Smart Sprinklers?

[JDuc]

I appreciate all of the replies! I guess that this is a topic that is quite intriguing for some based on the replies I see here.

What exactly do you mean by "only run when its needed"?

Are you trying to not have it run when its raining?

Do you only want it to run when the ground is at a certain moisture level?

Ideally, I don't want to be watering when it's not necessary. Not watering when it's raining is easy. Rain sensors are a dime a dozen.

This could be done by zone I would imagine.

I have been trying to find a ET or moisture based system that can be integrated into Elk. I really hope someone would build a self-contained (read low-cost) ET system that uses serial communications to provide input information to HA controller. I have yet to find such a product. The closest I can find is a Hunter SolarSync or Hunter ET-sensor. From my research, the ET-based systems produce a “good-enough” calculation with the added benefit of not relying on a single in-ground sensor. In addition, the in-ground sensors have to be calibrated.

As an alternative, I believe it would be feasible to integrate either of these products into a Rain8Pro. These products use moisture sensors which mayproduce unstable readings depending on the quality of moisture sensor.

You bring up a few points that make me realize I'm still needing to do some research. I'm guessing that there's a few differences with the ET systems and what you're pointing out with the Rain8Pro?

FWIW, I've been running a ET based water balance system starting earlier this year and found it to work quite well.

It turns out that since you water a whole zone at a time anyway, then you just have to model each zone based on the most stringent conditions within the zone to make sure that everything within the zone get enough water. Some parts of the zone may get over-watered, but that issue exists within a zone even if you schedule manually.

This is what I was thinking would make the most sense. I was under the impression that ET type systems might utilize moisture sensors as well as current, past, and future weather conditions to determine if you needed to water. My thought was to keep my current zones and just install moisture sensors in those zones in the ideal locations. Only water specific zones as necessary based on conditions. I would want to tie this into the Elk system in order to tie everything into one central location.

The CQC irrigation scheduler which I use and wrote does manage each zone independently. You can specify the parameters for each zone including soil type, vegetation type, sprinkler flow rate, sun exposure, slope, etc. The water balance is maintained for each zone based on rainfall received, weather forecasts, and evapotranspiraton. So far, it has worked well.

This sounds like it would be ideal for what I was envisioning. Sounds like you're already all over it though! I suppose I just need to figure out what hardware I can use for this set up.

Again, I really do appreciate all of the discussion on this topic!

 

[jkish]

Sounds like a nice system. I think I would get overwhelmed trying to account for the fact that you can't just look at things like sunny, windy, but which direction the wind is comming from, if it is sunny in the morning/afternoon only, etc. as residential yards are so full of obstacle to sun and wind from one direction but not another. Maybe all of those details just get averaged out in the end.

I think it is cool that you have it working.

At some point, you can't model every detail, but as you say, much of the variation tends to get averaged out.

In case it doesn't, most of the settings of the system are configurable so you can tweak zones based on experience. For example, by default the system waters when the water balance of a zone reaches 50% and waters long enough to bring it up to 100%. If you find that a zone tends to get too dry to wait for a 50% water balance, you can change that zone to water at 55 or 60 percent to compensate.

 

[jkish]

This is what I was thinking would make the most sense. I was under the impression that ET type systems might utilize moisture sensors as well as current, past, and future weather conditions to determine if you needed to water. My thought was to keep my current zones and just install moisture sensors in those zones in the ideal locations. Only water specific zones as necessary based on conditions. I would want to tie this into the Elk system in order to tie everything into one central location.

You don't have to use individual moisture sensors, just a central rainfall gauge.

The components of my system are a CQC server with a Weather Channel driver, a Rain8Net Pro irrigation controller and driver, an Irrigation Scheduler driver, and a Davis Vantage Vue weather station and driver. As I mentioned before, you can substitute another controller as long as there is a driver for it. The Elk would work with relays attached to control the valves.

 

[JDuc]

You don't have to use individual moisture sensors, just a central rainfall gauge.

The components of my system are a CQC server with a Weather Channel driver, a Rain8Net Pro irrigation controller and driver, an Irrigation Scheduler driver, and a Davis Vantage Vue weather station and driver. As I mentioned before, you can substitute another controller as long as there is a driver for it. The Elk would work with relays attached to control the valves.

Perhaps I'm interchanging 'moisture sensor' and 'rain gauge' where they shouldn't be?

