Self sustained garden « Jee Labs Talk

Self sustained garden

(24 posts) (6 voices)

Tags:

No tags yet.

pedro

This all started while reading about environment and recycling. Apparently about one third of the home garbage is organic. So I started looking around how to use this organic trash. The problem with standard compost machines is the smell and size. There is another type of compost machine that uses worms. This is much better to have in an apartment. So I got one here: Of course, if I'm going to have lots of compost then I need a garden where to use it. I got myself 2 containers of 100 liters of soil each and installed them in my balcony. I'm starting to plant some vegetables in them. Now the electronics part! I'll be on holidays for 3 weeks so until then I have to figure out how to water my garden while I'm away. For this I need:
- stand alone controller (JeeNode) - RTC module + output module - water pump - battery (+ battery charger) - soil humidity sensor - water container - some plastic pipe
I'm going to start by the soil humidity sensor. I got the idea here: I'm going to change it a little bit. I already built one with plaster of paris and I'm just waiting for it to dry out. I'll publish here the results of my testing with this sensor.

Posted 08/06/2010 21:39:45
mdamen

Cool, I am working on a small soil sensor myself.. not working that great so far :-( I want to measure my indoor plants moisture level.. I've bought a cheap moisture sensor from dealextreme.. and took it apart. Problem I have is that once the probe is in the water it reads fine, however when I remove it from the water I get very unstable readings (it goes up and down)
I still need to test it in real soil though.. right now I am just testing it in a glass of water.

Posted 08/06/2010 22:44:32
jcw

See also this contribution from a while back: moisturesensor/>

Posted 08/06/2010 23:06:56
jeroen

Pedro, I have the plaster sensor working now for a few months. First plaster module was to big. After some rain it took days to dry. Instead of the nails, (they rust) I use copper wire, that works fine.
regards Jeroen

Posted 09/06/2010 08:49:42
pedro

@mdamen: A plaster sensor should be more stable but with a big response delay taking a while to dry out after the soil is dry.
@JC: Yes, I already saw this one on your discussion group, in fact it came out as a related topic for this one...! By the way, how can I put some pictures in this topic?
@jeroen: Thanks for your feed back. In fact I'm searching the good compromise between size and response time. If too small then it will not last long. If too big then it want dry out. I just build the first one out of two galvanized nails and the mold was a 35mm film plastic round box. So it is 3cm diameter by 5cm tall. I hope that the resistance will be different for different levels of humidity. I'll publish the results here.

Posted 09/06/2010 19:52:07
jcw

> how can I put some pictures in this topic?
They need to be stored somewhere else, then you can point to them using \!\ - see the [Markdown syntax](http://daringfireball.net/projects/markdown/syntax#img) page.

Posted 09/06/2010 20:00:33
tankslappa

Pedro, go sign yourself for a free account on flickr, then you can host your pictures on there and link to them. I think the limit for a free account is 200 pictures, or something like that.
You might want to submit your pictures to some of the groups on there too, there are groups for everything, including arduino based projects... Although there doesn't seem to be one specifically for JeeLabs... yet ;-)

Posted 09/06/2010 20:07:57
jeroen

Pedro, I used also a 35 mm film box, and 2.5mm electricity wire. (I tried with smaller but the rusty nails broke the plaster, I created also larger sensors (plastic coffee cup), but they were much to big and were never dry. For me the response time is fine. The sensor is +/- 3cm below the surface.

Posted 09/06/2010 20:45:34
pedro
@jeroen: Uau, we had the same idea ;-) I was looking around the house when I came across the empty 35mm film box and thats when I had the idea of using it. So according to your experience this should work fine. Good.
@JC: humm... I'm using the ports library to read the voltage value of a voltage divider where I use a 100K resistor connected to PWR (4.15V) in series with the sensor (457K now) connected to ground. I'm using P2 AIO pin connected between the two resistors. I'm reading a value of 40 but I think this is too low. Am I using correctly the Ports library? The current code:
```
    //
    // Soil moisture sensor test
    //
```
#include Port soilSensor(2); int dSoilMoistureValue = 0; void setup () { Serial.begin(57600); Serial.println("\nsoil sensor test; soilSensor.mode(INPUT); } void loop () { dSoilMoistureValue = soilSensor.anaRead(); Serial.print(dSoilMoistureValue); Serial.println(); delay(1000); }
Posted 09/06/2010 21:57:38


	
		
			
								
					jcw

				
			
			
				
Looks ok to me - anaRead() is a thin wrapper around analogRead() from the Arduino/Wiring runtime.



				Posted 10/06/2010 02:00:26
			
		
	
	
		
			
								
					pedro

				
			
			
				
In fact, after some testing I realized that the resistance of my sensor is sometimes 450K and sometimes 2K...! I guess there is a bad contact somewhere...
@jeroen: Do you remember how much was the resistance of your sensor?



