Electrical DIY Atmospheric Carbon Dioxide Detector?

[jenny_shoars]

I thought it would be an interesting project to build a detector that will measure and log the outdoor atmospheric CO2 levels. These (global average) levels are currently about 400ppm and these (global average) levels vary by a few ppm every year. Looking for a sensor though, I can't seem to find any that have an accuracy much better than ±30ppm. That is, unless I setup a spectrum absorption detector or something similar, but this might be a bit out of my price range, especially, since I was hoping to have a small circuit which I could run 24/7, constantly logging data, and running off a solar panel so I didn't have to do much to keep it running smoothly. Anyone know of a better solution? Perhaps an accurate sensor I missed? Or might the variation per measurement be enough that the ±30ppm would still give good average results? Thanks!

 

[rootone]

As with any equipment, the more accurate you want it to be the more expensive it is.
I did a bit of browsing and came across these, (I am not in any way associated with the vendor).
http://www.temcocontrols.com/products/sensors/c02-hum-aq/co2-sensors.htm
These all see to be within a price range that's not excessive.

You could of course try to make something similar from component parts, but I'd say that will end up being more expensive than buying a manufactured unit, since you will be buying the parts at unit cost unlike a manufacturer who can buy them in bulk at a lower price.

 

[jenny_shoars]

Thanks. These are ±70ppm. The sensor I was looking at (MG811 Sensor) is ±50ppm and only $35. So with a homebuilt setup (add on a simple microcontroller and something to power it), I could probably get away with a setup just over $50 (excluding the solar panel cost).

If there isn't a better option, the only other question I would have for someone that might know, is if the variation from measurement to measurement would be more than 50ppm. That is, if the variability per gust of wind is more than that. In such a case, the measurements would still produce a useful outcome when averaged. However, if the variation is only a few ppm, then if the low accuracy is caused by some systematic error, the values might not be useful. Anyone happen to have a bit of insight on this?