The Monday Jolt is a new series about microcontrollers and electronics that appears in MAKE every Monday morning.
HarvestGeek is an ambitious project to automate a greenhouse or other growing environment through the use of a variety of sensors hooked up wirelessly to an Open Source, Arduino-compatible base station. An additional unit can trigger greenhouse fixtures like lighting, fans, and pumps.
For instance, you could have a crop of basil growing in your basement with sensors measuring temperature, humidity, luminosity, CO2, pH, soil moisture, and so on. The system is so much more robust than merely a glorified timer: you can monitor and control your setup remotely, and automate certain tasks like lighting, ventilation, and watering.
If your greenhouse gets too cold to support your basil crop, for instance, you could get a text letting you know, allowing you to correct the situation before it affects the grow. Meanwhile, you’d have the option to share your data online and collaborate with other farmers. In the hypothetical case of the basil crop, you could see if other people were having better luck with different temperatures, light levels, or moisture.
HarvestGeek founder Mike Alt is a software engineer who been working on HarvestGeek for the past year. As the project closes in on completion, Mike has created a Kickstarter campaign to help complete the project.
Anatomy of a HarvestGeek Setup
The Base Station is the brains of the HarvestGeek setup, communicating with its various substations via a 915 MHz JeeNode network.
The Automation Station controls the greenhouse environment, triggering fans, lights, pumps, and other devices as directed by the base. Each automation station can control up to four devices, each independently switched and rated for 15 amps.
Each Sensor Station is equipped with a wireless module and is powered via USB, and each base station can connect to up to 26 of these units. Inside each Sensor Station enclosure, multiple sensors can be installed.
All of these components are Open Source, and HarvestGeek will be sharing their files, code, and schematics after the Kickstarter campaign concludes and the units start to ship.
While HarvestGeek is open source, there will be some who will want an easier solution than hack-your-own. HarvestGeek will offer a HarvestGeek Subscription, an online toolkit that allows users to set up their networks, monitor readings online, and publish to the cloud, creating an environment of collaboration between HarvestGeek users.
An Interview With Mike Alt
Last week I interviewed Mike via email and phone:
JB: Could you tell us how the idea germinated (sorry!) to build HarvestBot, and ultimately to Kickstart it?
MA: The idea sprouted (sorry!) a few years ago, I had been tending a small indoor garden and I found I was spending a ton of money on equipment for my plants. Temperature and humidity sensors that were inaccurate, expensive pH and PPM probes, timers, switches, all kinds of stuff. I had a bin full of the tools of the trade and knew there had to be a better way. I started researching indoor monitoring systems and realized they were way out of my price range. There was a forum post or blog entry or something that was detailing his homebrew DIY grow system and I knew right then I had to have one. I had no electronics experience and a lot of free time so I set out to build my own.
JB: Give me an example of how you would use the HarvestBot modules to monitor a typical greenhouse, for instance.
MA: In my experience there aren’t many typical greenhouses, they tend to take on the personality of the one who tends it. I think a pretty common application would be a hobbyist grower with a sensor station or two monitoring their plants and checking on them from work. In our “lab” we have a few sensor stations, a hydro station monitoring our main reservoir and an automation station to manage our water pumps and a fan. I really thought the automation station would be a huge time saver, but it turns out the bigger time saver was not having to manually check the temperatures and pH levels every day. The prototype is still up and running in our lab, I expect I’ll have to take it back to the office in the next week or two to do some more work on the next revision.
JB: Can you tell us more about the technical details of the HarvestBot base station? Is it based on a familiar and hackable format like Raspberry Pi or Arduino?
MA: HarvestBot is fully Arduino compatible and runs with much the same specs as an Uno using the ATmega328p. Somewhere along the line in my research, I stumbled upon the JeeNode, which can be found here at http://www.jeelabs.org.
Much of my design incorporates the work of the very talented Jean-Claude Wippler. HarvestBot followed in its footsteps and uses integrated RF via the Hope RFM12B for wireless communication. The cost of these is much lower than ZigBee and in most applications just as well suited as a Zigbee module.
This is one of the remaining hardware decisions I have though. Switching to ZigBee does offer one distinct benefit. Being an open-source platform, it would be optimal to be able to integrate and communicate with other IoT devices, and it seems like ZigBee is more or less the standard at this point (until the cost of WiFi comes down).
JB: Tell about the evolution of the project from its infancy to where you’d like it to go.
MA: My prototype is currently running an ENC28J60 10BaseT ethernet chip. I was originally using the WizNet 5100 but realized it was overkill for the amount of data being transferred and scaled back in order to save a few bucks. The not-so-long-term plan is to go WiFi and incorporate the very awesome Electric Imp, it is actually a stretch goal if our kickstarter reaches $50k.
When I was first prototyping for my own use, I choose the best of all worlds. Cost wasn’t important to me at that point I was just building a cool toy for myself and it was a relatively cheap hobby in comparison. When I decided to make HarvestBot a “consumer electronic” there were a lot of very tough decisions I suddenly had to make that I was more-than-ill-equipped to do at the time, with my lack of experience in electronics and knowledge of manufacturing. Knowing where to save two dollars and where to spend five ends up being an important decision in the long term goals of any product.
In the end, I had to pick a number that I felt comfortable with, that I thought consumers would spend on the system. The number I came up with was $150 for the Base and a couple sensor stations. I quickly realized that I was nowhere near that number and a lot of the hardware I chose was really unnecessary for what I wanted to do.
The base station is Arduino compatible. The first few iterations were just Unos and sensors stuffed into a project box and it evolved from there, breaking out the stations into separate units. The base station has an LCD which reads out the data from the sensors stations and Ethernet which relays the data back to the cloud. It communicates with the other stations via RF, there is surprisingly little else to the base station. After deciding I wanted to Kickstart the project, I struggled often with feature creep wanting to add in every bell, whistle and neat thing I came across. I was reading a bunch of Lean startup books at the time and the thing that really struck home was creating a Minimum Viable Product. I pared everything back considerably. I don’t think what I ended up building was exactly that, but the base station and sensor station were pretty good compromises to keeping the project super simple and affordable.
JB: One of the parts of the project that particularly intrigued me was the possibility of small farmers sharing environmental data real-time. Can you talk about how that collaboration would benefit farmers?
MA: I’m pretty excited about this myself. Having the real-time and historical data from my grow created a feedback loop that let me quickly improve my grow. It let me take a more scientific approach based on quantifiable data. I’m really just one person with a hobby. The power of the feedback loop is derived from observing an event or events, analyzing the data, and improving your method based on the information at hand. Combining the power of the feedback loop with the power of the collective collaboration and crowdsourcing and I think the results will be very interesting. I think farming, especially on a small scale has a lot of room for improvement. I’m very anxious to see what innovations fall out of the process when real farmers begin to utilize this method.
How would you use HarvestGeek? Do you think it shows promise for urban farmers? How could the online collaboration benefit you? Leave a comment with your thoughts!