I'm seeing 'moisture sensors' as being inserted into the ground and utilizing their readings to determine if a zone needs watering. These sensors would be inserted into a location within a zone that is deemed the best location based on ground conditions, sun exposure, etc. While a rain gauge would, as the name implies, determine how much rain you have received or are currently getting.

The way I see it, this would negate unnecessary watering of zones that are already at adequate moisture levels.

Am I over thinking this?

 

[jkish]

Perhaps I'm interchanging 'moisture sensor' and 'rain gauge' where they shouldn't be?

I'm seeing 'moisture sensors' as being inserted into the ground and utilizing their readings to determine if a zone needs watering. These sensors would be inserted into a location within a zone that is deemed the best location based on ground conditions, sun exposure, etc. While a rain gauge would, as the name implies, determine how much rain you have received or are currently getting.

The way I see it, this would negate unnecessary watering of zones that are already at adequate moisture levels.

Am I over thinking this?

No, that is one way to do it.

For that type of system, there probably wouldn't be any ET calculations required although you could do that and then use the moisture sensors as a override.

The way the water balance / ET systems work is the amount of water leaving the soil is calculated from Evapotranspiration and the amount of water entering the soil is calculated from a rain gauge. The total amount of moisture that the soil can hold is based on your soil type and vegetation type. You start off with a guess at the current balance and then the hourly ET calculations, based on the weather, subtract from the balance and any rainfall add to the balance. When the balance for a zone falls below a certain percentage, say 50%, then the irrigation system runs to bring it back up to 100%. The zones are each characterized by soil type, vegetation type, sun exposure, slope, etc, so each zone only waters when it is time based on it's balance.

I think there are less things to go wrong with the WB/ET system as I have heard about a lot of problems with moisture sensors being accurate/reliable.

 

[JDuc]

No, that is one way to do it.

For that type of system, there probably wouldn't be any ET calculations required although you could do that and then use the moisture sensors as a override.

The way the water balance / ET systems work is the amount of water leaving the soil is calculated from Evapotranspiration and the amount of water entering the soil is calculated from a rain gauge. The total amount of moisture that the soil can hold is based on your soil type and vegetation type. You start off with a guess at the current balance and then the hourly ET calculations, based on the weather, subtract from the balance and any rainfall add to the balance. When the balance for a zone falls below a certain percentage, say 50%, then the irrigation system runs to bring it back up to 100%. The zones are each characterized by soil type, vegetation type, sun exposure, slope, etc, so each zone only waters when it is time based on it's balance.

I think there are less things to go wrong with the WB/ET system as I have heard about a lot of problems with moisture sensors being accurate/reliable.

Thanks for brining it all together for me. This makes complete sense now. Of course, I would assume that there are provisions in the software to prohibit watering between certain times of the day, no matter the temperature, and instead to defer the watering until after said time of day? I would guess that most of the time, the conditions during that time would prohibit watering at those times anyways, but I sure would like to avoid those pesky fines for watering in the middle of the day during the summer months while under watering restrictions.

Now, I need to get a parts list together...not sure on that one just yet.

 

[jkish]

Thanks for brining it all together for me. This makes complete sense now. Of course, I would assume that there are provisions in the software to prohibit watering between certain times of the day, no matter the temperature, and instead to defer the watering until after said time of day? I would guess that most of the time, the conditions during that time would prohibit watering at those times anyways, but I sure would like to avoid those pesky fines for watering in the middle of the day during the summer months while under watering restrictions.

Now, I need to get a parts list together...not sure on that one just yet.

Yes, you can schedule the watering, if needed based on the water balance, for a certain time of day and certain days of the week. There is even/odd, certain # of days, and certain days of the week scheduling available. There are also weather rules that can be enabled that prevent watering if it is raining, if the chance of rain is greater than a certain percentage over the next # of days, if the temperature is too low, if the winds are too high, etc.

You can control a certain zone based on a fixed schedule instead of via water balance if desired - maybe a vegetable garden needs to be watered every day no matter what.

