AQE Ridiculous CO and NO2 Readings

[Mark Gibbs]

I fired up my AQE 48 hours ago and despite waiting for the sensors to settle down I'm still getting ridiculous readings such as NO2 at -2147483648 and CO at 112264 (see http://airqualityegg.com/egg/97759). I've also seen that the NO2 value reported on airqualityegg.com often disagrees with that reported on Cosm (https://cosm.com/feeds/97759).

So, do I have faulty sensors or did I just skip the part in the set up instructions where I was supposed to sacrifice a chicken first?

[vicatcu]

This has come up on forum.wickeddevice.com a couple of times so. I want to copy my most recent response to the group here. I don't have an answer to why there might be differences between cosm.com and airqualityegg.com as they are drawing from the same source data - but maybe the Cosm developer's can weigh in, or you can always post an issue to the Git-Hub repository (http://github.com/cosm/airqualityegg.com).

Everyone, I'll write a blog post about this to help clarify things, but I want to address various questions and comments here briefly:

1. The fundamental measurements of importance for the gas sensors are resistance, humidity, and temperature. The concentrations being reported are the hardware's best effort to interpret the resistance (ignorant of the humidity and temperature, as these relationships have not been characterised so far).
2. The best effort calculations being made by the Egg depend on a 'nominal resistance' stored on the device (per sensor) which is referred to as R0. The value can be field updated, as can be demonstrated in the PollEggBus_WriteTest sketch, and the community is working on ways do calibration in that way.
3. On-line calculations are ultimately what is needed to take into account the humidity and temperature readings to calculate the gas concentration.
4. The readings reported by the Egg are not meant to be used for gauging things like safety of life. We can't assume that kind of responsibility with this project. This is NOT a substitute for a home carbon monoxide detector / alarm.

[Joel Dubiner]

The first number is one way of representing -1 -- it may be an error value propagating up the API chain.

http://en.wikipedia.org/wiki/2147483647_(number)

(From the article)

The number 2,147,483,647 is also the maximum value for a 32-bit signed integer in computing. It is therefore the maximum value for variables declared as int in many programming languages running on popular computers, and the maximum possible score, money etc for many video games. The appearance of the number often reflects an error, overflow condition, or missing value.^[8]

If I were you I'd verify the following:

1) Is the sensor fully 'seated' on the board?  if it's loose, you may get crazy results.

2) Is the 'heater' for the sensor getting good voltage?  Are you running from wall power or a battery?  Check that there's enough power for the sensors.  Remember, they 'cook' the air in order to get a reading.

3) If you can find the correct points on the board, try to get at the voltage coming off the sensor.  It's an analog voltage that the microporcessor reads and scales to give appropriate human-friendly values.  If the sensor is putting out good data, then it's upstream.

4) All this implies that you have access to the sensor specs, a multimeter, and the board's schematic.  Because this is an open source project, you should be good to go.

5) What type of Chicken did you use?   I got mine from Amazon.

http://www.amazon.com/Squawkin-Rubber-Chicken-Novelty-Joke/dp/B0009Z3MQU/ref=sr_1_3?ie=UTF8&qid=1358365355&sr=8-3&keywords=Rubber+Chicken%3F

Some of the 'budget' chickens you find on eBay just don't do the trick.

-- Joel

--

[Mark Gibbs]

On Wednesday, January 16, 2013 11:43:25 AM UTC-8, jdubiner wrote:
> The first number is one way of representing -1 -- it may be an error value propagating up the API chain.

Indeed. Unfortunately there are no diagnostics and I wasn't planning on debugging the AQE ... not that I have time to do so anyway.

> If I were you I'd verify the following:
>
> 1) Is the sensor fully 'seated' on the board?  if it's loose, you may get crazy results.

Everything looks OK.

> 2) Is the 'heater' for the sensor getting good voltage?  

No idea ... I bought this as a fully assembled Kickstarter reward and if there's a manual covering this I can't find it.

> Are you running from wall power or a battery?  Check that there's enough power for the sensors.  Remember, they 'cook' the air in order to get a reading.

Wall power.

> 3) If you can find the correct points on the board, try to get at the voltage coming off the sensor.  It's an analog voltage that the microporcessor reads and scales to give appropriate human-friendly values.  If the sensor is putting out good data, then it's upstream.

No manual.

> 4) All this implies that you have access to the sensor specs, a multimeter, and the board's schematic.  Because this is an open source project, you should be good to go.

Yeah, I've got tons of spare time to debug something that is supposed to work.

Seriously, I didn't expect to get a project and, I suspect, neither did many of the other people who ponied up for Kickstarter. Kind of disappointing.

> 5) What type of Chicken did you use?   I got mine from Amazon.

I didn't but I probably should have.

> Some of the 'budget' chickens you find on eBay just don't do the trick.

Mwah.

Thanks.

[mg]

[Joel Dubiner]

I ordered the DIY kit, which is still pending shipping.  I'll post back on this thread if I run into any similar anomalies, and (more importantly) if I can find workarounds.

I used the Parallax CO2 kit for another monitoring experiment:

http://www.parallax.com/Portals/0/Downloads/docs/prod/sens/27929-GasSensor-v1.0.pdf

it gave reasonably accurate results.  I'll run some samples with my AQE and this sensor when my AQE arrives and I'll see how they compare.

You might be interested in something more along the lines of:

http://www.co2meter.com/?gclid=CKziupXw7bQCFUdxQgodgkQAtg

AQE is, to me, a fascinating *network of* sensors.  The actual sensor itself seems unremarkable.

I think AQE is of merit just for the dialog it has started.  If I get actual working hardware, that's sort of a bonus.

I live in Utah where I know the air is horrid for at least 4 months out of the year.  Don't need a sensor to tell me that.  But if enough people *care* about what they are breathing, we may be able to get some changes made.

See also: http://www.utahmomsforcleanair.org/ 

-- Joel

--