(Bad :) air quality sensor

[Brad Midgley]

Some thoughts on the air quality project.

Let's make this the flagship product of the microcontroller meetup. It touches on every kind of problem we solve.

Let's select a platform. I'm interested in low cost wifi enabled so I think esp8266 is a clear winner. We need a sensor.

I'm willing to have data quality be secondary in initial phase. We can have people set these up in lots of different circumstances and not worry as much about rigorous deployment initially.

I'm contacting the University of Utah to see who we can partner with.

I imagine a sign saying a data collection point is nearby would be great publicity and might even encourage better behavior. Of course it will have a web page and QR code.

Unless there is a better model out there for connecting, let's have the device try to acquire a connection and if that doesn't work, go into its own access point mode. If users have a cellphone they connect to it, it can use geolocation API to ask the phone where they are and remember it along with access point password. It could register an email as well.

Let's talk power. I think we need to start with three AA batteries and go from there.

I'll buy a couple of these since I'm out of boards for new stuff. http://www.ebay.com/itm/like/201607584855

Brad

[Beth Sallay]

Can it tweet/email a few people every time it reaches a certain threshold? Then occasionally change the from address for when the governor block the messages. 

[Beth Sallay]

oh derp I just posted this to slack on the uaf channel but I do IT support for Occupational and Environmental Health and Safety for the U and I have a list of software these guys use, one of them is having problems getting the air quality one to work with windows 10, quest pro 2, . I know what software they use, but for their higher tech equipment. I'll ask for advice, one is a huge star wars fan.

On Thursday, February 23, 2017 at 8:40:24 AM UTC-7, Brad Midgley wrote:

[Scott C. Lemon]

Brad,

I like the idea, but the costs and type of air quality transducer are going to be the key for this.  As you said, it might not be the most accurate or truly useful, but could be cool to see working.  There are a couple of things to think about:

1. Although the ESP is an interesting module, it can only provide Wifi connectivity, and no GPS.

2. I'm not usually a proponent of Bluetooth and cell phone connected sensors, but this *would* provide the ability to get GPS, along with the data, and provide real-time access to the data.

3. There are a wide range of gas sensors, at a wide range of costs, but then we have to be honest about their characteristics.

 - CO2 sensors are still slightly pricey - the last batch I got were still $20 each

 - volatile gas sensors also are same range, and are not actually accurate unless you know the gas they are in the presence of

 - cheap volatile gas sensors could be $5-$10 each

4. All of this is without the ESP, Battery Holder, Batteries, etc.

Also, two AA would be good to start, depending on the other component requirements.

If you did the Bluetooth version, it can talk to their phone, and then to an app that can grab GPS lat/lon and send the data to a cloud app ...

On Thursday, February 23, 2017 at 8:40:24 AM UTC-7, Brad Midgley wrote:

[Scott C. Lemon]

Just for reference ... there are a number of Bluetooth modules out there pretty cheap ($3) that could be used.  You'd still need a MCU (ATMega?) and the gas transducer ...

http://www.martyncurrey.com/bluetooth-modules/

Again ... the advantage here is pair with a phone, and app on the phone can provide a UI, plus grab GPS and push to a cloud app.  We could then have a real-time "heat map" of the monitoring ...

[Scott C. Lemon]

Just as another reference point ... some cool lessons from MAGfest using the ESP ...

https://youtu.be/8WCIu-3OFEQ

[Duane Johnson]

On Sun, Feb 26, 2017 at 9:26 AM, Scott C. Lemon <scott...@gmail.com> wrote:

Just as another reference point ... some cool lessons from MAGfest using the ESP ...

https://youtu.be/8WCIu-3OFEQ

That RSSI mode(?) is really cool. Colorful lights that show how strong the "wifi" signal is!

[Brad Midgley]

Great ideas Scott!

I am worried mobile coding would be a big leap, but maybe there is a way a smart mobile web app could bridge the gap. Sounds like we need to start prototyping!!

Would co2 or volatile gas be generally useful? Honestly I am scrambling to find out what are some of the quantifiable things inversion air has that makes it bad for breathing.

[Beth Sallay]

http://www.purpleair.org there's other sites that have done similar, can this be unique to utah and do a better job or somehow get attention in a way that we aren't already getting. Also it would be useful to look at critiques of devices that monitor air quality and see if they are valid. 

[Duane Johnson]

On Tue, Feb 28, 2017 at 10:12 AM, Beth Sallay <macbe...@gmail.com> wrote:

http://www.purpleair.org there's other sites that have done similar, can this be unique to utah and do a better job or somehow get attention in a way that we aren't already getting. Also it would be useful to look at critiques of devices that monitor air quality and see if they are valid. 

We have a purpleair sensor at our Ardusat HQ in downtown SLC. I don't know a lot about it, but I can ask questions of my coworker who installed it if anyone has any.

[Scott C. Lemon]

We're also working with AQMD, and the purple air project appears to use the $17-$20 particulate matter sensor.

As we're talking here, I think you are beginning to see the complexities of what most people think are very simple issues. The transducer technologies are not exactly what and where you would think they are.

