Soil Respiration: CO2 chamber measurements


#1

Continuing the discussion from OurSci, collecting and sharing measurement data from sensors:

Let’s start a dediacated topics for such measurements setups, co2 chambers for measuring microbial activity soils.

I have ordered and tested a few different CO2 sensors, for cheap (relatively) from Ali-xpress and other online sources. notes about reading the sensors are documented on the hackteria wiki:
https://www.hackteria.org/wiki/CO2

sofar i have succesfully read the values from arduino and teensy, added a little OLED display and got quite confused about the sensor calibration and preheating… but getting there!

To Do:

  • make simple plug to connect sensor to board (ethernet cable?)
  • add ventilator to chamber
  • choose simple chamber, size? IKEA?
  • add humidity sensor into chamber
  • add something to take up the potential condensation of water?

First Measurements:


Soil respiration chamber Randelab
#2

More developements on using the MH-Z19 sensor on the Teensy 3.2 with OLED screen.
Got some nice code working now to show the CO2 measurement data and display the time when it reaches a certain threshold.
@Pannu what are reasonable values of soil respiration over time for soil?
and what kinda shape volume of a chamber?

still got to work a bit on the usability. maybe add another button to start the measurement, reset the timer, and export the full plot for further analysis.


#3

Marc,

I noticed your fine work on the CO 2 sensor. You asked about typical values in soil for CO2. I have worked with farmers and labs performing soil respiration testing and have developed a Jean understanding of CO2 respiration and its use. The CO2 value will be directly related to the temperature and moisture levels. No respiration at 32F and doubling for every 10C increase in temp.

Moisture is more of a curve that increases and is max at about 60-70% moisture. The base level of CO2 will be based on organic matter content and soil type.

If anyone has questions or need more information let me know. Keep up the good work all!

Bob

IngenuityWorx

207-692-2600

Bob@ingenuityworx.com


#4

That screen sure looks great. It’s really useful to see if the measurement makes sense from the beginning!

As mentioned, the amount of decomposing material, the quality of decomposing material, temperature and soil moisture are the main factors affecting CO2 production rate (if we assume that microbes are well and happy). The change of CO2 inside the chamber also depends on the (mean) chamber height (volume / surface area = (mean) height).

In good summer conditions, I would say that CO2 increase in a 10 cm high chamber could be around 1 ppm / s. This estimation is not accurate at all, but it should be easy to remember!

For CO2 I think in theory it’s good to use rather large chamber, but installing a large chamber “airtight” can be very difficult, because soil is never perfectly flat. Collars that are buried into soil well before the actual measurements can be used as well. Collar means an extension of the chamber that goes belowground and is permanently installed to the soil. The problem with collars is that the soil is disturbed during the installation, especially roots can be broken. Wet sand or something like that can be used to seal the holes between the collar and the soil, if digging the collar deep (5cm or so) into the soil is not an option.

So, I would recommend using as large chamber as possible, which in practice means quite small chambers (diameter 30cm or less). Because the signal of CO2 is so good, I would recommend tall chambers (20cm high or even more).

A perfect chamber is circular cylinder (best area/perimeter ratio). Also the formula for flux calculation (moles / m^2 / s^1) gets a bit simpler because h=V/A.


#5

#6


Great visit by @manuel to RandeLab and introduction into Our-Sci

Got it all installed and now ready to upload some data from my phone. thx for the intro!


#7

reading a bit more about that Soil Respiration on wikipedia…


#8

@dusjagr @Pannu
The Arduino code for the respiration chamber is available over here
https://gitlab.com/our-sci/measurement-scripts/tree/master/mh-z16-randen/arduino-mh-z16-randen

The Android app is found over here
https://play.google.com/store/apps/details?id=org.oursci.android.ourscikit

First time you connect via Bluetooth you need to pair the HC-05 with your phone first. You need to do that in the Android system settings, the pairing code is 1234.

I created a Survey with a sample ID and location question over here:
https://app.our-sci.net/#/survey/by-form-id/build_Randen-Soil-Chamber-Measurement-with-the-Soil-Chamber-built-at-Randelab_1525625829

It will show up in the Android app as “Randen Soil Chamber”.

Let me know how it goes and keep up the good work!


#9

Testing all the different sensors and soil-chambers using the RBSSA method (rotten banana skin soil analogue)


#10

In the graph below, some tests with two sub-100€ CO2 sensors: Airsense K30 (I think 2000ppm model), and MH-Z16. MH-Z16 is about half of the price of K30. Shortly: they seem to have very similar response, suggesting that they work as they should.

Both sensors were inside the same chamber that was kind-of airtight. The air space inside the chamber was mixed with fans. Using 9V instead of 3.3V seemed to decrease the response time of the sensors considerably (this is good news). MH-Z19B testing was inconclusive, because of some technical-practical issues. However, it seems that MH-Z19B (the golden box and also the cheapest of the sensors) has in practice very long response time (some minutes), and it does not get shorter by ventilation. K30 can explain only about 30% of the variance of MH-19b, which is low. This is because of the problems with the experimental setup, and the very different response time of the sensors.

Anyway here is the picture, showing that in the today’s test environment the sensors respond very similarly (which is great news).
Z16vsK30

Edit.: Here is an additional picture of the measurement, just to get the idea what was done.

