New Project guidance (fluid control with multiple valves)

Projects General Guidance
1

Hi,

Newb at forum here, thanks in advance for reading (and sorry for my english)
My doubt is not mainly about arduino but is semi related and I believe it can be interesting (I didnt find anything similar)

Ive seen several project related to water (fluid) manipulation in order to water a garden composed by several plants, these project can be classified into 2 main types :
-drip irrigation (not sure if im saying it correctly)
-Electrovalv + water tank + pumps : the tank is filled by a electro valve, from the tank there are pumps that push the water to their destination

Both of them are quite interesenting but also very limited for what I have in mind

I have a gallery (balcony ? )with a shape similar to a L , with 10-15 metres per band, what I want is fill the gallery with different plants (maybe 7-15 plants per band) and build a nice garden :slight_smile: . The target is clear, I need to water the plants in a proper way every plant individually (Im not looking for mass watering all the plants at the same time because they have different needs, different positions and different exposure to sun). The idea is to share the water infraestructure creating a distributed system with multiple valves and also quite a lot of sensors (light, humidity, temp, etc)

In order to get a proper idea I think something more graphic would be nice (the --- are the tubes from pumps to valves and plants ) :
tap ---- evalve --- tank --- pump --- Valve at T1 ----- Valve at T2 ----- Valve at T3

| | |
Plant1 Plant2 Plant3

The idea is that I would be able to run the pump for water any plant individually, for example if I want to water plant number 2 I would open the multi valve T1 at the entrance and the exit (without opening the exit to plant1) and then open the valve at T2 at the entrance but without opening the exit to the right and opening the exit to the plant 2

Not sure if the scheme is clear enough :S , the alternative to that is placing quite a lot of pumps and tubes from the pumps to plants.

My question is , what im planning is viable ? is reasonable in terms of fluid control ? I a system where you want to control fluids that is the proper way with distributed valves ? I need some guidance and advice, Im not even sure if valves with 3 ways out are used that way or if other more apropiate ways to control water are used in order to achieve a similar result (using other methods o other devices)

Just to clarify I know that the whole idea is quite crazy and weird, but I think we are here because we like crazy and weird things :slight_smile:

Regards and thanks for reading any advice would be apreciated

2

flecha:
My question is , what im planning is viable ? is reasonable in terms of fluid control ? I a system where you want to control fluids that is the proper way with distributed valves ? I need some guidance and advice, Im not even sure if valves with 3 ways out are used that way or if other more apropiate ways to control water are used in order to achieve a similar result (using other methods o other devices)

Controlling pumps and valves by an Arduino is no problem. But you need shields/modules for driving such devices, the output pins do not supply enough power. Separate power supplies are suggested for the Arduino and the other devices. It's a good idea to select a pump and valves of the same operating voltage, so that a single power supply is sufficient for these devices.

Controlling the amount of water spent may become tricky. Without individual feedback from each valve you better open only one valve at any time, and find out how much water the pump will supply in what time. Feedback from soil humidity sensors is not very reliable, in case you want to make your system more clever.

3

DrDiettrich:
Controlling pumps and valves by an Arduino is no problem. But you need shields/modules for driving such devices, the output pins do not supply enough power. Separate power supplies are suggested for the Arduino and the other devices. It's a good idea to select a pump and valves of the same operating voltage, so that a single power supply is sufficient for these devices.

Controlling the amount of water spent may become tricky. Without individual feedback from each valve you better open only one valve at any time, and find out how much water the pump will supply in what time. Feedback from soil humidity sensors is not very reliable, in case you want to make your system more clever.

I was thinking about using a relay (no shields) in order to control the valves, however Im not worried about that (dont misunderstand me, thats important but I know ho to approach that matter)

Agree with power supply.

About the amount of water I thought about time control, not sure if I would ever be able to control the amount of water thrown into the plants.

I thouth soil humidity sensores were reliable, so what if I want to know how often must I push water into the plants ? Should I make an aproximation and just release the water for a show determined time lease ?

And what about the valves distribution ? Is my idea reasonable ?