Particulate matter is a very important one to watch. But they are still expensive and very power hungry. CO2 is important, but more important indoors. The VOC transducers detect a broad range of VOCs, and so they do not yet identify *which* VOC is present, and without that knowledge your readings are almost meaningless. (e.g. You know one or more VOCs are present, but not which one, and so you don't know if the reading is "bad" or not.)

A basic mobile app could be created to talk Bluetooth, but the more important part is settling on what we want to measure and provide.

We could provide a solution that people take home and install, but at what cost will they agree to pay for it?

[Beth Sallay]

So is the project starting from scratch or working with something already out there? My boss is currently negotiating support contracts with the environmental health and safety  department right now and the guy I asked for advice never got back to me so I'm trying to figure out who I can beg for info that won't annoy anyone. 

[postmowoods]

Are there any clean air watchdog groups that could subsidize the costs knowing these things will get spread around? 

Just a thought

Kelly

Sent on a Sprint Samsung Galaxy Note® 3

--
You received this message because you are subscribed to the Google Groups "MakeSaltLake" group.
To unsubscribe from this group and stop receiving emails from it, send an email to makesaltlake...@googlegroups.com.
To post to this group, send email to makesa...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/makesaltlake/beec63b7-b31f-4c5f-b9d5-a1d138b8193d%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Scott C. Lemon]

My company and I already have experience in this I'm glad to share ... this is data from one of the commercial unit we've been selling into clinical trials:  https://personal.xively.com/feeds/457408885

This is done with a combination of ATMega and ESP8266 modules, and a sensor array.

The CO2 are Winsen, and the Particulate matter is Shinyei (I think).

If Brad wants to focus the microcontroller meetup on creating something like this, it's pretty easy to get something running.  I do want to make sure that if the data quality is secondary, we want to ensure that people know that.  We've had these in clinical trials for a year now, and have learned a huge amount on the transducers and their accuracy.  (e.g. what you are and are not able to learn from the values)

Also, I have a list of 5-10 additional transducers that we have in R&D and prototype forms ... but again cost and connectivity becomes the issue.

Another spin on this would be to focus on a LoRaWAN solution (instead of Wifi) and then drive the installation of a LoRaWAN network in Utah ( https://www.thethingsnetwork.org/ ) and then create the sensors to be LoRaWAN compatible.

[Beth Sallay]

So from your perspective, what do you recommend? What gets the biggest bang that will get us the most attention/$ as

1. make salt lake 

2. the air quality here to shame the people who could do something about it 

3. empower people who get these to participate in citizen science 

4. empower those that don't to invest ($ or brain power) into stem themselves as a way to help with the air quality and possibly climate science in general

[Brad Midgley]

There are so many cool things to think about and work on. And lots of ground work already done.

Next steps I think

Under $30 target prototype on the particulate sensor path, with an eye on having it outdoors.

Start with big-battery path initially with simple weatherproofing

Get advice from Scott so we avoid early problems

Use maker faire and arts festival and other events to promote participation

Get people to register interest -- will need a way to manage community and device registration

Secondary topics

Data collection will need to be solved, with wifi, bluetooth, LoRaWAN as possibilities

xively and possibly a custom app for citizen participation

GATT service for air quality, work from http://mynewt.apache.org/latest/os/tutorials/air_quality_ble/

LoRaWAN network in Utah ( https://www.thethingsnetwork.org/ )

Outdoor power: solar, POE, battery recharging

On Wednesday, March 1, 2017 at 12:03:24 PM UTC-7, Scott C. Lemon wrote:

[Scott C. Lemon]

I do think that particulate matter is a pretty easy one to do, and get decent readings.

Under $30 target is a good target ... who do we anticipate is going to pay for these?

I'm not sure that I agree with battery.  This type of device is going to wipe out the batteries in a few days, and people are not going to enjoy constantly changing batteries.  Even our project at the Mayo clinic grew tired of changing batteries every three months.  I think that if we really want "customer satisfaction" and a project that is going to keep running long term, we really ought to be looking at AC power.  We can get "wall wart" power supplies fairly cheap, with longer cords, that will ensure that the sensors can be installed and "forgotten".

If this first version is WiFi, then we promote people to install the units outdoors at their house, plug it in, configure to get on their home Wifi, and it talks to our cloud application.  They can then create an account and put in their address or general Lat/Lon and we can do all sorts of things ... long term.

I'm really afraid that a lot of expense could go into this and then the project simply is discarded when people give up changing batteries.  If people can "install and forget" then it's going long term, and we can email them reports and give them a portal to access and view the data.

I would be willing to have Wovyn sponsor some of the development of the firmware and cloud software since we have been doing this.  I wonder if the festival would assist in having people "pre-order" their sensor from Make:SLC for pick up at the festival?  We could have them come to the booth to learn about it, how to install and set it up?  Then they just take it home and install it.

[Brad Midgley]

Scott

I'm not sure that I agree with battery. 

Good point on battery. I like the feedback. It would be great to have help on the firmware, thanks!

