sensor is not enough, dynamic air sampling with filters and very sensitive circuitry are needed

[bruno aubert]

Dear all,
My name is Bruno and I'm the founder of Cairpol company.
I've just reading your posts and I really appreciate your project. I
only want to give you answers to some posted questions about our
sensors: for gases we're using Membrapor's electrochemical sensors
(Swiss manufacturer).
I wish to give you my opinion: of course you need the best (and
cheapest) sensors but it is not enough! In the aim to have the better
accuracy like references devices, you need to add dynamic air sampling
with stable airflow and filter to smooth interferent variations like
humidity. Moreover, generally all sensors can't reach so high
sensitivity needed for air quality monitoring (less than 50 ppb): you
need very sensitive and stable electronic circuitry in the aim to
measure nanoAmps currents.
I wish all the best for your project
Sincerely
Bruno

[Dirk Swart]

Thanks Bruno, this is useful information.

Cheers
Dirk


Follow me on Twitter now: dswart

[MattDance]

Below is the text of a conversation that Bruno and I had about AQ
monitoring in general, and a couple of technology deployments.

I am working on a project at the University of Alberta using the
sensaris.com AQ sensor, and of course I find the aqegg to be very
appealing for its cost and ability to be hardware hacked on a
weekend. Very cool. The issue that I am struggling with is the
context. What is the data context envisioned? How do we qualify AQ
data that might not have a correlation to a regulator run regional
system with a detailed reference method? I say this as an open data
activist (I did the data work on emitter.ca) who wants to deploy
similar sensors with an open data portal.

Thanks in advance
Matt

Bruno,
It is important to distinguish between the different types of AQ
monitoring that occur, and the niche that the AQegg is filling. I
think the AQegg is not a compliance monitoring tool, nor will it be
used by regulators to define a regional AQ (health) index.  At least
in the short term.  What you describe as a protocol for monitoring is
absolutely vital  for a regulator or industry that is monitoring AQ. I
think that the egg as two roles to play beyond the monitoring role of
gov't and industry:
1. The egg is about citizen engagement where the sensor is able to
provide some relative measure of a person's AQ in an urban setting.It
will be interesting to see what the numbers are for different cities,
and if they are higher (and therefore more detectable) than at a
regional level. For instance, the urban AQ monitor in Edmonton is at
the top of a building. I am curious what the NO2 numbers are at street
level - exactly where the egg will be deployed. In addition, as we get
more data (and at the egg price point which I expect will be less than
200USD, there is potentially a lot of data) there is a good
opportunity to explore the correlation of the egg numbers, with
climate numbers (the egg does provide RH and TEMP?) and the regional
monitoring.  You can do a lot with marginal data if you have enough of
it.
2. The egg is pushing the monitoring technology envelope as well as
the citizen science, open data and open government envelopes.  This is
a VERY important consideration.




Thoughts?




MattDance (@mattdance)
Dear Matt,




You're perfectly right: The AQEgg is not a compliant monitoring tool
and I think I didn't make misunderstanding about this.
I agree totally with you concerning the incredible advantage to have a
lot of sensors and even these sensors are less accurate the results
could be better than AQ monitoring.
My company is now developing wireless sensors for monitoring diffuse
emission (gases, odors...) at ppb levels around facilities or waste
treatment plants. To date only one compliant monitoring device was
working in the facility (because the high price) and computerisation
was to be made to survey these emissions outside the facility and the
result was not so good. With our sensors it is now easy to monitor on
line the emissions and to adapt immediately the process in the aim to
limit these emissions.
I was just trying to give you my opinion based on my short experience:
if your AQEgg doesn't reach a minimum requirement concerning accuracy
you'll probably the same problem I met few years ago when I started my
project (it was concerning a cheap and miniature AQ device dedicated
to the fragile persons like asthmatics).
When you're talking about "some relative measurement" I think it is a
problem: sensors (you can buy today) without right technical
environment (like my previous post) could give bad measurements
depending of temperature, humidity, cross interferent.....and these
variations can reach more then 10 times the right value (you can find
a lot of tests reports by internet especially for NO2).
Moreover you wrote you'll be very interested to compare AQ monitor at
the top of a building in Edmonton and values (by AQEgg?) at street
level: you can't (even you test the AQEgg before at the same place of
AQ monitor).
I agree it could be possible to decrease the price of AQEgg down to
200$ even with good accuracy.
I sincerely hope you consider this e-mail friendly and don't hesitate
if you have another questions.
Best regards,
Bruno
PS: I don't know if you want I post this answer, it is for this reason
I sent it by e-mail


