Soil Tensiometer Sensor Network

[aur]

My first post here. I am building a tensiometer sensor network. So far I've successfully built and calibrated a soil moisture measurement device (uses gypsum block sensors) which works on ATmega328 mini pro and supporting 4 sensor-blocks(on one node), measures soil moisture every 30 minutes. It's solar powered and also has a rechargeable battery.

While researching for different ways to create a network of these that can communicate wirelessly with a central console to log data, I came across 'MySensors' and cant seem to appreciate it enough! (No, Really.. Great Job!)

I was thinking of building my own console to communicate with the sensors (tensiometers - in my case) and log data, and after reading about "Vera" controllers, I am debating if I should use one of those instead. I could use some help with the following:

• Since I do not have any actuators in my project, (only passive data logging) is there an alternative way to communicate my data from the gateway to a computer terminal where I can log it? (instead of using of using a Vera Controller)
• If I do buy a Vera controller, can their GUI be customized at all? I am looking for ways to display the logged moisture data in graphs according to different sensor zones.

Appreciate any/all the help!

(P.S. Is this the right place to ask such a question?)

[hek]

@aur

This is a great place to present your project.
Will be interesting to hear more about the gypsum block sensor and how to to use it with arduino.

You don't have to buy a vera to get started logging data. There are several open source controller options that has logging capabilities running on low-tech hardware.

[olaeke]

@aur I don't think Vera is the ultimate controller for your project, I would have take a look an all open source controllers listed on the Controller page

[aur]

@hek From your knowledge and experience, is there a particular controller that you think is better suited for data logging and possibly displaying (PC interface) ?

Meanwhile I will investigate further into all the different controllers listed.

[korttoma]

We need pictures of your sensors and some more details about how you set this up:

@aur said:

solar powered and also has a rechargeable battery

[hek]

@aur

I wouldn't be much of a help here sine I haven't tried them all. Vera (which I run) is generally bad at logging data out of the box.

Maybe someone else here could recommend a system with good logging capabilities.

[tbowmo]

a perl/python script picking up the messages from the serial gw, and dumping them into a database, could be enough for what you want. And then some graphing engine to pull out the data for a nice view.

[epierre]

Hello,

It seems you are more advanced than me on the sensors, here i my project I've not yet completed and I would love to know how you did it:
soil-humidity-and-temperature-sensor-watermark-davis-granular-matrix

Regarding data logging, I have made my own perl script to read the gateway and use the data, I put it in a file, but I post it on a graphing application too so I can have any value I want. I've done that for my gas sensors (more than 10 !).

[pete1450]

Shamelessly self-promoting my post, but it really is cheap: http://forum.mysensors.org/topic/719/graphing-sensor-data/1

[aur]

I plan to make around 20 more of these sensors, so I am still in the process of optimizing the circuitry, but so far this is what I've done for my prototype node:

Regarding Solar powered:
I used a solar charger module from Seeed Studio (Li-Po rider); its pretty neat, in that it switches power from solar to battery only when needed. Moreover, v1.2 supplies up-to 350mA. I connected a 3.7V 2000mA Li-Polymer rechargeable battery and a 5V 100mA solar panel to this Li-Po rider which sources my Atmega328. The battery continues to charge while the sensor node is sleeping. I logged the sensor data in the ATmega EEPROM every hour. (As it’s not wireless yet! - that's what brought me to 'MySensors')

Regarding sensors:
After researching about the different options available, I bought a few different ones including a Watermark 200SS and Delmhorst GB-1. They are all based on the same principle – change in soil water tension – affecting resistance value. Watermark claims that their sensors do not dissolve. Their stainless steel jacket seems to gives it an increased life. I also created my own sensors, based on some information found on Van der Lee Vineyard blogsite. A simple series of resistance measurements from completely wet to completely dry for these sensors – home made as well as the ones I bought, helped me calibrate my home made sensors. I now feel comfortable using my own sensors and have acquired a reliable degree of accuracy and precision on my little tensiometer. Which is why I am ready to expand them into a network.

Now:

• I'm looking for different ways to log this data and have it available on simple graphs. (Currently, studying various controllers mentioned here. Btw, it is possible at all to use EasyIoT directly on a PC instead of on a Raspberry Pi?)
• I've bought some nRF24L01 modules with PA and SMA antenna and will perhaps have to design a simple 3.3V regulator to source it directly from the the Lipo Rider board (since the pro mini wont be able to meet my current requirements(?) for a longer range - I want to be able use them on maximum power for long range.)

