Don't get me wrong. Garden work can be relaxing, too, but everyday watering is not, especially, when you are not at home due to business trips or holidays. I know there have been irrigation computers virtually since the beginning of mankind, but with constantly gaugechanging weather conditions (as we experience in Central Europe) an irrigation computer needs a lot of luck to irrigate when it has to. So, a transition of irrigation computers into "smart" computers makes sense not only in terms of usability, but also in terms of convenience.
Based on my experience with consumer electronics, I have created a concept that offers great user experience at reasonable costs and development efforts. This article contains some technical details, so even if you are not interested in electronics (yet), you might get some valuable insights.
Smart Irrigation Hub
So what is a smart irrigation hub? "Smart" is an often (over)used term, but the general understanding is that a device is smart when it is connected, not necessarily to a cloud, but to other devices or apps. This means a smart irrigation hub can be set up and controlled via app, and it can connect to devices like moisture sensors and rain gauges using various wireless technologies. A smart irrigation hub must be able to accomplish everything an irrigation computer does, but with more focus on user-friendliness and "smartness". Computers switch valves on and off based on time events. A smart device also takes the weather forecast and the location into consideration.
This irrigation hub is not only a gateway to other wireless sensors and valves, it also contains an initial use case.
- System on module with ARM v7 processor
- 256MB RAM
- 1GB SLC NAND Flash
- WLAN/BT 4.2 SoC/SoM
- 24V power supply with 5.5 mm coaxial DC connector (1A/24W)
- 4 x 24V DC outputs for water valves, at least
- 100mA each
- 8 spring clamps in total (add an appropriate tool)
- DC-DC converter 24V to 5V
- 6 buttons (one for each valve + reset + connect button)
- 5 status LEDs (one for each valve + connection status)
In accordance with the old saying "everything is better with Bluetooth", communication with the sensors is based on
Bluetooth 5.0 utilizing the mesh capabilities. This means we can save an additional SoC for another RF
Incidentally, the latest mesh feature happens to be great for irrigation as it allows to cover a larger area.
Also, Bluetooth SoCs are cheaper than 6LowPAN or Z-Wave without any downside. Oh wait, Z-Wave has 868/912MHz, so it
might have a better wall penetration, right? Right. Only, most of the time, this is not the case.
If the user places the smart irrigation hub in their basement, for example, they will still have a hard time connecting it to their garden sensors. NFC is put to use to pair the sensors and other devices to the hub. Just hold the sensor above the hub and it connects automagically.
Do yourself a favor and use a Linux distribution that can be updated easily. Bare Yocto is great to save costs on the
flash, but well... I prefer Ubuntu Core due to its sandboxing technology, and the neat update staging and rollback
Regarding flash, just don't/NEVER EVER use MLC flash. If you can afford it, use eMMC if not SLC. Oversize the flash a bit, so it is future-proof for updates.
To switch the 24V on and off, I would use solid-state relays with a short circuit protection. Since the user can do
some unintended stuff with the outputs, I would use ye ole glass tube fuse to protect the input.
And what if the user wants more outputs? Some might think about adding an RS485 bus, but I would add a Bluetooth relay board. It is easier to use, and you can reuse the same Bluetooth modules you have already employed for the sensors.
You might ask yourself: "Why do I need an app for all that?" Well, it is definitely not about turning the irrigation on and off. It is about an easy setup. No need for an operation manual that nobody reads anyway: the app guides the user through the necessary steps and can advise them on best practices.
The old fashioned irrigation computers from famous brands are terrible in terms of UX. There are already some smart computers available, but the UX must get better. The average user doesn't know how to set up a rule based on the soil moisture to automatically turn on the sprinklers. Like I already said, the user needs guidance and rule templates, so setting up a rule becomes a piece of cake.
Besides, a rule should not only consider whether the plants have enough water. It should also save water. For example, if it is going to rain in a hour, it is not necessary to turn on the sprinklers even if the soil is dry. On the other hand, if it is really hot, and the sun is beating down, it is not effective to water, and irrigation should be scheduled to the evening when it is getting cooler.
We wanna make the setup easier, so don't bother your customer with another user name and password. Use push-button pairing instead: the app is paired by the push of a button, easy and secure. Wi-Fi setup can be done with access points, light signals, or Bluetooth. (There are other methods, but these are the best.) Out of the three options I've presented, I would go with Bluetooth. It is the most convenient in terms of UX, and the user doesn't have to change the Wi-Fi network, where every change causes some delay, or hold the display above the device. The app connects automatically to the Bluetooth device, and the user enters the credentials, and gets immediate feedback whether the credentials are correct.