Should I buy a sensor for the prototype?

New list:

Under $30 target prototype on the particulate sensor path, with an eye on having it outdoors.

AC power target (could use passive POE splitter or something if distance is a concern)

Simple weatherproofing needs to be addressed

Each host person will pay for the unit

Assembly party events

wifi configured through a cell and using geolocation API

The things needing discussion:

Use maker faire and arts festival and other events to promote participation. Decide exactly how this looks. Maybe assembly party at the events? It wouldn't be stuff people take home but they could register an interest and help with the builds. 

Tracking interest -- will need a way to manage community and device registration. This looks to me like a custom application that integrates with data collection at xively. I can start this part if there isn't anything already out there.

Brad

[Scott C. Lemon]

Just another data point ... great article today on air quality and pollutants.

Ozone is a critical measure, but the accurate transducers are still really expensive!

http://www.ksl.com/?sid=43381911&nid=157&title=smog-in-the-american-west-study-shows-asia-largely-to-blame

[Brad Midgley]

I tried some experiments on the esp12e-packaged nodemcu 1.0 board I have handy.

First, geolocation using a server on the esp is out. Modern browsers require https. We can do that from our own real server but it would be later in the registration process.

I was able to use this sample to connect to the esp in AP mode and tell it the wifi credentials: https://gist.github.com/dogrocker/f998dde4dbac923c47c1

My thought now is that we would have AP mode and set the wifi credentials and some kind of unique id in the initial stage, then the user would connect to our server from phone/computer and share the same token so we can link everything up.

[Beth Sallay]

Utah Arts Festival people are going to hook me up with ideas for funding. I'll be emailing Brad with info and posting more here.

[Beth Sallay]

At the meeting I went over the usual list of projects and then the possible (newer) ones ending with the wireless air quality sensor because I wanted to spend more time on it. A lot of people had ideas for funding and two approached me after saying to just give them a number. I don't know how much of a sure thing it was but I said that one member had one for a community garden at $6-$7 a piece and we're working out the modules for it now and it's in the design stages so we don't have a final design and $ yet. But at the very least I'd like a way to lower attrition when handing these out by hooking up with science teachers or classes so they don't get tossed. I'd like some way for the public to participate in citizen science by using these and seeing the effects the air quality is having with the data. I have been talking about this everywhere and keep getting ideas for funding. I told the group we'd do a gofundme if needed. So there's a lot of support. And in the end I got an email address and was told to give them a dollar amount. I'm hopeful. 

[Scott C. Lemon]

Are you looking for a "per unit" dollar cost?

[Scott C. Lemon]

Brad,

A few thoughts:

1. Geolocation should be done when the user registers their device on-line - in the portal - and they provide an address of where it will be installed.  We can then use the Google APIs to look up the Lat/Lon by address.  I think it's critical to go through this to accurately locate where they are putting the sensor, and also to have good contact info for them.  (e.g. future contact, status reports, issues/contact)

2. I've got an entire library that I developed that is based on the WiFi Manager code here:  https://github.com/tzapu/WiFiManager  In my library what I do is give the user a chance to press a button during boot to enter into WiFI config mode, and we can grab WiFi *and other* parameters and then I save them to the SPIFFS flash memory.  I've also integrated the OTA firmware updates so that we can update them in the field after installation.

3. The ESP modules already have a GUID (unique ID) so there is nothing else we have to do there.  We usually display it in the WiFi Manager set-up, and also on a sticker on the device ... they use that to "claim" the sensor during the registration process.

Is there some time that you want to set-up a meeting to start to create some real project management documentation for this?

[Brad Midgley]

Scott

 

Is there some time that you want to set-up a meeting to start to create some real project management documentation for this?

Yes, I can work it out Wednesday nights around the meetup time ~6:30 or I can make a lunchtime work. 

I need to talk to Beth too. I'll start checking in Wednesdays.

Brad 

[Brad Midgley]

Some thoughts tonight

If we use nodemcu or a similar platform with a pushbutton gpio0 programming button, we could reuse this button. It could be used at runtime to indicate the node should go back into access-point-configuration mode. Is that what you do Scott? I'd love to avoid adding buttons or making a custom board.

I'd like to put a firmware in an open source repo that we could work together on. I could put what I've done up and maybe it's ok if we'll trash it and start over.

I contacted the takingspace/airbeam/aircast.org people to see if there is overlap with their work. The software at least should be similar and they might want to share data. They have a $250 unit (!!!) for sale on http://www.takingspace.org/aircasting/airbeam/ . Probably to help fund their infrastructure and efforts on the whole platform. But they have published a 3d printed case and we might be able to use that and their pm2.5 sensor of choice to get those parts of the problem.