I will try to keep posting more updates as I have them.

Meanwhile, any help/ideas are appreciated.

[aur]

@hek:
My radio communication using the MySensors library is now successful. I am able to transmit my soil moisture data from the prototype sensor node to a gateway - which I have connected to a computer. I am looking into building a basic windows app (using visual studio, C#) that will read from the serial port and log data into a database. Later this data can be made accessible via queries/graphs.
I think this should be doable given that serial protocol used by MySensors library is a semicolon separated string of values. Is this is a reasonable assumption?
From your experience(and expertise!) do you foresee any difficulties/incompatibilities in implementing this?

P.S: I am also looking into using Vera - just bought a lite version off of Amazon - but if the Visual studio plan works out, I will return the Vera. Btw, Vera's customer service rep told me about this 'datamine' plugin that can log and display graphs. Thought you'd like to know.

[hek]

@aur said:

From your experience(and expertise!) do you foresee any difficulties/incompatibilities in implementing this?

Nope

[epierre]

@aur how is your project by now ? anything to share ?

[NeverDie]

I looked into this once. Those gypsum sensors are easy to use but not very accurate, not even the ones you buy. Plus, they do eventually fail, and how do you know when that has happened unless you dig them up every year to inspect them? Two failure modes:

1. The gypsum (or granular matrix or whatever) dissolves away, exposing the electrode metal. Corrosion ensues, and accuracy plummets.
2. Similar to 1, except sensor loses contact with the surrounding soil. This is why they tell you to dig them up every year and replant them. It's supposed to help ensure good sensor to soil contact.

Checking soil moisture levels using a circuit that measures capacitance seems like it would be more robust.

[epierre]

@NeverDie this looks like what Vinduino says.

are you using 3 sensors at one place ? are the sensor auto powered ?

[NeverDie]

@epierre said:

@NeverDie this looks like what Vinduino says.

that's the first II can remember hearing about it.

are you using 3 sensors at one place ?

yes

are the sensor auto powered ?

yes

[epierre]

good, yes Vinduino is the only ... and has been awarded for this !

How do you auto power ?

and how deep do you make each sensor ?

[NeverDie]

@epierre said:

good, yes Vinduino is the only ... and has been awarded for this !

???

How do you auto power ?

Not sure what you mean by "auto power." Each soil moisture sensor runs from batteries.

and how deep do you make each sensor ?

2", 4", and 8" deep. I picked those depths because the purpose is to monitor the soil moisture content for my lawn. Probably 80%+ of the grass roots are within the top 8 inches or so of soil, so that's where most of the water gets absorbed and used by the grass to drive photosynthesis.
Here's a real-time plot:
https://plot.ly/~WhiteRabbit/997
As you can see, there's little change in soil moisture at night, when there's effectively no photosynthesis happening.

[epierre]

@NeverDie auto-power I mean sustainable energy, solar pannel... nothing to change and dispose of batteries too fast...

I have the same issue with lawn, and I would like to have something to look for because it is not flat but that doesn't explain my problem in whole.

[NeverDie]

These look like they would be interesting sensors, but it would be expensive if the foundation for a large number of soil moisture sensors deployed in a home network:. http://www.vegetronix.com/Products/Soil-Moisture-Sensor-Probes.phtml I believe it may use some sort of high frequency capacitance measurement, in which case I wonder whether the Arduino's capsense library would yield results which are just as good? The high frequency apparently has some advantages.

Adafruit sells this: http://www.adafruit.com/products/1965?gclid=CjwKEAjw6IauBRCJ3KPXkNro1BoSJAAhXxpyN6VAQf445eQ7-3EGkxjcmui4-FUFiQGbd73y2PEBrxoCu_7w_wcB
which is meant to be hackable, and it's a lot more affordable too. However, it's not fully waterproof, like the vegetronix is, so that's a major negative. On the other hand, it is open source, and the details are here:
http://wemakethings.net/chirp/
It even gives Arduino code for reading it. It actually uses an attiny44 to do the capacitive measurement!
If you could make the entire thing waterproof (or at least the probe), wireless, and "auto power," then you'd have a very useful sensor!