Hi Bruno,
Thanks for your email! Is this you? http://www.cairpol.com/index.php?lang=en.
It looks like a great product. I plan on exploring your website to a
greater extent over the weekend.

I recently purchased the Sensaris EcoPod. What do you think?

Also, please feel free to post the conversation stream. It may be
useful.

Best,
Matt


bruno aubert bruno....@cairpol.com
Feb 25 (13 days ago)

to Matthew
Hi Matt,

Yes, its me. Thank you.
Concerning the Ecopod, I think it could be a good product especially
concerning wireless communication system but it is not so hard for
good engineer to design it (I made it too).
As I wrote the most important is concerning the sensor. All companies
like Sensaris and even bigger one like Honeywell or Draeger are making
the same error: they believe you have just to connect sensor on the
electronic circuitry. It is not true: if you want sensitivity down to
few ppb and good accuracy you have to design specific arrangement
around the sensor.
Few years ago, Sensaris launched field test in Paris with his sensors
for NO2 monitoring (project name "la montre verte" the green watch).
Sensaris gave the results on its Web site and I checked the pollution
levels with the several regulatory AQ stations near this field test
and I observed differences up to 3 times. Consequently the agencies in
charge of the AQ monitoring had a bad opinion of these kind of sensors
and I met myself some difficulties to recover the confidence of these
agencies: it is for this reason I try to warn you concerning the using
of sensors.
By the way, what is the price of Ecopod?
OK, next time I'll post my answer.
Best regards,
Bruno

[Ed Borden]

Matt, thanks for posting this. Huge amount of great stuff in here. I
have to tell you, this is the right conversation to be having and I'm
really excited about that. First, there has to be an understanding of
what we are trying to accomplish and there has to be empathy with
that. We are starting in a place that is extremely imperfect from the
point of the view of the sensors, but we are making a statement as a
community. Once we get some legs as a community, the refinement of
our methodologies will get better and better.

Every deployed Egg is like a beacon. It says, "Hey, I care about this!"

Every other AQ project starts from the other side and THEN moves
toward people/users/citizens. We've flipped that model here.

[Cesar Garcia]

Maybe this is also a way to make clear that cheap sensors are not available for these citizen initiatives and hopefully someone will come  with low cost better alternatives.

Let's hope so and go on forward!

[Matthew Dance]

Ed & Cesar, thanks for your comments!  I have a couple thoughts about the data end:

1. I am working with a local airshed zone - the Alberta Capital Airshed Alliance (http://www.capitalairshed.ca/) based in Edmonton, Alberta, to potentially run some tests on the sensaris NOx sensor comparing it to the a standard regional AQ NOx sensor. This would be a great opportunity to test an AQegg as well, and I would love the opportunity.

2. If we are able to get a significant sensor penetration within a limited geographic area, there is a spatial-statistical case that can be made for what the data is saying.  We would need a lot of AQegg data points, and a reference to a standard method (i.e. from above) to do this work. But it is worth it as a powerful validation tool.

3. I believe there is a case to be made for various levels of sensors providing a range of fidelity, as long as we know what each sensor is saying. For instance, perhaps we have three levels of sensors deployed on one monitoring system - (1) many AQeggs as a low cost citizen deployed method; (2) fewer of Bruno's sensor (they are more expensive and therefore less accessible); (3) perhaps a dozen compliance sensors.  There is a way to reference the sensor type  spatial-temporally to the compliance sensors such that a more robust suite of data is generated - in a sense, qualify the probability of any one data point being accurate.

4. We will never move entirely away from a regulated or compliance sensor network.  I agree that we are making a statement of caring intent be developing and deploying the AQegg, but we also need to work with other monitoring bodies (I believe) to put those data into context.