Brad

[Brad Midgley]

Project requirements

* construct the prototype below

* mod the firmware so it can use captive AP to program it on first boot

* mod the firmware so flash button gpio0 starts AP programming mode (but don't erase eeprom)

* acquire a domain name

* web site and basic web pages to advertise the project (main page should also work for promotions and posted near sensor)

* select message queue and data collection server side (budget)

* devise an enclosure

v1 prototype

=========

parts: Plantower PMS3003 ($17), nodemcu v1 ($3.20), microusb ac adapter ($7).

https://rishav006.wordpress.com/2016/06/22/pm2-5-a-sensorweb-project-by-mozilla/

Reference firmware: https://github.com/sensor-web/arduino-station/

https://www.aliexpress.com/item/Laser-PM2-5-DUST-SENSOR-PMS3003-High-precision-laser-dust-concentration-sensor-digital-dust-particles-G3/32371229255.html

https://www.aliexpress.com/item/1PCS-Wireless-module-CH340-NodeMcu-V3-Lua-WIFI-Internet-of-Things-development-board-based-ESP8266/32665100123.html

https://www.amazon.com/Adapter-YIFENG-Switching-Charger-100V-240V/dp/B01M4LRSC5/ref=sr_1_2

[Beth Sallay]

Do you think you could have a prototype or 5 by mini maker Faire?

[Beth Sallay]

To give out.

[Brad Midgley]

It might be a race. The sensor comes from aliexpress and I only ordered one since I don't know how well it will work. I can't track it at the moment.

[Will Blaylock]

Brad will likely win, I have been working on some projects to help a
couple of kids who have been hurting themselves... I have not even
started looking at finding the money to build my prototype.

planned on probably April unless I get lucky money wise LOL

When is the faire? I do not see it right off?

[Scott C. Lemon]

Hello ... sorry, was out of the country and got back late Monday.

Yes, I'm re-using the "program" button for WiFi config.  It's the best way to do things, and minimize parts.  Upon power-up I give the user 5 seconds (can be changed) to press the button and enter config AP mode.  Agreed on doing any type of custom board ... it's not worth it.

We ought to use the MakeSaltLake repos on Github for all of the open source.  Like you, I'm waiting for the arrival of the PM sensors from China to put together my initial prototype.  I've already started to create the base project, ready for pretty much any sensor to be created.  I'll see if I can post my code by this weekend.

I'd question 3d printing enclosures ... only because of cost and time.  There are a lot of inexpensive off-the-shelf enclosures from china that we could get, and then simply laser-cut the internal mounting panels from acrylic ... but if someone else wants to make that call they are welcome to.

I'm also curious if XMission would be open to donating some cloud server(s) and storage for the data, since we're making this an open and public solution?

[Scott C. Lemon]

A few thoughts here ...

1. * acquire a domain name
   

1. Should we create a new domain name?  Or put this under MakeSaltLake?
1. Power
1. we'll have to look at the enclosure - particularly for outdoor - and see if we really want to expose the USB
2. I've been using cheap 5.5mm x 2.1mm power supplies and connectors
3. These also work well for off-the-shelf PoE splitters
4. I got these power supplies for <$3 in 100 quantity
1. https://www.aliexpress.com/item/The-best-5-V2A-power-adapter-5V2A-LED-lights-for-power-5-V2A-AC100-240-V/32637139365.html?spm=2114.13010608.0.0.U6xcxT
   

[Scott C. Lemon]

WHen is the Mini Maker Faire?  Dates?

On Sunday, March 19, 2017 at 11:08:03 AM UTC-6, Beth Sallay wrote:

[Beth Sallay]

You should document the everloving crap out of this and put it online
yes, everything, articles before you had the idea, that drone picture
with the brown air, the current attempts to get rid of protections,
etc. How it affects you personally, sore throat, it's making me have
more migraines, we are beating out the worst 3rd world polluters some
days, your kids are getting sick. Then, everything technical, all
iterations of ideas. You had X idea but realized it wouldn't work, so
that others can take it and not make that mistake.You want a better
design, better air, a citizen science, not something cheap you can
sell and walk away from. You are telling a story to get people
interested in giving you money (to give these away) and time so that
they stay invested and keep the momentum going. You want them to share
this with other people like they do that giant chicken video. What the
hell was that thing? Tell personal stories, that's the best draw.


Checking with the board on adding it to msl.

Mini Maker Faire is April 15th

> --
> You received this message because you are subscribed to a topic in the
> Google Groups "MakeSaltLake" group.
> To unsubscribe from this topic, visit
> https://groups.google.com/d/topic/makesaltlake/Cs9zvKMywXs/unsubscribe.
> To unsubscribe from this group and all its topics, send an email to

> makesaltlake...@googlegroups.com.
> To post to this group, send email to makesa...@googlegroups.com.
> To view this discussion on the web visit

> https://groups.google.com/d/msgid/makesaltlake/8350e112-bc85-4d7b-b762-252f351bdd21%40googlegroups.com.