Use a 1" 24DC NC (normally closed) internally piloted solenoid valve (the one made out of plastic). Doesn't have to be a fancy one, all the "special" membrane materials for water valves with great sounding marketing are more or less the same, and you probably guessed it: water is not the most corrosive fluid.
Smart Garden Sensors
Measuring moisture is tricky in terms of accuracy. I prefer the electrical resistance soil measurement method. There is no such thing as one standardized soil that has a specific electrical resistance. As moisture sensors aren't completely accurate, I recommend to measure in three categories: wet, medium, and dry. The results should certainly not be displayed as a percentage for residential irrigation. The water distribution is not necessarily even, and the measuring takes place at a specific spot in the garden, so it is not 100 % representative for the overall average soil moisture. The measurement is done by sending short electrical impulses through the soil and measuring the voltage with an analog input of the microcontroller (uC). By the way, don't underestimate the corrosion on the metal parts stuck in the soil: use stainless steel skewers.
Just to mention other methods: With capacitive sensing, the moisture influences the capacity, and by measuring the time it takes the voltage to reach a certain level, you get a clue of the moisture content. Some devices measure the temperature at the surface and different depths of the soil, and then determine the moisture by calculating the temperature difference.
Battery measurement is not as easy as it seems if you want an accurate percentage, and it is not necessary either. The user just wants to know when they need to change the battery, so a critical battery alarm is certainly enough. Two AA batteries in series deliver around 3V, which is enough for most Bluetooth SoCs. You need no additional voltage regulation - just a brown-out circuit, which is integrated in most SoCs.
I don't see (yet) how temperature and light measurements can benefit the plant care, but it is certainly neat to have this information, and those sensors don't cost much.
You want to use a solar cell to power the sensor? I support every idea that helps reduce waste, especially dangerous kinds like batteries, but keep in mind that you will need a rechargeable buffer battery, and a few more electronic components that will increase your development and testing costs.
The sensor communicates with the hub via Bluetooth 5.0, and the pairing is done with NFC or Bluetooth discovery. The device needs one button to reset bonding data, plus one LED to give the user feedback about the status (advertising, paired, idle).
- Soil moisture senor (stainless steel skewers)
- Optional temperature sensor
- Optional light intensity sensor
- 2 x AA batteries (incl. battery holder)
Rain gauges based on the tipping bucket principle have proved their worth. Take a look at this page for a deeper understanding of the tipping bucket principle: http://park18.wakwak.com/%7Eweather/archive/RainGauge.htm
The electronics part is super simple: take a good old transistor, a pull-up resistor, and a hall effect switch. Every time the bucket tips over, the magnet passes by the switch, and the microcontroller counts up.
As with battery measurement, communication takes place via Bluetooth 5.0, and pairing is done with NFC. The rain value is transmitted in millimeters. Again, the device needs one button to reset bonding data, plus one LED to give the user feedback .
- Bluetooth 5.0 SoC
- Reed switch
- 2x AA batteries (incl. battery holder)
Possible consumer services are:
- Weather service
- Remote connection
- Notifications (push or SMS)
- Alexa integration
When it comes to online services, there is no need for planned obsolescence. Do our planet a favor, and design the product to last as long as possible! I know you're probably thinking: "But what about my future revenue?" The solution for your headache are subscription based services. These services do not need to be limited to the very basic ones I mentioned. Maybe you can come up with a sophisticated image processing solution to monitor the plants' health coupled with automated fertilizer and pesticide ordering on your favorite e-commerce platform.
Clouds are cool, but don't put basic functionalities in the cloud. If I can't irrigate my garden because of an Internet outage, I'm pissed. Aside from that, company servers might be shut down in the future, leaving me with inoperable pieces of electronics. Cloud solutions are okay for remote connection, notifications, or any other ADDITIONAL services, but again, not for basic functions. Period.
The smart irrigation hub does not need to be restricted to irrigation - even if the name suggests otherwise. It could also be put to use for swimming pools (water sensor, pool pump) or garden lights, for example.
In addition, as I have already mentioned under "Services", you may come up with new nifty ideas as to which services you can excite your customers with.