Thoughts?  Thanks!

Matt

--

| Matthew Dance | Graduate Student | University of Alberta |

| 780.554.9222 | @mattdance |

[Ed Borden]

1) I will get you a prototype asap. Send me your contact and shipping
info directly.

2), 3) Agreed.

4) You sound like the perfect liaison for the community to those
bodies. You have my nomination :)

[charline]

Hi guys,

Your messages are really interesting and I think we are all following
the same will: giving to people the possibility to understand the air
quality around them.
I work for Sensaris (the ECOpod sensors) and I think it is a bit wrong
to think small sensors are just connected electronics. We have data
correction implemented into our devices but it is not enough to make
it accurate. I guess the AQEggs share the same problems than us
regarding calibration and MOS sensor sensitivity to humidity and
temperature. But this is a matter of time before we find both solution
about it (because I guess this is also a strong concern for Pachube
team). We know that it is possible to have 3% of accuracy compared to
fixed stations with ozone sensors so we can see that it is possible to
use those small sensors for good results.

I think it may be really interesting to work on platforms that would
be able to integrate different kind of data. I think the advantage of
the ECOpod is to let people using the application they want, and if
the AQegg is designed the same way, it will offer a better choice in
sensors. This flexibility seems necessary to me to find the right use
of those sensors.

Matt, if you agree to share the results about our sensors, it would be
great since we are looking for improving this ECOpod. We are also
making a comparison with fixed stations here, we could both share the
results to see the differences if you'ld like to. We are strongly
working into a solution that is easy to use for all people since it
would help to build meaningful networks but we have to improve the
ECOpod again and again to do so, there are several european projects
about to start.

Bigger networks would also help to strongly decrease the sensors costs
and make it available for citizens. Well, hope some initiatives could
help to do this !

> On Fri, Mar 9, 2012 at 08:56, Cesar Garcia <cesargarcias...@gmail.com>wrote:
>
>
>
>
>
>
>
>
>
> > Maybe this is also a way to make clear that cheap sensors are not
> > available for these citizen initiatives and hopefully someone will come
> > with low cost better alternatives.
>
> > Let's hope so and go on forward!

> > El 09/03/2012 16:41, "Ed Borden" <borden.edw...@gmail.com> escribió:
>
> > Matt, thanks for posting this.  Huge amount of great stuff in here.  I
> >> have to tell you, this is the right conversation to be having and I'm
> >> really excited about that.  First, there has to be an understanding of
> >> what we are trying to accomplish and there has to be empathy with
> >> that.  We are starting in a place that is extremely imperfect from the
> >> point of the view of the sensors, but we are making a statement as a
> >> community.  Once we get some legs as a community, the refinement of
> >> our methodologies will get better and better.
>
> >> Every deployed Egg is like a beacon.  It says, "Hey, I care about this!"
>
> >> Every other AQ project starts from the other side and THEN moves
> >> toward people/users/citizens.  We've flipped that model here.
>

> >> On Thu, Mar 8, 2012 at 4:17 PM, MattDance <dance.matt...@gmail.com>

> >> > bruno aubert bruno.aub...@cairpol.com

[Matthew Dance]

Hi Ed,

1. Please let me secure the funding before you fire a prototype my way! I don't want to draw on resources until I know about the cash, and I will let you know ASAP.  Thanks for this!

2 & 3. I will keep you and the group in the loop on this!

4.Thank you ;). I accept the nomination!

[Matthew Dance]

Hi Charline,

Thanks for your note.   I am curious about the data correction process that is implemented into the Ecopod.  What is it and how does it work?

Yes, I am very interested in sharing the data that is generated from the testing that will hopefully occur (it is contingent on funding).  I can only speak for myself in saying that I am working with the intent of openness and transparency, but I trust that the AQegg group shares the same principles!

Thanks,

Matt

[Gustavo Olivares]

Very interesting discussion!

As with most projects, the key is the objective.

I am part of an air quality research group in New Zealand (www.niwa.co.nz) and we've been exploring "alternative" methods for a few years now and we always come back to the "why" and "what for" questions.

