Hong Kong project

[PenguinSix]

Greetings from Honk Kong,

Just saw your Kickstarter and thought I'd say hello. About a year ago
some members of the Hong Kong Hackerspace started some initial work on
a low-cost air quality sensor. I believe they were working off an
Arduino board solution but due to the rather transient nature of Hong
Kong living some of the folks involved were relocated away and our
project sort of fizzled. I'll see if I can scrounge up some of the
work they did if it might be of any use to this project.

Here in Hong Kong and in much of China, the biggest problem we face is
PM. PM2.5 and PM10 readings are consistently well over the W.H.O.
maximums for daily limits. I just released an iPhone app that takes
pollutant levels here and then runs it against various indexes from
other countries. It's helpful to the expat here who might be more
familiar with a different standard for air pollution (i.e. an American
hearing 'code orange' knows more what that means rather than hearing
'the Hong Kong API is 100').

http://itunes.apple.com/us/app/hong-kong-air-pollution/id504536152?ls=1&mt=8

We'd love to find a low cost solution that we could offer to schools
and other institutions around the city so that we could build a much
better map of the current pollution situation here. The government
only has 14 monitoring stations focused in the business districts,
while several more residential areas live in a bit of a mystery as to
the overall levels.

[Dirk Swart]

Hi.

We would love to hear what you have. Especially if you have PM 2.5 related solutions. Doing a cost effective measurement of that is really hard.

Cheers
dirk



Follow me on Twitter now: dswart

[joe saavedra]

Absolutely.

PM2.5 is perhaps the most important air pollutant to measure, according to the World Health Organization.

It is notoriously difficult to measure -- I myself worked on a funded project which required PM2.5 measurements -- we ended up *renting* a $10k sensor for $300 a week.

Recently there have been development using refracted light to get low resolution readings on particulate matter of small air samples.

We'll keep updating with ideas, but if you have resources, please pass along!

joe

[Gustavo Olivares]

I've been experimenting with the SHARP  GP2Y1010AU0F  dust sensor for an indoor prototype (codename PACMAN: Particles, Activity and Context Measurement Autonomous Node ... might change it if Namco starts noticing ;-) )

After some headaches with actually getting the damn thing to work (I'm a chemical engineer with some very old and basic electricity/electronics knowledge ... trying to augment it with MITx course!) We run it in parallel with a DUSTRAK and a SIDEPAK from TSI.

Both instruments are "industry standard" for personal exposure and have active sampling systems (i.e., a pump) with a small cyclone in the inlet so they "claim" that they actually measure PM10 or PM2.5, or PM1. We had them measuring PM10.

Now, the really interesting thing is that for indoor sources (cooking oil, incense, cigarette smoke, gas cooking) the little PACMAN was giving very comparable results!, in one of the tests we got a 98% of agreement between the $10 Sharp unit and the $5k instruments.

We have a manuscript in the next Healthy Buildings conference in Brisbane later this year for what was version 2 of the PACMAN. I'm now starting to build version 5 that incorporates a few enhancements.

I haven't been too good with documentation outside my desktop but I could send the manuscript to anyone interested and the schematics and arduino code are in:

https://bitbucket.org/guolivar/pacman 

I'm in New Zealand so I don't think I'll ever make it physically to a workshop and I'm about 15hrs ahead of where the action is around the Egg but I'm more than happy to share my experience and learn from yours.

/El Gus

[Ed Borden]

This is great... I really think we are going to offer an Indoor air
quality sensor box that can connect to the Egg. It would be great to
incorporate something like this. Maybe you can give some more details
on size/capabilities/etc? We could use this for sure.

[nopolabs]

I am very interested in the Air Quality Egg and have shared information about it with a local tech group (dorkbotpdx).

One thing that responders seem interested in is measuring particulate matter (PM2.5 and PM10).

Here is an example of a component that might fit this requirement: http://www.seeedstudio.com/wiki/Grove_-_Dust_sensor

I also sawOptical Dust Sensor - GP2Y1010AU0F on one of your wiki pages. Has that been tested?

Also, to what extent will the hardware and software developed by this project be open source and/or available for extension and modification?

Best Wishes,

Dan

[joe saavedra]

All files will be open sourced!! 100%!  We support Open Hardware as well, I was a part of the open hardware summit the past two years.

In my experience, the two "dust" sensors are good at giving you an idea of measuring how clear the air is of particulate, but neither are good at actually giving a reading of true PM2.5 or PM10.  In general, PM sensors are extremely expensive, on level of $5k or more for accurate readings. Although, what El Gus is saying sounds super promising -- if we added a sensor or two, would anyone be interested in upping their pledge?

There are people working on developing low cost solutions, we'll post here if we start to consider any.

The shield will leave lots of space for hacking and modding, and like i said the source files will be available as well.