Thanks for the answer :slight_smile:

4

Relays are okay, valves distribution is okay as long as the pump will not be damaged when all valves are closed. What exactly bothers you with the valves distribution?

Most (cheap) soil humidity sensors measure conductivity, which depends on the amount of ions (salt, fertilizers...) in the soil, which are hard to control. Check yourself what other sensors are available.

5

If you have mains water you do not need a pump.

One valve is all you need , individual requirements can be met by adjusting the drip rate at each plant.

6

Boardburner2:
If you have mains water you do not need a pump.

One valve is all you need , individual requirements can be met by adjusting the drip rate at each plant.

As far as I understand what you say is to push up the water tank (potential energy) and just let a valve open and close in order to allow the water flow to the plants, what I dont get is how do you adjust the drip for each plant ?

7

If the valves are small and easy to turn maybe they could be moved by servos. Or maybe you could use servos to pinch a piece of soft plastic tube - no need for a valve.

...R

8

Robin2:
If the valves are small and easy to turn maybe they could be moved by servos. Or maybe you could use servos to pinch a piece of soft plastic tube - no need for a valve.

...R

I like the idea but problem is valves are 5 € (the cheapest chinese valves I found), servos are more expensive

9

flecha:
, what I dont get is how do you adjust the drip for each plant ?

You use an adjustable dripper.
however they are more expensive than the $5 valves you seem to have found.

You can make the same by clamping a piece of silicone tube with a screw and clothes peg.

Also check the operating pressure of those valves they may require high pressure to operate.

10

flecha:
I like the idea but problem is valves are 5 € (the cheapest chinese valves I found), servos are more expensive

You should be able to get servos for a similar price and you would not need relays or control boards.

...R

11

the idea that you have a single pump
if plant 7 needs water
open valve7
start pump

you can have lots and lots of valves.

if your solenoid valves are small enough, it would be easier to make a board with a lot of FET's on it and control them.
if you want to use relays, that is fine, the only draw back is the physical size. they are cheap, easy to use, good way to get started.

with your quantity, I would also recommend you learn about shift registers.
the 595 has lots of examples and tutorials. this will allow you to use 4 pins to control 8 outputs.
the beauty is you can add a second one, use the same 4 pins and control 16 or 24 outputs. in multiples of 8

one of the most expensive parts will be the solenoids. they do not make tiny ones and they typically cost around $10 each. this will be the largest investment.

if you want to tinker, consider how a peristaltic pump works. it uses a roller to seal the tube, not crush, jus seal.
imagine a cam shaft with 8 lobes. you could rotate that with one stepper and each lobe could be placed to allow one pipe to allow water to run. pump off, set to the correct zone, pump on, time out.

I am thinking of a series of wooden (plastic/metal) discs with a relief all on a shaft.
a pushrod assembly so each disc pushes on one rod,
the business end of the rod pushed on the tubing.

the shaft rotates to position and only when in position does the push rod release and allow water to flow.
it could be quite compact and control lots more than 16 drip hoses. only one at a time.

12

flecha:
As far as I understand what you say is to push up the water tank (potential energy) and just let a valve open and close in order to allow the water flow to the plants, what I dont get is how do you adjust the drip for each plant ?

first off, if you have 1 or 100 plants, you can water from one pump. that means one water flow sensor.
you can monitor gph, not gpm as I suspect you will use low flow rates as you said high density of plants and not large plants.
one flow sensor, one pump and then you need to select how you disperse the fluid.
if you look at drip irrigation for gardens you will find small hose and small drip heads.
flow can be as simple are pressure on a calibrated orifice.
take a hose with a cap.
heat a pin and 'drill' a hole into the cap.
turn on the pump and time how long it takes to fill a cut or a quart.
calibrated !

put in two holes, repeat. now you have a higher volume water unit.

one key here is that the longer the hole the greater the resistance and therefore the better the repeatability.