 

[Michael McSharry]

Some additional clarification. Lou is describing a system that is primarily based upon time and can run totally within the ELK. jkish is describing a system that runs on a PC as an extension to CQC. Note also that the Penman Moth algorithm is based upon use of Solar Radiation which is not a generally available parameter unless you have an onsite Davis or equivalent weather station. mcsSprinklers also does ET (and/or time and/or moisture) scheduling with the ET based upon data available from the internet (or local weather station). It runs as its own application or as an extension to Homeseer. It is also now running in a Linux VM and hopefully soon as an appliance in a ARM-based controller. http://mcsSprinklers.com/mcsSprinklers.pdf

The effectiveness of ET scheduling is based upon the variation of climate. If you are in the desert then there is very little weather variability so time-based scheduling will be as efficient as ET-based. There is a mcsSprinklers chart at http://board.homeseer.com/showthread.php?t=140625 showing the weekly ET variation and from it you can see how wastefull a time-based schedule would be for this area in the Pacific Northwest.

What other posters in this thread have indicated and also what I think you also have concluded is that you really need to determine what your needs and expectations are as there are many different approaches. The Rain8Pro and typical consumer irrigation controllers approach to using a mositure/rain sensor is to cut off the 24V from the valves. This is typically not effective with smarter systems since the intelligence will not know when the valve has power cutoff unless you also route the sensor status to another input. If you do this then no need to cut off the 24V. The moisture/rain sensor can be effective when there is small ET variation and enought rainfall during the irrigation season to make a difference.

Advice pointing you away from moisture probes is also good. It is expensive to install an effecive probe network and even then there can be variations in the flow patterns under ground. One mcsSprinklers user benchmarked a tensiometer-based set of moisture probes vs. the mcsSprinklers ET scheduling algorithm and found the two track very closely. He eventually reverted to the ET schedule because of its simplicity and lower maintenance. This past year I benchmarked the Vegitronix moisture probes and found that they worked quite well for tracking soil moisture in desert clay, but were useless in rich soil in the Northern climate. I would consider this moisture probe to be very good to schedule irrigation in desert environment where foilage is sparse and the sun intensity changes over the months.

Lou is also correct in that experience with your plants goes a long way to tuning the irrigation control. The software-based controllers will provide considerable configurability, but it is still up to you to learn what each part of your area actually needs. When you lay out your irrigation system you have more to consider than just water pressure and valve density if you want to optimize water utilization. You want to have valves under similiar control based upon the environment and soil and this may not always be physical proximity. In my case I have 16 valves which covers about 10,000 sqft and this allows water balance thresholds and zone characteristics to be adjusted appropriately. To further illustrate consider the shade variation cast by your house in your area based upon the sun angle as you progress from Spring to Fall.

 

[swaggy]

Hi Michael

Thanks for the informative response!

Where can I find instructions on how to run mcsSprinklers on a Linux VM? All the references I found were for standalone Windows or Windows VM installations.

Thanks

 

[Michael McSharry]

The Linux VM is only an intermediate development step toward mcsSprinklers being part of an embedded device such as one of the Marvel plug computers. I disclosed it just to indicate the state of development. I'm working with the $25 Dockstar now as a target test bed. mcsSprinklers will not be made available as software-only for execution in Linux or Apple environments. It is just too much of a support nightmare.

 

[JimS]

MCSsprinklers looks interesting so not saying anything against it... But if you are interested in rolling your own for Linux (or for other platforms for that matter) you can find plenty of info and others doing similar things. I have been thinking of doing something like this for some time. Here are a few links:

http://members.cox.net/networks1/irrigation/index.htm

http://blog.allanglen.com/2009/07/building-a-smart-irrigation-controller-part-1

Following is for misterhouse and dated but mentions ET as well as looking at the forecast to reduce watering.
http://old.nabble.com/Fun-new-problem-td9726959.html

A more recent misterhouse thread:
http://old.nabble.com/Reducing-Garden-Watering-due-Rainfall-td23628522.html#a23628522

 

[linuxha]

MCSsprinklers looks interesting so not saying anything against it... But if you are interested in rolling your own for Linux (or for other platforms for that matter) you can find plenty of info and others doing similar things. I have been thinking of doing something like this for some time. Here are a few links:

http://members.cox.n...ation/index.htm

http://blog.allangle...ntroller-part-1

Following is for misterhouse and dated but mentions ET as well as looking at the forecast to reduce watering.
http://old.nabble.co...-td9726959.html

A more recent misterhouse thread:
http://old.nabble.co....html#a23628522

Excellent information, thanks for the links!