Even though Rob kept insisting we start documenting our work, I figured I had time and I'd do it in the end and what resulted was a very painful last-minute rush yesterday. Thankfully, there's a 1-week extension and I plan to update this blog on a daily basis, starting today.
My name is Aashish. My idea for this competition is a low-cost, digitally controlled, Personal Hydroponics Farm. Hydroponics is the science of growing plants without soil (directly nourishing them through nutrient solution) and what makes this a "Personal" farm is that it is portable, small enough to fit inside a kitchen or a bedroom and designed to plug directly into a regular wall socket found in our homes. I've also decided to call it the HYGROMAX 630 since it's a 630 liter unit (but mostly in the hope that having both "MAX" and "630" in the name might attract more points![:)]( "Smile :)")
).
I'm going to write this blog with a view that it might help anyone who might be inspired to try building this on their own. So the information may be unnecessarily detailed and there might be ramblings about decisions that I took and why I took them.
Below is a side-by-side comparison of the design sketch I came up with right at the start of the competition and a picture of how the unit finally looks. A few features had to be modified (removed) for various reasons (for starters, the load cells weighing the plants seemed to be overkill, even though the MAX32630 seems to have no problems taking on any additional tasks you throw at it).
![sidebysidesmall.jpg]()
It will, eventually, have white, acrylic walls and a door to seal the unit, minimizing any threat to the growing plants (3 trays of nutrient solution, each measuring 100cm x 100cm x 21cm and a total of 48 plants - each level housing 16 (4x4) plants).
A pictorial BOM of the equipment that is actually being controlled by the electronics is given below.
![picBOMsmall.jpg]()
3 Humidity/Temperature Sensors - DHT-11s (I seem to have misplaced one), a pH electrode (need to pick up two more, but the reviews on Amazon for this one were suspicious, so I'll try it out first), 9 LDRs (3 per tray), an MH-Z14 CO2 sensor (was going to be 3 of them, but at those prices, my idea would no longer qualify as 'low-cost' ), an aquarium pump with tubing (for oxygen) and air stones (to enhance the amount of dissolved oxygen) and a set of T5 LED lights (the main source of light is going to be fluorescent, but this system will support configurable profiles such as "low power" in times of electrical failure, when backup power is being used - these LED lights would kick in at that time. Also included are 3 fans - mainly to circulate the CO2, when released into the system and a secondary role would be to serve as dehumidifying agents, when used with vents.
What is missing is a CO2 cylinder which I'm having trouble sourcing. The prices online are frightfully intimidating and there are a whole lot of local vendors willing to 'refill' my CO2 cylinder at a very small cost, if I have an empty one. Might have to resort to a CO2 boost bucket (synthesize CO2 with chemicals and pipe that into the unit) - but that is my last resort.
I wish the next picture was as pretty as the previous one. But it isn't. It's my workbench with most of the hardware built.
![junglesmall.jpg]()
The downside to the FTHR board offering such awesome features like SD card, Bluetooth, PMIC, Gyro, etc. is that only a few pins are available to use as GPIO. But with a processor this fast, multiplexing a whole lot of these signals becomes a very comfortable option. Two 74HC4067s open up 32 lines to be used for data acquisition and to further reduce the need for pins, a 74LS164 shift register is used to generate the 6 control lines needed to work the multiplexers. 3 8-channel relay boards provide 24 lines for the various actuators and here, again, the pins are minimized by using three shift registers.
Thanks to Ollie Milton for the 'ShiftOut' library on mbed. It cut down hours of downtime that I would certainly have encountered if I had gone at it alone. I know I should be using shift registers with latches on the relays, but I had a bunch of 164s lying around and was in a bit of a rush. The final design will have them, of course - that is also the reason I'm living with a jungle of wires and not getting these PCBs printed just yet. But that is not to say that I haven't done anything about the wires. A lot of time and solder has gone into getting pretty close to the real thing.
![manualPCBsmall.jpg]()
This approach did not, however, work with the 4 MAX7219s that I'm using to control the 32 7-segment displays to monitor the various parameters like humidity, temperature, pH, light etc. The wiring turned into something really nasty and debugging the problems on that started becoming pure rocket science (No offence, Dan Julio...![:)]( "Smile :)")
if you're reading this, Dan, please know that I am insanely impressed by your project ). I eventually got the PCBs printed.
![printedPCBsmall.jpg]()
Thanks to the team at Maxim for the MAX7219 library. For someone like me, who's always been accustomed to building everything from scratch and re-inventing the wheel a lot, it is a whole new, refreshing experience to see the kind of contribution that is being made on the mbed platform. Libraries that can be picked up and plugged right into any project and they just work seamlessly, right off the bat. I'm so looking forward to doing my bit of contributing as well![:)]( "Smile :)")
The unit is configured through the memory card. The format of the file is continuously evolving. There was a huge dependence on the file since I wasn't able to get the BLE to work even after a ton of effort. The ideal thing would be to have an android app managing the whole configuration and leave the SD card for data archival/storage. I hope there's enough time for me to try this out, especially since there now seems to be hope in the BLE department. It would also be a lot of fun to be able to log this data into the cloud and retrieve a report from anywhere around the world (also because, in my intial enthusiasm, I ended up buying the domain - hygromax.com for this purpose).
On a final note, here's a picture of my 3-week old babies eagerly awaiting the completion of this project, the completion of their new home...