				Posted 10/06/2010 20:42:13
			
		
	
	
		
			
								
					pedro

				
			
			
				
Here are a couple of photos:
http://www.flickr.com/photos/pedronf/4689195414/
http://www.flickr.com/photos/pedronf/4689195356/
I guess the bad contact is because of the screw terminals I'm using... I'll remove them and solder the wires directly to the anodized nails.



				Posted 10/06/2010 21:00:00
			
		
	
	
		
			
								
					pedro

				
			
			
				
The previous post with the correct syntax...
Screw terminals
Inside mold
?!? Why the hell the photos are not visible?!?
(Because you're not linking directly to images... sir - so I've removed the !'s)



				Posted 10/06/2010 21:03:04
			
		
	
	
		
			
								
					jeroen

				
			
			
				
Did not remember, but I have two spare, so I tried a minute ago.
Dry (it was for weeks in the livingroom with +/-50% relavitive humidity)
60 MegaOhm, 
Wet (after 2 minutes in a glass of water) 50 KiloOhm.
regards Jeroen



				Posted 10/06/2010 21:18:33
			
		
	
	
		
			
								
					pedro

				
			
			
				
The sensor when dry behaves like an open circuit and after 5 minutes in water is at 3K ohm. I left it plugged to the JeeNode while drying out so I can have the resistance curve in time while drying out. I'll publish the results tomorrow.



				Posted 12/06/2010 20:49:54
			
		
	
	
		
			
								
					pedro

				
			
			
				
Some more photos and a picture of the data collected while the sensor was drying out, after 5 minutes in water. It started out at around 3K ohm and after 22 hours it was an open circuit. I didn't measure the temperature, it is pretty hot in my small lab when I have the PC on all the time.

nf/sets/72157624247303342/detail/>



				Posted 13/06/2010 18:39:40
			
		
	
	
		
			
								
					raalst

				
			
			
				
Hi, I'm not sure how you measure the resistance but it is worthwhile to measure once
with + on pin A and - on pin B. then do the next measurement in reverse
( - on A and + on B).
Why ? because DC corrodes metal very quickly in water.
when you reverse then the chemical processes are also reversed and corrosion is a lot less.



				Posted 13/06/2010 19:39:53
			
		
	
	
		
			
								
					pedro

				
			
			
				
@raalst: I measured the resistance with a multimeter for a couple of seconds. The voltage applied by the multimeter in ohm measurement is small I guess. For the normal working I'll be applying 3.3v to a resistor divider of a 10K resistor and this sensor, using a digital output, and then measure the voltage to the sensor using a ADC. Do you think in long term the metal will corrode, even if it is galvanized?



				Posted 13/06/2010 20:52:16
			
		
	
	
		
			
								
					pedro

				
			
			
				
I think I re-invented the diode!!! After raalst post I tried measuring the resistance both ways, that is, exchanging the multimeter connections. Guess what? I have 22K ohm in one direction and 470K ohm in another! Thats a huge difference! How is this possible?



				Posted 16/06/2010 19:16:58
			
		
	
	
		
			
								
					jeroen

				
			
			
				
I measure both ways the same value. So something is wrong.



				Posted 16/06/2010 20:34:09
			
		
	
	
		
			
								
					pedro

				
			
			
				
Yes, something is really wrong... maybe I tested for too long with one polarity and the metal / plaster combination chemically changed... or maybe while I was soldering the wires to the nails, too much eat on one of the nails...
Anyway I'm building a new one using copper wire instead of galvanized nails and smaller. I'll publish the results in a couple of days.



				Posted 16/06/2010 20:54:57
			
		
	
	
		
			
								
					tankslappa

				
			
			
				
That's quite curious, I would have thought with two identical bits of metal, no electrolysis would occur, you need dissimilar ones for that.
I would avoid copper, it will readily corrode, especially with the sulphur compounds rain water picks up on it's travels, and then you end up with the green copper we all know from building roofs and completely different resistive properties.
Personally I'd try with something a little less likely to react. Unfortunately gold is a tad expensive, so maybe try some stainless steel screws?



				Posted 16/06/2010 23:32:29
			
		
	
	
		
			
								
					pedro

				
			
			
				
Well, my testing showed me that this kind of sensor is not accurate at all. After powering the sensor with 3.3V it takes a while to stabilize the reading. So I can't do what I wanted, just power up and read the sensor each half hour.
Also after watering the soil, its resistance goes down and then up again after a short time. This makes very difficult to rely on the sensor to know when to water the soil. Unless it waits that the soil is really dry, but I don't want that.
So for now I'm not using the sensor, I'm using the RTC and watering for 90 seconds (with my pump this corresponds to 3 liters) each day at 21h00. I'll try other techniques of soil humidity sensor after my holidays. Maybe this one: 



				Posted 27/06/2010 22:48:01
			
		
	
	
		
			
								
					jcw

				
			
			
				
Another link - just pictures 



				Posted 02/07/2010 21:29:05