This document: http://dspace.library.colostate.edu/webclient/DeliveryManager/digitool_items/csu01_storage/2012/02/14/file_2/120523 compares a number of soil moisture sensors. It's worth noting that the watermark wasn't deemed accurate enough to simply use out of the box but instead required a correction factor to be developed. However, if you need to do that anyway, maybe make your own or use a different technology altogether--one that won't dissolve on you and require annual digging up and re-implanting.

It probably does make sense to get a tensiometer to calibrate whatever you end up buying or making. That's something I'd like to do, maybe next year.

[epierre]

@NeverDie we have the same reading and objectives, so far I didn't finished the sensors because I was stuck with solar and LiPo/LiOn powering, but we may task up to make something interresting !

[NeverDie]

I'm presently using soil moisture sensors that measure soil conductivity, which might be OK if I never fertilized my lawn. However, when I do fertilize, it seems to gradually affect conductivity, and the measurements lose calibration. Definitely not ideal.

Apparently each method has its own limitations, but I'd like to try capacitance and see if it maintains calibration better.

Most likely someone out there has already solved this using an arduino, or with a simple circuit connected to an arduino to get the required accuracy. These guys seem close to an answer like that: http://zerocharactersleft.blogspot.com/2011/11/pcb-as-capacitive-soil-moisture-sensor.html
Based on that discussion, it appears that high frequency is required to avoid salinity (such as from fertilizer) throwing off the readings.

In any case, there seems to be a loose consensus that 80Mhz is the fright frequency (e.g. http://www.surechem.com.my/products/901003-100105-PDF.pdf)

[epierre]

or you have the other way around: Infragram DIY Plant Analysis Webcam
http://store.publiclab.org/products/infragram-diy-filter-pack

As the vegetronix, I have in RFX433 the Imagintronix soil moisture sensor: http://www.imagintronix.co.uk/shop/4587637347/soil-moisture-sensor-check-water-level-in-your-pot-plants/9163185

This is already quite useful to see water propagation in a 2x2 kitchen garden.

[NeverDie]

@epierre said:

or you have the other way around: Infragram DIY Plant Analysis Webcam
http://store.publiclab.org/products/infragram-diy-filter-pack

Doesn't work. I've been down that path already. In actuality, it mainly just tells you if your grass is alive or dead. That's because dead grass doesn't have chlorophyll and isn't green. You'd get the same results looking at a color picture and noting how green or brown it is. Checkout http://www.publiclab.org/profile/WhiteRabbit and the follow-up discussion in those posts.

You could try looking for other indicators of grass stress. For instance, grass does change color the drier it gets. If you have the skills, you might try measuring that. It might be a fun science project, but in the end it's going to be more complicated and expensive. However, it also changes color based on fertilization, so that's going to throw off your calibration.

As the vegetronix, I have in RFX433 the Imagintronix soil moisture sensor: http://www.imagintronix.co.uk/shop/4587637347/soil-moisture-sensor-check-water-level-in-your-pot-plants/9163185

This is already quite useful to see water propagation in a 2x2 kitchen garden.

I'm pretty sure that one isn't capacitive but actually measures resistance. Take a look at the probe near the tip. Do you see exposed metal contacts, or is it all sealed up? Pull it out and have a look. I'll wager you'll see two small metal pads near the tip, and it's along that gap between them that soil conductance is being measured.

[epierre]

@NeverDie in fact I'm colorblind so I need something to tell me between green and brown...but that's a very interresting discussion !

[NeverDie]

In that case technology is your friend.

Regarding # of soil probes and depth, I'm finding that the deeper one is probably the most important, at least for established turf. You can keep the upper layer fairly moist by watering for too short a time, and interestingly, the water never makes it very far down. Here we have clay soil, and it may be different for sandy soil. Here the result is that the roots at the deeper layers keep pumping the water out, and it just gets drier and drier at the deeper layer. It's invisible: you can't know that from just looking at things or feeling the top of the soil. In fact, without a soil probe (or a shovel) I don't think there's any way to know that such a thing is going on.

On the other hand, most grass roots are shallow, and if push comes to shove, I'm not sure to what degree deeper watering saves water or is a waste of water. I think it might save water, though, because the shallower soil is more prone to evaporation. The standard advice seems to be to water infrequently, but deep.

[epierre]

@NeverDie in engineer school I couldn't go and grab resistors nor check the values... a pity...

you said about 2", 4", and 8" deep, but given the sensor size are 2 and 4 not too close this way given their own size ?