![lettucesmall.jpg]()
You can't make out clearly from the picture, but they're extremely grateful to Maxim Integrated and AllAboutCircuits.com for giving them this opportunity to embark on this awesome journey of Life.
Also, they can't seem to shut up about that Hololens.
My name is Aashish. My idea for this competition is a low-cost, digitally controlled, Personal Hydroponics Farm. Hydroponics is the science of growing plants without soil (directly nourishing them through nutrient solution) and what makes this a "Personal" farm is that it is portable, small enough to fit inside a kitchen or a bedroom and designed to plug directly into a regular wall socket found in our homes. I've also decided to call it the HYGROMAX 630 since it's a 630 liter unit (but mostly in the hope that having both "MAX" and "630" in the name might attract more points
).I'm going to write this blog with a view that it might help anyone who might be inspired to try building this on their own. So the information may be unnecessarily detailed and there might be ramblings about decisions that I took and why I took them.
Below is a side-by-side comparison of the design sketch I came up with right at the start of the competition and a picture of how the unit finally looks. A few features had to be modified (removed) for various reasons (for starters, the load cells weighing the plants seemed to be overkill, even though the MAX32630 seems to have no problems taking on any additional tasks you throw at it).
It will, eventually, have white, acrylic walls and a door to seal the unit, minimizing any threat to the growing plants (3 trays of nutrient solution, each measuring 100cm x 100cm x 21cm and a total of 48 plants - each level housing 16 (4x4) plants).
A pictorial BOM of the equipment that is actually being controlled by the electronics is given below.
3 Humidity/Temperature Sensors - DHT-11s (I seem to have misplaced one), a pH electrode (need to pick up two more, but the reviews on Amazon for this one were suspicious, so I'll try it out first), 9 LDRs (3 per tray), an MH-Z14 CO2 sensor (was going to be 3 of them, but at those prices, my idea would no longer qualify as 'low-cost' ), an aquarium pump with tubing (for oxygen) and air stones (to enhance the amount of dissolved oxygen) and a set of T5 LED lights (the main source of light is going to be fluorescent, but this system will support configurable profiles such as "low power" in times of electrical failure, when backup power is being used - these LED lights would kick in at that time. Also included are 3 fans - mainly to circulate the CO2, when released into the system and a secondary role would be to serve as dehumidifying agents, when used with vents.
What is missing is a CO2 cylinder which I'm having trouble sourcing. The prices online are frightfully intimidating and there are a whole lot of local vendors willing to 'refill' my CO2 cylinder at a very small cost, if I have an empty one. Might have to resort to a CO2 boost bucket (synthesize CO2 with chemicals and pipe that into the unit) - but that is my last resort.
I wish the next picture was as pretty as the previous one. But it isn't. It's my workbench with most of the hardware built.
The downside to the FTHR board offering such awesome features like SD card, Bluetooth, PMIC, Gyro, etc. is that only a few pins are available to use as GPIO. But with a processor this fast, multiplexing a whole lot of these signals becomes a very comfortable option. Two 74HC4067s open up 32 lines to be used for data acquisition and to further reduce the need for pins, a 74LS164 shift register is used to generate the 6 control lines needed to work the multiplexers. 3 8-channel relay boards provide 24 lines for the various actuators and here, again, the pins are minimized by using three shift registers.
Thanks to Ollie Milton for the 'ShiftOut' library on mbed. It cut down hours of downtime that I would certainly have encountered if I had gone at it alone. I know I should be using shift registers with latches on the relays, but I had a bunch of 164s lying around and was in a bit of a rush. The final design will have them, of course - that is also the reason I'm living with a jungle of wires and not getting these PCBs printed just yet. But that is not to say that I haven't done anything about the wires. A lot of time and solder has gone into getting pretty close to the real thing.
This approach did not, however, work with the 4 MAX7219s that I'm using to control the 32 7-segment displays to monitor the various parameters like humidity, temperature, pH, light etc. The wiring turned into something really nasty and debugging the problems on that started becoming pure rocket science (No offence, Dan Julio...
if you're reading this, Dan, please know that I am insanely impressed by your project ). I eventually got the PCBs printed. 
Thanks to the team at Maxim for the MAX7219 library. For someone like me, who's always been accustomed to building everything from scratch and re-inventing the wheel a lot, it is a whole new, refreshing experience to see the kind of contribution that is being made on the mbed platform. Libraries that can be picked up and plugged right into any project and they just work seamlessly, right off the bat. I'm so looking forward to doing my bit of contributing as well
The unit is configured through the memory card. The format of the file is continuously evolving. There was a huge dependence on the file since I wasn't able to get the BLE to work even after a ton of effort. The ideal thing would be to have an android app managing the whole configuration and leave the SD card for data archival/storage. I hope there's enough time for me to try this out, especially since there now seems to be hope in the BLE department. It would also be a lot of fun to be able to log this data into the cloud and retrieve a report from anywhere around the world (also because, in my intial enthusiasm, I ended up buying the domain - hygromax.com for this purpose).
On a final note, here's a picture of my 3-week old babies eagerly awaiting the completion of this project, the completion of their new home...
You can't make out clearly from the picture, but they're extremely grateful to Maxim Integrated and AllAboutCircuits.com for giving them this opportunity to embark on this awesome journey of Life.
Also, they can't seem to shut up about that Hololens.