[Brad Midgley]

Notes on our discussion 3/22/17

• Need to find out who at makesaltlake can administer domain to add a subdomain and register the custom name for the space to own
  
• Choose a good domain name, something conveying mapping, sensing, visualizing, environment, people, space
  
• Will avoid tying the name to air or Utah to keep it generic
  
• Need to define the application behavior and a platform to work with a programming team
  
• research enclosures
• github project space (same as domain name?)
• github collaboration

Getting an environment going:

• serial driver download (search for nodemcu setup)
• download arduino ide ( www.arduino.org/downloads )
• launch arduino, preferences, add http://arduino.esp8266.com/stable/package_esp8266com_index.json to additional board manager URLS
• tools->board->board manager
• scroll down to esp8266, click it and click install
• get sample code
• create a project with the right board descriptor
• tools->board->nodemcu 1.0

[malcolm.s...@gmail.com]

Nice to meet everybody last night, at the microcontroller meetup.  Thank you for helping me start learning about this technology.  I'm looking forward to continuing on the project!

Malcolm

[Brad Midgley]

Great meetup folks!!

Rather than making another todo list, I think it might be good to use the existing makesaltlake space in trello.

https://trello.com/makesaltlake/members

Please make an account there and ask Joel to add you do the tream. It can use google to sign in if you want to avoid making yet another password.

Brad

[Brad Midgley]

I submitted this to openwest and started writing slides. I'll need to replace the pictures or ask for rights to use them.

https://docs.google.com/presentation/d/1r7VYOgbqw14KDcmXZjbTtutMpEnE_ZIQqukuaHpiU2Q/edit?usp=sharing

[Brad Midgley]

I get to speak at openwest. Should be fun.

Scott, the workflow for wifimanager doesn't make sense. Their docs:

Instead of calling autoConnect() which does all the connecting and failover configuration portal setup for you, you need to use startConfigPortal(). Do not use BOTH.

I found this confusing, because what it looks like I want is to to call autoConnect and then in the loop, call startConfigPortal if GPIO0 goes low. Is that allowed? Or could I instead watch for GPIO0 to go low and clear the eeprom and use autoConnect? The problem is you may want to reconfigure in the presence of the previously configured AP.

[Malcolm Schongalla]

Brad-

Interesting. Do they waive your registration fee for Openwest then? :)

Minor point, but you could add to slide #6, cardiovascular and pulmonary health complications.  Those correlations are much more strongly established in the research on particulate exposure.  The link to neurological problems is more speculative (so far, I believe).  If you want some deeper background, for context in your presentation, I could do some digging to produce some talking points.

(Dr.) Malcolm

--
You received this message because you are subscribed to a topic in the Google Groups "MakeSaltLake" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/makesaltlake/Cs9zvKMywXs/unsubscribe.

To unsubscribe from this group and all its topics, send an email to makesaltlake+unsubscribe@googlegroups.com.

To post to this group, send email to makesa...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/makesaltlake/c828a8f3-3b4d-4069-a5c7-13ee6d649ab8%40googlegroups.com.

[Beth Sallay]

So if you don't get far enough along by the UAF would you be willing
to put together some info and talk about it at the tent? We can work
on documenting personal stories, ideas, where you are so far, etc and
then the concept and how far you have to go. I think this is a good
platform and audience for it.

> --
> You received this message because you are subscribed to a topic in the
> Google Groups "MakeSaltLake" group.
> To unsubscribe from this topic, visit
> https://groups.google.com/d/topic/makesaltlake/Cs9zvKMywXs/unsubscribe.
> To unsubscribe from this group and all its topics, send an email to

> makesaltlake...@googlegroups.com.

[Brad Midgley]

Yes--it will get me into the conference :)

I updated my slides. Thanks for the tips.

On Sunday, April 23, 2017 at 11:33:11 PM UTC-6, Malcolm Schongalla wrote:

Brad-

Interesting. Do they waive your registration fee for Openwest then? :)

Minor point, but you could add to slide #6, cardiovascular and pulmonary health complications.  Those correlations are much more strongly established in the research on particulate exposure.  The link to neurological problems is more speculative (so far, I believe).  If you want some deeper background, for context in your presentation, I could do some digging to produce some talking points.

(Dr.) Malcolm

On Sun, Apr 23, 2017 at 11:57 PM, Brad Midgley <bmid...@gmail.com> wrote:

I get to speak at openwest. Should be fun.

Scott, the workflow for wifimanager doesn't make sense. Their docs:

Instead of calling autoConnect() which does all the connecting and failover configuration portal setup for you, you need to use startConfigPortal(). Do not use BOTH.

I found this confusing, because what it looks like I want is to to call autoConnect and then in the loop, call startConfigPortal if GPIO0 goes low. Is that allowed? Or could I instead watch for GPIO0 to go low and clear the eeprom and use autoConnect? The problem is you may want to reconfigure in the presence of the previously configured AP.

On Sunday, April 2, 2017 at 7:41:23 AM UTC-6, Brad Midgley wrote:

I submitted this to openwest and started writing slides. I'll need to replace the pictures or ask for rights to use them.

https://docs.google.com/presentation/d/1r7VYOgbqw14KDcmXZjbTtutMpEnE_ZIQqukuaHpiU2Q/edit?usp=sharing

--
You received this message because you are subscribed to a topic in the Google Groups "MakeSaltLake" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/makesaltlake/Cs9zvKMywXs/unsubscribe.

