Remote flower watering and monitoring
ArticleCategory: [Choose a category, do not translate
AuthorImage:[Here we need a little image from you]
TranslationInfo:[Author + translation history. mailto: or
original in en Guido Socher
AboutTheAuthor:[A small biography about the author]
Guido likes Linux because it is a really good system to
develop your own hardware.
Abstract:[Here you write a little summary]
Every time I go on vacation I find myself investigating who could
look after the flowers in the house and the vegetables in the
This year I decided to built an electronic system.
Now now it is possible to water indoor potted plants and flower/vegetables in the garden
remotely over the internent.
With this system one can monitor temperature and soil humidity. It
gives then the possibility to water the plants remotely over
the internet. All this can be done from any PC or
mobile phone. The tuxgraphics Ethernet board with the appropriate
web server software makes it possible.
ArticleIllustration:[This is the title picture for your
ArticleBody:[The article body]
Not only a toy system
I did my first experiments with automatic flower watering
as a boy. I used cotton
strings to connect my plants to a bottle of water.
This system is very simple but works never because it does not adjust
to the demands of the plants. It is either always too wet or always
When I was older I experimented with Hydroponics.
Here the plants grow on artificial soil with a bit of fertilizer in the
water. It's not a very natural way of growing plants.
It makes a difference especially for edible plants.
Vegetables grown on real soil taste much
better and are probably much healthier.
We need a system to remotely water the plants rather than watering
them automatically. Sometimes you know from the weather forecast that it will
rain the next day. In other words you don't want to unnecessarily overwater
the plants. The system should be compatible with normal potted plants and garden plants
(no special soil).
Monitoring flowers and other plants
There are many sources of information that you want to use when
you monitor plants remotely. The temperature in the area next to the plants,
especially when plotted as a graph over time, can give you a fairly good feeling
when to water the plants.
Soil humidity is the other important component. If it is not too hot
then plants can grow in fairly dry soil. However when the sun shines straight
onto the plants then they need moist soil.
The weather forecast is also a very good source of information. You might
not need to water the plants if there is a rain or thunderstorm coming. There are
many people who apply too much fertilizer and then they apply too much
water to save the plants from burning in the sun on salty and fertilizer rich soil.
Grass can generally be mistreated this way but herbs, vegetables and flowers
need a more natural treatment to grow nicely.
I recommend to water the plants in the evening. The water can then soak
in during the night. If you water during the day then most of the water
will evaporate immediately especially on a sunny day.
How to measure soil humidity?
I have seen many soil humidity measurement systems that try to use just two
metal bars and measure the ohmic resistance between them. One can also buy
cheap humidity meters in garden centers and they use metal electrodes
with two different
metals. When the soil is wet then this will become a "battery" and they
measure the current that is generated.
Don't use metal electrodes. You will have to run a small current
between the metal bars even if you want to just measure the ohmic resistance. This will cause the metal to de-compose during the
electrolysis that was started here. Electrolysis can corrode almost any metal.
If we use stainless steal for the
metal bars then it will be even worse because we would poison the soil
with heavy metals.
Those heavy metals in the soil might be tolerated by the plants for a while
but we really don't what to have those in the vegetable garden.
A long 5mm graphite electrode
Pure graphite is a fairly good conductor, it does not corrode
in salty environments and it does not decompose under electrolysis.
I used two 5mm diameter graphite electrodes and a potted plant
for testing purposes.
The graphite electrodes in a potted plant
I observed the behavior over a couple of weeks with an ohmmeter and I was
fairly disappointed. The change in ohmic resistance depends to a fairly
small degree on the humidity of the soil. The ohmic resistance changes only by a few percent
when going from dry to wet soil. What made me finally discard this idea
was the observation that the ohmic resistance changed a lot when I added
a few drops of fertilizer to the water. In other words you really measure here
how salty or sour the soil is. You don't really measure humidity at all.
