom.png


Finished picture with velcro attached (minus second strip of velcro for attaching incandescent lights


Part 3: Assembling pieces within enclosure

First, the lighting/heating. Coil the incandescent rope light into a loop that fits on the left side of enclosure, then use a zip tie to hold the form. Use two loops of Velcro to attach to the top, the attach to the two lines of Velcro "hooks" on top.



Pic 20151204_112535.jpg

Second, everything else. It is going to be a tight fit. You need to add one strip of Velcro "loop" side along the bottom of the Pi, then one strip "loop" side at the top of the Dylos and one at the bottom.

Diagram Boxlabeleddiagram.png



Make sure to remember to turn the Dylos on before putting it in place!

Feedback Needed/Next Steps

2 major things I am concerned with/ looking for more elegant solution

  1. Worried about re-circulation/separating intake and exhaust. I could not find a good way to create a separate tube or method for separating intake and exhaust for the Dylos, which is a big deal because I don't want the same air to continually recirculate through the system. To test it, last night I lit a cigarette next to one of the intake holes in the Primex to see how long it would take that particulate plume to circulate. Was pretty happy with results, you can check it out at, occured on night of 12/3/15. Would love to use a better method though.

  2. Heating method Using waste heat from incandescent lights is less than ideal. Would love to hear ways other people have made it work, I also tried a seedling heat map (link), but it took up too much space. Major constraints are space and cost. The method needs to keep the box above 40 degrees when the outdoor temp is down to 0 degrees.

My next step is installing a humidity sensor in the Pi to make sure it does not get too humid in the enclosure.


I plan on installing one of these at a school, likely an elementary school, to get some real readings and see the effect of cars and buses on particulates. Schools are ideal places for a couple reasons:

  1. Since it is inclusive for all parents/kids, great place to have a conversation about environmental health.
  2. Showing trends caused by personal decisions like car driving is a very accessible way to present the data, minimizes need to discuss atmospheric chemistry, brings the personal decisions/air pollution creation feedback loop down to a very understandable level.

Got any questions/comments? Would love to hear 'em. Leave a comment or reach me at kevin (at) joinvivergy (dot) com.

I did this Help out by offering feedback! Browse other activities for "open-air"

People who did this (0)

None yet. Be the first to post one!


if you're interested, we made Dylos drivers for Open Pipe Kit for the Pi-- its a flexible platform for pushing data to Phant and any other database, based on command line scripting.

Reply to this comment...

awesome resources and case! hadn't seen this summary of low cost PM sensors:

Reply to this comment...

Thanks @mathew! Yeah @jeff showed those to me, the AQICN one was pretty straight up too and I just went with theirs because I read their tutorial first and I was pretty happy with the documentation.

Reply to this comment...

@kevinvivergy great detailed post.

I'm not that familiar with the Dylos--haven't handled one--but looking at some photos maybe just cut out holes in the enclosure corresponding to the back of the instrument fan intake and exhaust vent. Hole saws are great for neatly cutting openings. Ideally that would expose the intake to the outside air and then the exhaust is plumbed out, yet you can still heat/cool the instrument enclosure without having to directly heat/cool the ambient air you're testing.

I do have concerns that the Dylos might not be made to operate 24/7, on constantly outside, even if kept from the elements--but try it!

For heating, a small silicone flexible heater makes sense. You can buy smaller versions than the one you have, but the cheaper ones also need thermostats. Watlow makes them but their online sales is virtually non-existent. McMaster sells similar Flexible Silicone-Rubber Heat Sheets which I'm not sure if these are Watlow or not. They require a thermostat wired in series, something like this. I haven't used these parts specifically, but have used the Watlow components to heat instruments. Alternatively, some pet supply stores my have little heaters made for lizards and other sm animal enclosures.

Like most on/off thermostats, the heat cycle is somewhat dependent on the distance between the heater and the thermostat so you might need to play around with the spacing to get it to work right--not over heat but not cycle constantly.

Ideally you'd have the heater PID controlled and record the enclosure temperature, etc. but that's not always practical or needed. Yet, some instruments are sensitive to temperature fluctuations and the heater cycling on/off can sometimes cause electrical line noise creating an artificial blip on your instrument reading, which the PID controller can help avoid that instant on/off electrical surge.

The low, low tech approach would be to really undersize the heater wattage then just plug it in without any thermostat and hope for the best, which I think is what your light bulb approach is doing. Inline fuses are always a good addition to a heating circuit to prevent a complete meltdown!

For summer you're probably going to want a cooling fan with filter. With these outdoor enclosures sometimes you just need to winterize them each fall (unplug fan, cover fan opening, plug in heater) and then summerize them each spring (unplug heater, remove fan cover, plug in fan).

Also, put a placard on it identifying it as a scientific instrument with your contact info just in case someone finds it unattended and calls the bomb squad.

Reply to this comment...

@Davidmack all good thoughts, thanks for the detailed response.

-The big challenge with intake thing is that, if I understand your approach correctly, the intake cannot be exposed directly to outdoor air in the winter because it will not give accurate readings if incoming air is less than 40 degrees. So that is why it is currently on the opposite side of the container, on the other side of the heating. I do like the exhaust suggestion- cutting a separate hole just to handle exhaust correctly.

-Really like the heater ideas, I don't have any experience really with the electric wiring stuff, but the pet store idea is a good one for an easy way to try it out.