To unsubscribe from this group and all its topics, send an email to makesaltlake...@googlegroups.com.

To post to this group, send email to makesa...@googlegroups.com.

[Brad Midgley]

Yes, great idea Beth. Let's run with whatever we can. I'm coding the firmware without the sensor at the moment for example :)

[Curtiss Porter]

Hey Brad just saw this on the net 

  https://opensource.com/article/17/3/build-smog-sensor-ESP8266-microcontroller?sc_cid=70160000001273HAAQ      

On Thursday, February 23, 2017 at 8:40:24 AM UTC-7, Brad Midgley wrote:

Some thoughts on the air quality project.

Let's make this the flagship product of the microcontroller meetup. It touches on every kind of problem we solve.

Let's select a platform. I'm interested in low cost wifi enabled so I think esp8266 is a clear winner. We need a sensor.

I'm willing to have data quality be secondary in initial phase. We can have people set these up in lots of different circumstances and not worry as much about rigorous deployment initially.

I'm contacting the University of Utah to see who we can partner with.

I imagine a sign saying a data collection point is nearby would be great publicity and might even encourage better behavior. Of course it will have a web page and QR code.

Unless there is a better model out there for connecting, let's have the device try to acquire a connection and if that doesn't work, go into its own access point mode. If users have a cellphone they connect to it, it can use geolocation API to ask the phone where they are and remember it along with access point password. It could register an email as well.

Let's talk power. I think we need to start with three AA batteries and go from there.

I'll buy a couple of these since I'm out of boards for new stuff. http://www.ebay.com/itm/like/201607584855

Brad

[Brian Goossen]

First off, congratulations to Brad for being chosen to speak at OpenWest!!! 

Brad, if you point me in the direction of the pieces I will get them ordered and will be happy to do some testing.  Also, any thoughts yet on an enclosure? This is very cool thanks for driving this project forward.

[Brad Midgley]

Brian,

Here is the slide with links to buy things.

I LOVE the enclosure they are using in Curtiss' post. We can do this too--just connect a couple of large PVC elbows and aim them downward!

The git repo is up and currently reports analog input values to an mqtt server. I'll put the slides and everything on github.

https://github.com/bmidgley/citizen-scientist

Brad

• Plantower PMS3003 ($16)

• nodemcu v1 ($3)

• 5v AC adapter ($7)

• Enclosure ($?)

[Brian Goossen]

Ordered. 

Shipping time is 60 days... so it might be a while. 

I'll spin through the plumbing isle next time I'm at the hardware store.  

I also really liked their map. 

[Brad Midgley]

Sensor arrived yesterday at long last.

The IoT sensor is fully armed and operational!!

monitoring the message bus, 2-second intervals before and after I lit a match near the sensor:

{"pm1":1,"pm2_5":3,"pm10":5}

{"pm1":1,"pm2_5":3,"pm10":5}

{"pm1":1,"pm2_5":4,"pm10":6}

{"pm1":1,"pm2_5":3,"pm10":6}

{"pm1":1,"pm2_5":3,"pm10":5}

{"pm1":1,"pm2_5":3,"pm10":5}

{"pm1":1,"pm2_5":3,"pm10":5}

{"pm1":107,"pm2_5":414,"pm10":788}

{"pm1":133,"pm2_5":461,"pm10":851}

{"pm1":153,"pm2_5":491,"pm10":888}

{"pm1":156,"pm2_5":496,"pm10":896}

{"pm1":90,"pm2_5":177,"pm10":252}

{"pm1":53,"pm2_5":87,"pm10":114}

{"pm1":39,"pm2_5":64,"pm10":83}

{"pm1":28,"pm2_5":42,"pm10":53}

{"pm1":28,"pm2_5":42,"pm10":53}

{"pm1":17,"pm2_5":24,"pm10":32}

{"pm1":9,"pm2_5":14,"pm10":18}

{"pm1":6,"pm2_5":10,"pm10":13}

{"pm1":5,"pm2_5":8,"pm10":11}

{"pm1":5,"pm2_5":8,"pm10":11}

{"pm1":4,"pm2_5":8,"pm10":9}

code is pushed at https://github.com/bmidgley/citizen-scientist

I did have to adjust the code sample we had for this sensor. Their code was trying too quickly to read the serial port after draining it.

An interesting problem is the sensor when connected constantly sends uart signals that interfere if you want to flash the esp8266. I added a switch to manually ground the sensor reset while programming it, but there has to be something better for that. Another reason to get OTA updates working as then you don't use the uart to reflash.

Brad

[Brad Midgley]

So I could use hardware help at this point. I haven't found a better esp8266 packaging and I'll need help if we have to make a board.

Where should we be investigating enclosures? I could mock up the PVC elbows but that is surprisingly expensive. They can charge it since usually PVC deployments are not all about the elbows :)

I need a clear enclosure for the arts festival. It can be a one-off, but it would be best if it could impress.