The graphite does not contaminate the soil but a DC current between the
graphite bars still causes an electrolysis. If there is e.g a bit of
NaCl (table salt) in the soil then you get chlorine gas on one electrode
and hydrogen gas on the other. The amount is very small. You can't smell
it but you can notice it on the ohm meter. The longer you measure the higher
the resistance becomes. The ohm meter starts to drift. This is because the tiny gas bubbles that
form along the electrodes insulate them a bit.
Although this experiment failed it was still very interesting. It means that
you can't really trust any equipment that uses two metal or graphite
electrodes for soil humidity measurement.
I started to investigate other ideas such as optical humidity measurements.
You might have noticed that paper becomes slightly transparent when wet and
my idea was that I could use a similar effect in other materials.
One day I had a pack of electrolytic capacitors in my hand and suddenly I
knew how to measure humidity. Those electrolytic capacitors consist essentially
of two insulated metal plates and in-between there is a salty liquid. In
poorly engineered computer motherboard you might have noticed that electrolytic
capacitors that are too close to a heat source such as the CPU leak over
time. If you then take one of those leaking capacitors out and measure them
then you will notice that the capacity values are off a lot. There is a huge
capacity difference between a new "wet" electrolytic capacitor and one that has dried
out. It should be possible to use this effect for soil humidity
Two pieces of PCB insulated and painted with clear coat.
I took two unused sheets of PCB (2.5cm x 10cm [about 1inch x 4inch]) and
sprayed them with a clear coat paint can. Normal transparent car paint. Any paint spray for cars will
probably work. This gives a nice thin and well insulating film of paint.
If you can't find any unused PCBs then you can take as well some sheet
metal. It should be brass, iron or copper. Something that is solderable
because you need to connect wires to it.
You can cut e.g two sheets out of a vegetable oil can and round off any sharp
with a bit of sand paper.
Solder a wire to each of the sheets and then spray the sheets with a paint
can. Take care to not hurt yourself when cutting the sheets out of the
vegetable oil can. Wear good gloves.
Dig a little hole in the soil and insert the two metal sheets. They should
be about 1cm to 2cm apart. Fill back the soil and compress it just a little bit.
The capacitor is now ready. It will have a capacity of around 5-10nF when
the soil is dry and up to 20nF when it is wet.
There is a little bit of a dependency on the concentration of salts and
minerals in the soil. The capacity seems also to depend a
bit on the temperature. These disturbances are however small
and a change in the soil's humidity changes the capacity values
As we do not run any permanent DC current through the
soil we will not decompose the minerals.
I had also the idea that one could not put soil between the electrodes
but rather some sponge or membrane like material such that only the water
from the surrounding soil gets between the electrodes and not any "dirt".
The above design with two simple insulated electrodes in the soil is however
working very well and it is robust enough for outdoor use.
The capacitive humidity sensor in action
All chemical processes such as plant growth depend on temperature.
A rule of thumb is that the "speed" of a chemical reaction doubles when
the temperature increases by 10'C. When it is warmer the plants will need
more water because they grow faster and more water will evaporate.
Over time one gets a good feeling for the amount of water that will
be needed just by looking at the temperature curve.
We use here the DS18S20 sensor. It is a calibrated digital temperature sensor.
The electric watering can
The most obvious solution is to attach a small electric water pump to a relay.
When it is time to water the plants then the relay controlled by the
tuxgraphics Ethernet board will close its contacts and provide
power to the pump.
The pump can be connected to a small canister or even a rain barrel.
Note that many water pumps require the water to flow freely into
the pump and then they can pump it up to a higher level. In other
words you should connect any pump to the bottom of the rain barrel.
Pumps will not totally block the water flow when they are off. A lawn
sprinkler connected to the pump will therefore need to be placed higher
up on a post. At a level that is equivalent to the top of the rain barrel.