Thanks for checking out the project!!

joe

[Nafis]

Paul Deng has already put together/documented an Arduino /Sharp Dust
Sensor that posts data to Pachube (http://sensorapp.net/?p=479)

On Apr 2, 11:34 pm, nopolabs <danre...@gmail.com> wrote:
> I am very interested in the Air Quality Egg and have shared information
> about it with a local tech group (dorkbotpdx).
>
> One thing that responders seem interested in is measuring particulate
> matter (PM2.5 and PM10).
>
> Here is an example of a component that might fit this requirement:http://www.seeedstudio.com/wiki/Grove_-_Dust_sensor
>

> I also sawOptical Dust Sensor - GP2Y1010AU0F<http://www.sparkfun.com/products/9689>on one of your wiki pages. Has that been tested?

>
> Also, to what extent will the hardware and software developed by this
> project be open source and/or available for extension and modification?
>
> Best Wishes,
>
> Dan
>
>
>
>
>
>
>
> On Saturday, March 31, 2012 5:54:02 AM UTC-7, Ed Borden wrote:
>
> > This is great... I really think we are going to offer an Indoor air
> > quality sensor box that can connect to the Egg.  It would be great to
> > incorporate something like this.  Maybe you can give some more details
> > on size/capabilities/etc?  We could use this for sure.
>

> > On Sat, Mar 31, 2012 at 10:42 AM, Gustavo Olivares <guoli...@gmail.com>

> >http://itunes.apple.com/us/app/hong-kong-air-pollution/id504536152?ls...

[Gustavo Olivares]

Good for the openness!

Now, back to the PM issue.
Actually, measure PM is relatively simple and cheap (like those dust sensor examples have shown), the problem comes with size. If you want a PMx then you need to ensure that the particles you're measuring comply with that requirement and that's why the expensive instruments have pumps ... the rest of the cost goes on sensitivity of the electronics and measuring principle (e.g., the sharp sensors do not respond to concentrations below 100micro-g/m3 but the Grimm Aerosol spectrometer that costs 1000 times more but it is basically the same thing can measure down to 1micro-g/m3)

PM10, PM2.5, PM1 are not different things but 3 different metrics of the same aerosol population. The reasons for using one over the others are largely historic as there is loads more info on PM10 than there is on PM2.5 and on PM1 but these are all indicators and not in themselves the goal of any measurement.

What my experiments show is that for combustion sources PM~PM10~PM2.5~PM1 because most of the mass of the particle distribution is below 1micron and therefore I can forget about segregating by size and use the cheap sensor. If you don't have information about your sources, then you can't make that assumption and therefore have to measure with something else. Now, in urban areas, except in desert regions, aerosols are mostly from combustion and therefore of sizes below 2.5microns.

More importantly, the choice of measuring particles is rooted on the fact that they get into our lungs and cause very specific responses but they are just one indicator and we don't know what about the particles is what causes the most harm. It could be their size (below 100nm they get into the blood stream), or their shape (pointy ones cause more acute responses) or their coating (what's on the surface of the particles is what interacts with the tissues in the lungs) or their chemical composition (if it can be dissolved it will be in the wet environment of the lungs but if it cannot it may cause phlem as an acute response) ...

Now, I would recommend just pick a dust sensor and go with it. The optical ones are the less power hungry and if it is IR based then its sensitivity is to small particles. The main thing would be consistency and so far the sharp sensors are "somewhat" consistent but their baseline is unit-depending so data de-basing would be necessary.

[Ed Borden]

I'm pretty sure you just blew my mind.

[Liz Barry]

Hi Gustavo, 

This is fascinating. Could you send a link to your manuscript? I would love to read more about your comparisons and testing of the $10 Sharp unit and the $5,000 instrument. 

Also, your discussion of the health effects of the smaller particles is more detailed than the conversations I've had with air quality scientists, which would be more like "PM2.5 can get into the bloodstream and it's bad," could you send a link to a good reference study on this?

thanks!

Liz

liz at publiclaboratory dot org

[Gustavo Olivares]

Hi Liz, as promised in the other forum, here is the manuscript I'll present (well, not me, one of the other co-authors will actually make the trip)

I leave it here for everyone's benefit but don't overuse it as the conference is not for a few months and I expect to publish a technical note in another place before then (a scientist's got to eat ;-) ).

Now, a caution note. The focus of my pacman is more the relationships between human activities and indoor concentration and that's why I added motion sensor and range finding capabilities which may or may not be useful for a more general "indoor-egg" kind of application. At the moment the "pacpersons" talk ascii through serial to relay data to a base but I haven't spent too much in the wireless design to add mesh networking to it but it is in the wish list (I'm doing what I can being a chemical and not electrical engineer ;-) )

Finally, this week there will be a "science" event in Auckland where I'll have a couple of the "version 5" units so probably my media office will take some pictures which I'll share here as soon as I can.

Regards!

El Gus

[Ed Borden]

Gus, we are planning on an indoor sensor box. It will probably be
offered as part of this current Kickstarter campaign. We'd love to
have you collaborate on the design (we can provide the engineering
that you are having trouble with) and probably achieve both of our
goals. Would love to have you involved there!

What we really need is to start a discussion here on specifications of
such a sensor box. Do you want to lead that discussion based on the
work you've already done?

[Gustavo Olivares]

Sure thing,

I'll start the ball rolling sharing what we've learned (it hasn't been just me) and hopefully we'll get somewhere :-)

/El Gus