[Brian Goossen]

Hey Guys, I got my esp8266 board and sensor in the mail yesterday! I downloaded and tried to compile Brad's code from git and it breaks saying I need "FS.h". I found instructions here: https://github.com/esp8266/Arduino 

Unfortunately, I can't get the board to show up as an option in the boards manager with these instructions. Any suggestions? 

Also, brad do you have the wiring layout between the board and the sensor? 

Looking forward to getting this up and running. 

Thanks!

Brian

[Brian Goossen]

Ok here is the "sorta" fix I found:

https://ndsybkiot.wordpress.com/2017/05/05/how-to-fix-error-downloading-httparduino/

https://github.com/esp8266/Arduino/releases/download/2.3.0/package_esp8266com_index.json

This kind of worked as I was able to install the board but now I get "WiFiManager.h" is missing... any suggestions appreciated! 

[Brad Midgley]

Brian,

In the arduino IDE, try to load the additional libraries with this:
	// sketch->include library->manage libraries
	// WiFiManager, ArduinoJson, PubSubClient

[Brad Midgley]

Also we will work on it Wednesday! I have fixed some bugs in the client and it can also report back a GPS location if you're on a moving hotspot (must install an NMEA relay).

We can do a code review on Wednesday and work on the next feature together.

[Brian Goossen]

Hi Guys, Sorry I missed you last night! 

I'm banging my head against the wall trying to get the going. 

1) I didn't have all the libraries but that was an easy fix

2) I downloaded the driver from the wemos.cc website... It crashed my Mac! I had to restore from a back up...

3) Then I tried this driver still without luck... at least it didn't crash my system:  http://kig.re/2014/12/31/how-to-use-arduino-nano-mini-pro-with-CH340G-on-mac-osx-yosemite.html

I'm using the Arduino IDE and choosing board "Generic ESP8266 Module" and getting errors:

warning: espcomm_sync failed

error: espcomm_open failed

error: espcomm_upload_mem failedu

4) Then... hold the "FLASH" button on the board, press and release the "RST" button, then after a few more seconds release the "FLASH" button. SUCESS! Upload complete. 

5) "GeothunkAP"  popped up right away. After, joining wizard opened immediately. I input my home WIFI credentials and eventually GeothunkAP disappeared.

6) I'm not seeing any activity in the serial monitor or serial plotter... next steps? 

I hope my notes can save others some time and frustration. 

Also, I went looking for documentation for the sensor to wire it but came up empty. I know it has to be out there. Any suggestions or wiring diagrams? 

Thanks! 

Brian

[Brian Goossen]

Here are some specs I found on our sensor including the pin-out:

http://aqicn.org/sensor/pms3003/

[Malcolm Schongalla]

Hi everyone, 

I have been out of the loop a lot lately, thanks to a very busy work and travel schedule. (But I am still following along...)

Brian's article brings up a very good point by saying,  "There is quite often a big difference between what is said to be possible and what is actually possible (and this is why we are setting up all our experiments)."

And with regards to sensor accuracy, in this vein, has there been any discussion yet of device calibration and validation? If we want our data to be meaningful and comparable from one installation to another, or between our data sets and other communities, this step is not "optional."

Malcolm

--
You received this message because you are subscribed to a topic in the Google Groups "MakeSaltLake" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/makesaltlake/Cs9zvKMywXs/unsubscribe.

To unsubscribe from this group and all its topics, send an email to makesaltlake+unsubscribe@googlegroups.com.

To post to this group, send email to makesa...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/makesaltlake/501be73e-5840-4901-8a4e-6779a892b66d%40googlegroups.com.

[Brian Goossen]

Brad, I pulled the new code and loaded it however the GeothunkAP wifi will not show up. I waited a good long while, power cycled and even re-flashed without any luck.  It could be something I'm doing wrong on my end but I don't know what. Suggestions? 

Malcom, what would the process of calibration look like? 

Thanks! 

Brian

To unsubscribe from this group and all its topics, send an email to makesaltlake...@googlegroups.com.

To post to this group, send email to makesa...@googlegroups.com.

[malcolm.s...@gmail.com]

On Friday, May 26, 2017 at 3:07:37 PM UTC-4, Brian Goossen wrote:

Malcom, what would the process of calibration look like? 

Well, I was hoping that someone else had a magic solution already :)  ...I suspect other groups have ignored this step for a variety of reasons.  Validation is important if we wanted to collect scientifically-publishable data.  And it strengthens our argument, if we wanted to submit our data to influence public policy.  If we just want to build a cool gizmo network, then never mind.

The goals are to get a set of calibration curves for individual sensors (or prove the unnecessity), and to understand the error ranges/limitations of the sensor.

1. The default option is to say "heck with it," (at least for now) and live with the uncertainty.

2. Later on, we could build a calibration apparatus in the maker space.  Ideally, it would get certified on-site by a reference sensor.  Then, we could individually bring in sensors to compare them against our standardized setup.  I'd bet we could get a grant to help support this.

3. An alternative to #2, is to have individual sensors brought to a company that does the calibration in their own facility (this would totally wreck the cost curve, though).