For indoor plants and all potted plants I recommend to use a small electrical
To get such a pump was more difficult to get than I thought. There is no measure for
the quality and efficiency of such pumps. You have generally no idea what
power consumption it has until you really test it.
My first idea was to use a car windscreen washer pump. Those pumps are cheap and they can
be found everywhere. I was very disappointed when I tried different
models. It seems the car industry had never a need for a clever and efficient
design. All of them consume at least 5 to 10 ampere at 12V before they even
move one drop of water. What a waste of energy. They are unusable for our
purpose because it has a knock-on effect on all other parts. You need a stronger
power supply, a relay rated for 20A or more and appropriate shielding to protect
the sensitive electronics and the Ethernet magnetics from the electric and
magnetic fields that you generate when you switch the pump on or off.
After searching for quite a while I found two online shops that sell
suitable water pumps.
- http://www.lemo-solar.de/, a German
online shop. Sells only in europe and the web site is only in German but they
have a very interesting "Zahnradpumpe 12 Volt". It works already at 6V and
consumes less than 100mA.
siliconsolar.com is an online shop in the US and they sell world wide. This
replacement 6V solar pump is a submersible pump. You stick it directly into
a small water tank.
As you can see a good pump needs less than 100mA at 6V. This does of course
not generate a strong water jet but we don't want to excavate our
I ordered from both shops but I used for the potted plant watering system the
pump from siliconsolar.com.
The submersible water pump from siliconsolar.com I attached a short
plastic tube to the hose and I bended it like a little hook at the end.
This little pump goes then into a 3L plastic container and the
electric watering can is ready. 3L of water quite sufficient for a single potted plant.
It will be enough for several weeks.
For my first tests I used a small palm tree. Here is my
"test plant" with the electric watering can.
This electric watering can is made for a single plant. If you want to water
two plants then you use a T-tube connector and run from one pump hoses to
both plants. This solar water pump needs very little energy but it is
als not very strong. To water more than two plants you would need more pumps.
When I am absent during vacation then I will rather put most of the plants outside
where I have the lawn sprinkler
attached to the above magnetic water valve. I can then water the potted plants
at the same time as the vegetables in the garden. You just have to be
careful where you put the plants outside. Some indoor plants are not
used to direct sun light. They grow better in shade.
Lawn watering and electric sprinkler valves
Pumps are however not the only solution. All washing machines and dish washers
have a magnetic water valve. It opens when you connect it to straight line power
and otherwise it is closed. In most cases this valve is somewhere inside the
machine. Some european Miele dish washers have an external valve. It can
be connected directly to the water tap. It is easy to take it off an old machine
on a scrap yard or when somebody puts an old one on the curb for the city
to collect it.
Such a valve allows you to water a larger area such
as a vegetable garden. Note that washing machines and dish washers are build for
indoor use. Make sure you protect this valve as well as the relay to control it
from direct exposure to rain otherwise you risk to get an electrical
A magnetic water valve taken from an old dish washer. It connects
to the water tap and you can attach a garden hose to it.
Valves from dish washers and washing machines make a nice hobby solution but
the have the big problem that they require full line power (120V or 230V) and
that means you can't use them safely outdoors. If you want to build a full
professional garden watering system then look for 24V electric sprinkler valves.
They come in various forms sometimes even in little boxes to hide them in
ground at the corner of your lawn.
Different 24-volt lawn sprinkler valves.
The advantage of those sprinkler valves is that they run with 24V which
is more or less save for outdoor use.
The ethernet board with a relay to control the valve can then be located inside
the house. Some generic laptop power supplies can be used as an inexpensive
for the sprinkler valve. You would then
run the cable from the sprinkler valve into the house and connect
it there via the relay to the laptop power supply brick and to the tuxgraphics
ethernet board. After that you are
ready to water your lawn over the internet.
The sprinkler valves have a solenoid (electromagnet) that can work with
both AC and DC voltages. Most control equipment uses 24V AC. If you use
DC then the valves work with much lower voltages. 16V to 20V DC will normally be
enough for reliable operation of the sprinkler valve. A sprinkler valves consumes about 0.4A to 0.6A
current when powered with 16V to 20V DC.