[NeverDie]

Not sure what you mean.

[epierre]

@neverdie 2'' and 4'' are pretty close in depth (unless you put them horizontally, not vertically).

in international units that makes 5, 10 and 20cm depths, my sensor is already 8 cm in height.

[NeverDie]

@epierre said:

@neverdie 2'' and 4'' are pretty close in depth (unless you put them horizontally, not vertically).

in international units that makes 5, 10 and 20cm depths, my sensor is already 8 cm in height.

At present, I'm undoubtedly using a different sensor than you are, and I mounted it horizontally at the depth indicated, not vertically, if that's what you're doing. At the time I installed it, my goal for each sensor was just to measure soil moisture at a given depth, not across depths.

[epierre]

@NeverDie have you gone to use the EvapoTranspiration calculus such as described below (example in the second part after the theory)

http://www.konza.ksu.edu/weather/ReferenceET.PDF

[NeverDie]

@epierre said:

@NeverDie have you gone to use the EvapoTranspiration calculus such as described below (example in the second part after the theory)

http://www.konza.ksu.edu/weather/ReferenceET.PDF

I haven't. It would be a great help if there were a free library that worked with generic weather station gear. It would seem that Davis has a non-free library that works with their particular weather station gear, but aside from that, I don't know of much else.

[Lawrence Helm]

Also looking at soil moisture probes, either https://www.tindie.com/products/miceuz/i2c-soil-moisture-sensor/ which is based on the chirp one. Just need to seal the sides, or otherwise http://www.aliexpress.com/store/product/Soil-temperature-and-humidity-sensors-SHT10/1773013_32356151782.html

Is anyone using the SHT ones ? (hopefully no pun intended)

[epierre]

@Lawrence-Helm sht10 are expensive and are sensible to humidity, they are thus not made to stay in the ground or water.

the soil humidity need to reach the captor but this one is not meant to support being in a too high humidity environment.

Otherwise they are good ones

[Lawrence Helm]

@epierre the SHT10 that I put a link to has the following statements:

• Can be completely submerged.

and

• Usage:The probe can be directly embedded in the soil

Also the following link: http://www.adafruit.com/products/1298

Are your thoughts still the same, before I bite the bullet

[ahhk]

read the description:

"but it's always best to avoid long-term (over 1 hour at a time) submersion, if you need something that can be submerged for over an hour you may want to find a different sensor. "

[epierre]

@Lawrence-Helm I've got a collection of them and looked at their assembly, clearly don't trust the ill traducted or too much over-selling descriptions when you put electronic in a very humid environment

[haro]

@epierre Hi...it is ameasing your work!!

I am wondering about Mq-135. We have troubles with reading differents gases.
From the datasheet we read the slopes
Gas X" X0" "Rs/R0 "Rs/R0 pendiente (log y- log y0)/(log x-log x0)"
Air 200 10 3.6 3.6 0.00000
co 200 10 1.3 2.9 -0.26783
nh4 200 10 0.79 2.7 -0.41024
c02 200 10 0.8 2.4 -0.36673
3/4/ 200 10 0.72 1.9 -0.32391
1/4/ 200 10 0.64 1.5 -0.28432
+- 200 10 0.59 1.4 -0.28845

these data are in the next worksheet https://docs.google.com/spreadsheets/d/18Z6XyxwXVsWdMIk9vbUuCucaLJDdmR0QUFSTYG79ftw/edit?usp=sharing

¿may you share with us your code for this sensor in order to see your slopes and Rs/R0 values taken from the curve?
¿do you know the meaning of the name go gases with 3/4 or 1/4 or +- ?

thank you in advance

[nervusvagus]

following this thread loosely.

I have the following sensors and they seem to be accurate to some degree:

http://www.aliexpress.com/item/DFRobot-Capacitive-Analog-Soil-Moisture-Sensor-3-3-5-5V-Corrosion-Resistant-with-Gravity-3-Pin/32574020064.html?spm=2114.01010208.3.2.P99ddH&ws_ab_test=searchweb201556_0,searchweb201602_1_10037_10017_507_10033_10032_10040,searchweb201603_1&btsid=8d3b5e46-a069-441a-a1f7-491ae7e317e1

and they are much easier to work with Arduino compared to Watermark stuff. Anyone agree?