4. I don't see an "at-home" validation process working out, though that would be nice.

#2 would take an investment in equipment, reagents, and workspace.  Here are 2 informative documents:

ftp://185.72.26.245/Astm/2/01/Section%2014/ASTM1402/PDF/F328.pdf

ftp://185.72.26.245/Astm/2/01/Section%2014/ASTM1402/PDF/F649.pdf

At first they may seem overwhelming. But in practice, these usually get reduced down to a series of basic steps, which are easy to follow once someone has been trained on them -- getting to that point is the challenge.

It would be worth seeing if any of our members have experience in quantitative lab analysis, and to talk to other DIY communities that have gone before us.  I'm sure this will go on the back burner, but I was mostly curious if there'd been any discussion of it already.

Malcolm

[malcolm.s...@gmail.com]

Also-

The PurpleAir.org website briefly mentions they have a "smoke chamber" for validating their sensors before deployment.  They use a similar sensor to ours.  I emailed them with a few questions about it, I'll post back here if I get a response.

[malcolm.s...@gmail.com]

I got a reply from a Mr. Adrian Dybwad of the www.PurpleAir.com project.  One key takeaway is that they have a method, of sorts, that we could duplicate the makerspace, with very minimal resources.  Another take-home point is that he's willing to demonstrate his setup to the group.  The scientist in me still proposes a comparison to a known benchmark (either a known-calibrated device or a standard aerosol), and his technique doesn't really do that, but it's a pragmatic compromise.

If anyone wants to take him up on his offer to demonstrate, let me know and I'll try to coordinate. (And I'll be back from work to attend another meeting at some point, I promise!)

Q) Does your group have its own calibration chamber (and did you design it yourself), or do you outsource to a commercial calibration facility?

A) "We use a plastic box as a 'test chamber'!"

Q) Do you publish your validation protocol?

A) "All we do is compare 12 sensors at a time to make sure they all read the same."

Q) Where do you source your standard aerosols?

A) "Smoldering a piece of twisted up paper!!"

Q) Your website says that "So far, all sensors we tested have produced consistent output"--  Would you be willing to to share examples of your error ranges other testing results?

A) "I do not have any results to share since we do not make a separate not of them other than the data and real time display that each sensor has during testing."

Q) If you do calibration in-house, would you by any chance be able to host an educational "show-and-tell" for our small group to come see it?

A) "I would be pleased to come and demonstrate how this all works"

Q) Do you ever need volunteers, or are there other ways to get involved in PurpleAir, other than purchasing a sensor (The sensor, unfortunately, being beyond the budget of many of our makerspace participants)?

A) "We may be looking for help soon in making sensors. We are experiencing some growth."

[Brian Goossen]

Malcom, thanks for the research!  I'ts is a big decision and the start to a good discussion. 

Brad, have you had a chance to see the Purple Air housing? What do you think? A PVC end cap might work as well. Also, have you had any thoughts about added protection for the microcontroller like heat shrink or other protection from humidly, dust, bugs, etc.?

Thanks!

Brian

[Brad Midgley]

Brian, I do like the purple housing. It looks custom/expensive. Some of the other people doing this stuff seem to be trying to make money on it.

I have my ten sensors and boards in hand!

I'm going to try to get a clear laser-cut box whose sides lock together and we could seal along the top to keep water out. Have you guys done that? I'll ping Duane.

I've also gotten a $9 esp8266 board that includes a display. That would be a quick and simple way for the node to display what it's up to, but wouldn't be essential. I would just prefer to go with cases that could work with the board that has a display if we do want to do a few like that. 

The code still needs real uuid generation and over-the-air updates. I'm trying to get those done this week. I'm thinking updates would be done by sending out a message on a topic the clients listen to that says to update their firmware if they have a certain range of UUID or some other qualification. That would allow us to update a few devices at a time.

[Beth Sallay]

How do you think it went with the writer last night?

[Brad Midgley]

I think it went well. It also became apparent how useful it is to have the display on the device so we have more to talk about. I've gotten some basic stats displaying now.

For the record, I'm using the smaller d-duino board for showing what is happening. It adds $7 to the cost.

https://www.aliexpress.com/item/new-NODEMCU-wifi-NodeMCU-forArduino-ESP8266-wemos-for-OLED/32802182457.html?spm=2114.10010108.1000014.8.GwNKS4&traffic_analysisId=recommend_3035_null_null_null&scm=1007.13338.84300.000000000000000&pvid=69efae30-d8ae-40b7-b62b-ade8ad49205a&tpp=1

[Beth Sallay]

It did: https://www.theutahreview.com/backstage-utah-arts-festival-2017-art-technology-venue-bustling-shop-make-take-projects/ 

[Brad Midgley]

FTR... on Wednesday, we wired up a couple of the "v3" nodemcu boards but it appears there is a power problem. I will spend a little more time on it but since we didn't really expect to use the v3 board in the long run, I will work it out on the d-duino boards with the little screens when that order shows up.

[Brian Goossen]

Driver I found for the D-duino with screen: 

https://www.silabs.com/products/development-tools/software/usb-to-uart-bridge-vcp-drivers

If you find a better one please let me know.