An Orbit sprinkler valve connected to a graden hose for testing
The other side (water output) is just left open.
Power-on and water flows. The sprinkler valve works. We can now install it properly.
We have now discussed external components such as sensors and pumps. How do we
connect all this to the tuxgraphics Ethernet board?
To use the capacitive humidity sensor we need to generate a tone and then run
it through a capacitive voltage divider. One of the capacitors in this voltage
divider is the humidity sensor. All we need to do then is measure the voltage
after the divider and it is somehow proportional to the soil humidity.
The only problem is that we don't have here a DC voltage. It has to be an
alternating voltage (a tone) and the microcontroller can not measure AC
voltages. The solution to this is to add a precision rectifier circuit
and convert AC to DC. This is what the circuit in the upper left part of
the diagram does. The rest of the circuit is just the standard tuxgraphics
Ethernet board. To simplify the drawing I have omitted most of the
components on the Ethernet board.
Click on the diagram to open it as high resolution PDF file.
The amplifier circuit looks bigger than it is. All this fits onto the dot matrix
field of the tuxgraphics Ethernet board.
The whole circuit is very compact and fits easily into a little plastic box. You just attach the sensors and the
watering can via a 6V relay. I recommend to not try to design this water
proof. It is better to have air circulating freely. The best place for the
circuit is in the house or in a small garden shed.
Switching on/off an electromagnet
Switching on and off a devices that contains a large coil causes
high voltages to be induced. The sprinkler valves contains large coil.
If we just switch it on and off with the relay then the contracts
on the relay will wear off quickly, you will hear a click on
the radio and it can cause interference with the electronics on the
ethernet board. It can even destroy the insulation around the wire
of the coil.
To prevent this one needs to add a suppressor diode known as
flyback diode. If you power the coil with AC voltage instead
of DC then you can insert a resistor instead of the
diode. The best solution is however a diode and DC current as it
is much more efficient. The below circuit diagram shows how
to insert the diode.
A flyback diode is used to eliminate the sudden voltage spike across an
Flower and lawn watering over the internet
The remote watering system is completely web based. You can access it from
any desktop PC as well as any mobile phone with integrated web browser.
The tuxgraphics Ethernet board provides a web server. To access it remotely
over the internet you will need a DSL router that allows for port
forwarding. Most DSL routers can do that. To know where your remote watering
system is you will either need a static IP address or a DSL router that
can interwork with dyndns.org. A static IP is the best and
most reliable solution. Some DSL service providers offer static IP
addresses for a small extra fee.
Here are some screen-shots of the remote watering system.
This shows the main window. From here you can get to all other pages
and you can see current sensor readings. We don't measure the humidity
in absolute calibrated units. Humidity is displayed as an 8bit value
(0-255). After observing the values for a couple of days you will
be able to give a meaning to those values and decide what is dry, normal and
wet. Those ranges can then be entered in the "levels" dialog screen.
You can adjust the humidity levels. The red LED on the Ethernet board
will start to blink when the soil humidity goes below dry.
The water control dialog. For a potted plant you would specify
in the order of a few seconds. To water the lawn you can enter large
values like 1200s (=20min).
You can see several days of history in graphical form. A trend can be
The most recent reading is shown on top.
The graphs show a history of 10 days.
The software for the tuxgraphics avr ethernet board is called
eth_rem_flowercare. You can download the software at the end of
this article. To load it you need an avr-gcc compiler and a programmer
like e.g the avrusb500.
On the road: flower and lawn watering with the phone
Here is how you can take care of your flowers while you are on the road. It
works even with a basic data phone:
... of course it works also on the iphone:
It's really good to be able to water the garden where ever I am.
It's even fun to sit in the living room and turn the lawn sprinkler
on with the mobile phone.