Z16vsK30_temporal


#11

Updated the app to display the measurement zoomed too (auto axis):
https://app.our-sci.net/processor/randen-soil-chamber?formId=build_Randen-Soil-Chamber-Measurement-with-the-Soil-Chamber-built-at-Randelab_1525625829&instanceId=uuid:dcaa3549-061c-4c2e-bd03-4451001c5746&filename=measurement_2018-05-07T15%3A33%3A57.885%2B02%3A00.json

Should now be easier to detect trends in the curve.


#12

hmmm… making our lives easier. or is it cheating? only 40$:
https://www.aliexpress.com/item/Free-ship-by-quick-Singapore-post-5pcs-lot-Raspberry-Pi-Night-Vision-Camera-OV5647-raspberry-pi/32644108014.html?spm=2114.10010108.1000013.3.4336101f1EREEQ&traffic_analysisId=recommend_2088_2_90158_iswistore&scm=1007.13339.90158.0&pvid=e4741bc8-2b62-4dd5-8912-715e292bdcb6&tpp=1

i wonder what chip is on there. strangely the same system gets up to 120 $ on other ali-express sellers.

it’s using this NDIR co2 sensor:
https://www.amphenol-sensors.com/en/products/telaire-air-quality/sensor-modules/3215-t6700


#13

The sensor (T6713) looks really cool, it seems to have very low drift and does not require calibration. The sensor alone seems to be around 65 USD at least (mouser and co. ask over 100 USD).

Pretty sure the system for 40$ is without the sensor though.


#14

hmm… maybe.
we should just try to order that ali thing.

about callibration… it says it doesnt need… bt that means it’s also in ABC mode. auto callibration. and it even says that the calibration doesnt work, in case you use it in low co2 concentration. the ABC assumes regularly vented houses that once in a while are at atmospehric average of 400, and uses this as a baseline.


#15

Somewhere in the description of the cheaper kit, it says that CO2 sensor is not included.

The auto-calibration on MH-Z -sensors seem to require the sensor to be on constantly for 24/48h. By powering off the sensor (via software?) every 23h should disable the feature (just like some anti-calibration command). I wonder if the calibration changes the slope, or the offset. If it just changes the offset (‘a’ parameter in the calibration line: [co2]=a+bx), then it should not even matter in chamber measurements.

Ideally, the sensor should be able to measure concentrations under 400ppm. This would mean that we could measure photosynthesis. SenseAir K30 can do it (but it also autocalibrates).

It would be interesting to do some further testing on the sensors, namely response time with/without a fan, response to CO2 concentration / temperature / air humidity / how does the calibration actually works. At University of Helsinki we have CO2 calibration gases, that could be used. Furthermore, we have some serious CO2 analyzers, that could be used as controls (although I have no contract, I have contacts). It would probably take a couple of days for some semi-serious tests. I could to it at some point. I now have K-30, MH-Z16. It would probably be a good idea to have MH-Z19B and T6713 as well.

Testing/calibration CO2 sensors is one thing, and the next thing to test would be the chamber + CO2 sensor combination. We have a test setup for that in Hyytiälä. It would take about a week to do some serious testing, so this won’t be happening instantly.

P.S. Related to open research (open science): Under the Hyytiälä web page, check the Open data page, SMEAR online data and https://openscience.fi. SMEAR II is the measurement station in Hyytiälä. The new version of the site is just put online, and some information still seems to be missing.


#16

I am finally back to working on the Soil Respiration chamber. As i left all my working prototypes with you guys… @julian and @Pannu i had to make a new one.

still not so sure about how and if at all to make a PCB, wanna make some more tests first with hand-made prototypes. here is my newest creation: The BambooBox

Using the same hardware connections as the KiljuBox, 2 buttons, audio-output, bluetooth etc…
this time i added the Dallas Temperature sensor on Digital Pin 6 to also measure soil-temp.
no more use of that other CS8211 eCO sensor. it’s crap anyway.

next week, tuesday / wednesday our friend mat from Global Urban Forest will visit jogja and we’ll do some testing and comparison. Really wonder what’s in this BlackBox!


#17

It would be great to have some remote video-calls together next week! especially on tuesday.
@gbathree
@manuel
@Pannu
@badgeek


#18

Finally made that wiring diagram for the CO2 Soil Respiration sensor, aka BambooBox, formerly lknown as KiljuBox. It’s easy to reproduce, all components can be ordered from ali-express and only a bit of wiring and soldering is needed.

See the details of the connections to the Teensy here aswell, for reference and arduino code.


#19

I have not done much with the system now, except that I discovered a problem with Bluetooth, and also that my multi-meter does not anymore measure resistance.

In my opinion it’s better to give a 12V fan 5V instead of 3.3V (and definitely not 12V). Normally the mixing (fan power) should be as low as possible, but with the MH-Z16, proper ventilation will improve the response time considerably.

A small teleconference sounds good!


#20

Had a great session with Matt from G.U.F. and shared some methods, visited local permaculture farm and made some playful measurements with our new hardware.

See documentation growning here:
https://www.hackteria.org/wiki/CO2_Soil_Respirtion_Chamber#Soil_Geeking_.40_Lifepatch_with_Matt