I could picture the AQEgg in several scenarios (they are non-exclusive and definitely not exhaustive):

1.- Ed's "Beacon" idea. Where the main goal of each unit is to be used

2.- "Crowd sourced AQ". Here, the main goal of each unit is be part of a network and use the power numbers to have robust estimates of (let me call it) traditional AQ

3.- AQ mapping. In this case, the goal of each unit is to be consistent with the rest that are deployed nearby. By consistent I mean that the differences between the readings of 2 units are because they are in different places and not because they're different units. Note that this does not require the same baseline for all the sensors, it only requires that the sensors respond identically to changes in concentrations

4.- Population exposure. Too often we forget that the purpose of AQ management is not to measure concentrations but to reduce the health effects of pollution on the population. There are hundreds of articles documenting the shortcomings of fix-point measurements as indicators of population exposure because we move so a dense network of sensors can overcome that by effectively filling the gaps where people are moving. The ultimate expression of this is "wearable sensors" that provide information of individuals "total" exposure.

Now, in this set, only cases 1 and 4 are of relevance to the individuals operating the eggs and that's when the performance of the individual sensors is very important, but I think that its importance is related to the "so what" question. This means "what could the user do with the information?" and both cases are different. In the first case, the relevant feedback to the user is a heartbeat that indicates that it is in fact working.

The 4th case is a little trickier. What if your sensor (that you're carrying or wearing) tells you that your habits are poisoning you but those habits are imposed by external factors? For example, I carry my sensor all day and it says that while I'm driving to work my exposure is 20 times higher than when I am either at home or at work (we've done some experiments and that's about right) but that's the only way I can get to work ... other than awareness about my slow death I haven't gain anything from it. One could argue that I have and can now put even more pressure on local politicians to get a decent public transport system in place but how is that outcome different from the "beacon" scenario? The answer, in my opinion, is that if the instrument is sensitive enough to capture changes in my exposure due to changes in my habits, then I (personally) can use it, otherwise it's just a novelty.

Finally, and coming back to some of the points made by people here, ease of use, low cost, robustness, "networkability", sensitivity and accuracy are all important  but for different reasons and in different situations. As a way of example, our latest "instrument development" project has been an "indoor characterisation" unit (version 2 of the files are in: https://bitbucket.org/guolivar/pacman but version 5 should be up in a few days) that measures air quality (dust, CO2 and CO) together with "context" information (motion sensor and ultrasonic range finder) and we made the decision at the beginning that the user (the houses where we'll deploy this) will have no feedback from the units because we don't want them to change their behaviour. On the other hand, we had a failed proposal that was aimed at quantifying changes in behaviour by fitting a group of cyclists with personal exposure monitors that would give them immediate (light on the sensor) and delayed (once the data is downloaded a map can be created) feedback and we would observe and quantify the effects of that information on the cyclists behaviour.

If you got to here reading ... congratulations it wasn't my intention to be this long it just happened!

Regards

El Gus

[Cesar Garcia]

I think you post raised several interesting points about the project, so thanks for sharing your experience!

Regards,

César

| 780.554.9222 | @mattdance |

--
Cesar García - @elsatch

Ando con encolamiento para responder correos y los proceso lunes, miércoles y viernes. Si es algo urgente/rápido contáctame por Twitter. Gracias!

[charline]

Hi Matt,

About the data correction, we can vary sensor voltage by using
relative humidity and temperature. We are making tests to check our
sensors sensitivity and we are looking for a calibration partner. We
have some clue but we won't implement it into the software before to
complete the calibration process. So having the data from your side
would a great thing to compare it with the results will get from here.
If everything is coherent, it will help a lot into the improving
process of our sensors.

Gustavo, your points are really interesting and I think all the design
around the final product is important. There are plenty of
applications for wearable devices like the EcoSense, each applications
change a lot of things : design, accuracy, gases measured, IT infra,
softwares...concerning point 4, I guess it is a really valuable info
if you can change your route to your work. If you're able to see that
using a car or a bus change the air quality, if you see differences
concerning the hour in the day, you can make choices. I think it's a
beginning :)

Thanks,
Charline

> ...
>
> plus de détails »