-The summer/winter version is definitely a good idea, that is why I did everything with Velcro, so it could all be moved or taken out. Cooling fan should be an easy add.

-You are very right about the placard! I was thinking, since I am planning on doing this at an elementary school, that I could let one of the kids paint the outside with whatever design they wanted. Would make it more approachable/attractive for kids instead of a black box.

Reply to this comment...

I was thinking you could expose the intake to ambient temperature in winter and by heating the enclosure you'd be warming the instrument and air indirectly--but I'm not really sure if that would work or not. It's a little crude and you would probably still have issues in extreme cold and when there's a lot of fog. Below freezing the air is usually not wet enough to cause an issue unless you have snow or ice blowing in but you could prevent that with a baffle.

Does the Dyos have a temp requirement?

Moisture is always a problem with light scattering devices and the cheap way to compensate for condensation is to heat the air to lower the relative humidity (RH). So ideally you would only heat the intake air when the air exceeds ~95% RH. But this would require a RH sensor with a heat controller. You could offset the intake with a pipe (anodized aluminum, non magnetic is best) and then wrap the flexible heater around the pipe, which a metal pipe would conduct the heat better than say PVC.

In general you want to "handle" the air sample as little as possible. When you heat the air you risk changing its composition: particle sizes can change; and carbon, particularly semi VOC can burn off. You want to warm the air when necessary but not over cook it or burn it. Plus anything in the enclosure is a potential PM emitter and source of contamination as you move it from site to site.

Also, not to be a kill joy, but you don't want to paint the enclosure. It would be another potential source of contamination.

Is this a question? Click here to post it to the Questions page.

Reply to this comment...

Yeah I was thinking the same with the paint, introduces another source for pollutants.

When I called the Dylos folks, they said that the device will not break at temperatures below 32 Fahrenheit, but it will become increasingly inaccurate, so they recommended shooting for 40 degrees Fahrenheit to be safe.

Reply to this comment...

I'm revisiting your note, just want to say again how cool this is! great to see you're still up and posting data.

Reply to this comment...

Thanks @mathew! Has certainly been an adventure. Only consistent issue has been school internet security protocols- they treat the Raspberry Pi as an unknown device and block the posts. Fortunately we have been able to find a workaround at both schools.

I was planning on blogging about the results, but there has not been anything particularly concerning thus far! I am trying to figure out if the monitor is placed in the wrong place, or maybe the buses are just not putting out that much PM2.5. Trying to do the same thing with a NO2/CO sensor as well, but I am having a hard time finding one that is accurate and reasonably priced. Monitor currently located about 100 feet from bus dropoff/pickup.

Reply to this comment...

hah! school networks are always a problem.

I posted a note recently on hotspot identification, the first stage of which involved a systematic survey with an optical monitor (like the dylos). I'm new to the idea, but wondering if these kind of 6-minute "snapshots" of different spots couldn't help pin down where the emissions are going.

Reply to this comment...

Wow that is super interesting. That would definitely cover the variation over space... but what about if the levels vary over time in each particular place? Even a couple 6 minute sample times would not be able to cover variation over the course of the day.

Unless that is part of the author's point, that variation over time is minimal compared to variation over space?

Is this a question? Click here to post it to the Questions page.

Reply to this comment...

the authors did two levels of survey-- first the 6 minute 'snapshots' and then they used that to determine where to place passive monitors for long-term characterization. The second heat map is based multi-week monitoring at each site, using 7-day passive samples. So lower temporal resolution, but capturing the week-to-week average exposure hotspots.

Temporal variation is a big deal, but here they were just trying to demonstrate a means of characterizing spacial variability on the cheap using monitors that don't require power.

Reply to this comment...

Ahhh I see. And it looks like both ways showed that there is a big plume of particulates coming out of the rock quarry that dissipated over the right side of town.

Reply to this comment...

Great discussion and a lot of good ideas. The AQ-SPEC group ( has published results on several PM sensors other than the Dylos. Particularly interesting is the performance of the "Purple Air" sensor ($150/unit) and the widely used Alphasense OPC-N2 monitor (which provides 3 size fractions and 16 size bins). I am personally more familiar w/ the hardware than the software (just stare using the RPi) and would be more than willing to collaborate with people that have good ideas on how to build a small sensor network. I have been looking into AWS IoT and products like Splunk for structuring, sorting, and plotting sensor data. Anyway...I am digressing. Great job w/ this Dylos project and keep up the good work :)

Reply to this comment...

Thanks @drapoli! Yeah, when I built this thing I had not heard of Purple Air, I am really interested to learn more about it. The AQMD site has done a really good job of staying up to date and keeping up with new tech.

Reply to this comment...

Hi @kevinvivergy, i have just tagged this note as an activity that others could try as well. I did this by adding the tag activity:open-air. If you could scroll up to the tag area on this note and click "click here" in the blue box to fill in the other details, it would help people find these instructions and try to build one themselves. Thanks and let me know any questions, i can support.

Reply to this comment...

Can do @liz, any in particular you would recommend? Not familiar with the system!

Is this a question? Click here to post it to the Questions page.

Reply to this comment...

Hey @kevinvivergy it's so great to hear from you!

Here's the blue tagging assistance box that you will see when you click on "click here":


...Within those questions, there are clickable links for finding out more about what each tag means.

Any more questions? Also, feel free to call me at +1 336-269-1539.

Is this a question? Click here to post it to the Questions page.

Reply to this comment...

Login to comment.