13

flecha:
About the amount of water I thought about time control, not sure if I would ever be able to control the amount of water thrown into the plants.

calibrated orifice over time = volume ,. you can get pretty accurate
there are a few ways to sense moisture in the soil. conductivity will do for general plant watering.
I would use the calibrated finger. use some simple plants, use the sensors you want, the cheap ones use lead and will leach lead into your plants. fine for ornamentals, but not for anything you plant to consume such as herbs or even aromatic plants you burn.
capacitive sensors are available, but costly. I think they are only costly because there are so few suppliers.
capacitive are totally sealed, so better for any food plants.
I would put in half a dozen plants and set each sensor for some value such as 2.50 volts
let it run for a day, then make up one portable and test each plant soil to see the variation. with not too much problem you should be able to get a crude calibration for each plant to get within the range the plant likes for growth.

14

dave-in-nj:
first off, if you have 1 or 100 plants, you can water from one pump. that means one water flow sensor.
you can monitor gph, not gpm as I suspect you will use low flow rates as you said high density of plants and not large plants.
one flow sensor, one pump and then you need to select how you disperse the fluid.
if you look at drip irrigation for gardens you will find small hose and small drip heads.
flow can be as simple are pressure on a calibrated orifice.
take a hose with a cap.
heat a pin and 'drill' a hole into the cap.
turn on the pump and time how long it takes to fill a cut or a quart.
calibrated !

Yes
Largely the problem is mechanical or physics related.
It can be done with expensive electromechanical valves if you pocket allows.
Cheapest is to use existing tubes etc. and use simple on off .
At low pressure from a water butt expect the valve to be expensive.

15

Wow I read the posts and Im impressed, good ideas.

Thanks a lot guys, I will post a proper answer later the way you deserve

16

for my drip irrigation, I had to add a filter. the orfice's are tiny and get clogged all too easy.

I just went onto aliexpress. DC pumps are less than $2 each. tiny submersibles.

it will take the same control logic to drive a pump as it will to drive a solenoid.
pumps look like they can take tubing direct. some of the solenoids I had looked at will require pipe fittings. an added expense.

you can run water to a bucket and put in 8 pumps and then water 8 plants.or 10, or 12....
if you have multiple buckets, you could tie them all together so one fill valve would handle them all at once.

17

Just to clarify, yes the plants are small plants, not for eating just ornamental.

I have done a little schematic for the two main ideas you guys gave me, I think they may be interesting for others with similar projects

The first variant, a multipump system, without valves, push waters from multiple pumps connected to individual PVC pipes, or also better, the system you sugested using just one pump which is moved between multiples PVC pipes, I realize is the same concept but improved in some way (also makes the whole thing harder to implement)

Thats my main and original idea, lightly modified in order to use a lower amount of valves

Both ideas are quite good for what I want to do in the project and have a lot of merit, one uses a lot of valves and shares the PVC pipe, I think that one is more expensive because the valves. On the other hand the first idea seems to be quite good, more advanced, with less expensive parts, with less electronics for manager, but also introduces a LOT of PVC pipes, honestly I think the no-valves system is better, but the idea of using lots of pipes is worrying me.

DrDiettrich:
Relays are okay, valves distribution is okay as long as the pump will not be damaged when all valves are closed. What exactly bothers you with the valves distribution?

Most (cheap) soil humidity sensors measure conductivity, which depends on the amount of ions (salt, fertilizers...) in the soil, which are hard to control. Check yourself what other sensors are available.

Not sure if the pump is going to be damaged, but I think I can handle that opening the final valve BEFORE turning the pump on. Definityvately I will have to make more tests with humidity sensors, you are right Im using the cheap ones.

[/quote]

Robin2:
You should be able to get servos for a similar price and you would not need relays or control boards.

...R

Nice :slight_smile:
You mean classic valves with servos pushing it on and off, what you say its a point to take into consideration

[/quote]

dave-in-nj:
you can have lots and lots of valves.

if your solenoid valves are small enough, it would be easier to make a board with a lot of FET's on it and control them.
if you want to use relays, that is fine, the only draw back is the physical size. they are cheap, easy to use, good way to get started.

with your quantity, I would also recommend you learn about shift registers.
the 595 has lots of examples and tutorials. this will allow you to use 4 pins to control 8 outputs.
the beauty is you can add a second one, use the same 4 pins and control 16 or 24 outputs. in multiples of 8

one of the most expensive parts will be the solenoids. they do not make tiny ones and they typically cost around $10 each. this will be the largest investment.

if you want to tinker, consider how a peristaltic pump works. it uses a roller to seal the tube, not crush, jus seal.
imagine a cam shaft with 8 lobes. you could rotate that with one stepper and each lobe could be placed to allow one pipe to allow water to run. pump off, set to the correct zone, pump on, time out.

I am thinking of a series of wooden (plastic/metal) discs with a relief all on a shaft.
a pushrod assembly so each disc pushes on one rod,
the business end of the rod pushed on the tubing.

the shaft rotates to position and only when in position does the push rod release and allow water to flow.
it could be quite compact and control lots more than 16 drip hoses. only one at a time.

First, I dont know what a FET is, about registers I understand that going to multi valve system is going to need some "auxiliar device o devices" go manage the valves.

I really like the idea of a single movable pump fitting into different PVC pipes (thats what I understand from your explanation, please correct me if Im wrong), however at the moment Im not even sure if Im able to build such a complex device, I should need some "base" in order to fix a pump and guide the pump between multiple pipes, maybe with a servo or a stepper. About the amount of water used I think is interesting to measure, but I would like to fit that with moisture sensors (cheap ones conductual), I understand what you say, and thanks for the info about leach, didnt think about that.

One thing I will need to control is the "pressure", I didnt take into consideration, I bought some parts in order to do tests :
Valves : Valve (amazon) 5€

Rated voltage: DC12V
Coil Resistance : 37x3A9 ; À 0.25x3A9 ; ( 20 x2103 ; )
Switch Type : CC continues
Operating pressure : 0.02Mpa - 0.8Mpa
Average temperature: 1-75 ? C
Response Time: The x2264 ; 0.15 seconds , versus x2264 ; 0.3 seconds
Electric strength : Withstand voltage between AC 2500V
The isolation of the property: The insulation resistance between conductor and non-conductor is over 100M
Download Feature : 0.02Mpa x2265 ; 3 L / min , 0.1 MPa x2265 ; 12 L / min , 0.8Mpa x2265 ; 35 L / min
Working life: x2265 ; 0.5 million times
Thread diameter inlet and outlet : 4/5 " (20 mm )
Input and output angle : 180 degrees

Pumps : Pump (ebay)

Hmax : 40-220cm
Qmax : 100-350L/H
Power Connector : 3pin connector
Voltage : DC3.5V-12V
Power : 0.5-5W
Noise : 30DB
Working life : 50,000 hours
Out of the faucets : 9mm
Water injection plug : 10mm
Dimensions : 37 x 47 x 40mm / 1.46" x 1.85" x 1.57"
18

Your second diagram is the more typical one.

Variable drip feeds cost about £$10.

If you get the old type pegs the gypsies used to make and a bit of fuel tube from the model shop they cost cents.

Pressure control is not needed with a header tank.

Your valve consumes 250 ma though so its not a battery application.

Commercial ones use motor driven ball valves that only consume power when they are operating.

So if you want battery power then a plumbing ball lever valve and servo is the way to go.

Electric aerials or door locking car actuators are handy for this.

19
20

if you look at a drip irrigation system, about 25 m of 1/4 OD hose and lots of fittings.
even a hose adapter.

there are lots of kits available for these that will offer more options.

you can put about 10 of these in a bundle the size of a large garden hose.

if you look at the underground watering system, they run the large runs in hard plastic, but the end sections in a softer plastic. you can buy a puncture kit and 'plug' in fittings to allow you to have a small line run to one plant.

one idea is to make a planter. a unit that can hold plants, i would have one pipe connection and then multiple output pipes.
you could put your electronics, solenoids, valves, pumps, etc, into that and then have short runs of pipes come from that.
the top half would be a planter, the bottom your electronical box.