$300 MVP Food Computer Variant - My Build



I just completed a variant of the $300 MVP Food Computer. I encountered some challenges based on my design changes, and I encountered a few also because this is a new design that is still being perfected. It’s taken me about a month to gradually complete my build; I’ve taken photos and notes, and over the next couple of weeks I intend to document my progress to this point. The project sponsors have been extremely helpful with the issues I have encountered. You can review my questions and their answers here: $300 Food Computer.

My name is Dave, and I’m an amateur (ham) radio operator, holding a General class license. I have a Master’s degree in a non-STEM field. My coding knowledge is extremely limited, but I am handy with a soldiering iron, and I understand basic electrical and electronic theory.


I decided to build an MVP variant using a commercially produced metal shelving system. There were several reasons I decided to do this, despite the increased cost (the shelving unit runs around $80 at home improvement stores, compared to maybe $20 using PVC and a plywood piece for the top).

First, using the shelving allowed me to easily double the size of the chamber. I’m certain that you could increase the size using PVC pipe, but you’d probably have to add some cross members to support the additional weight.

Second, I am running this project in my basement, and I occasionally have mice come for a visit. Using the shelving allows me to have the grow chamber off the floor (yes, the mice can still climb it but it might be a challenge), and also having it off the floor helps stabilize the temperature, and the extra shelves provide me with space to store the consumables (fertilizer, PH adjuster, etc.) as well as a location for a monitor and peripherals if I need to access the Raspberry Pi directly.

And Third, I’m considering eventually adding an automatic pumping system (ebb and flow method), and having my grow chamber elevated will give me a gravity return for excess nutrient solution.


While I think it is possible to build the MVP at near the $300 cost as stated, be aware that variations, changes in availability of materials, and simple mistakes may increase your cost. I haven’t kept detailed records of costs associated with my build, but I estimate that I’m close to twice the project cost using the original design. Regardless of how you plan to complete this project, be aware that if you intend to keep your costs low, you may have to delay your build to obtain the lowest priced materials due to availability. Also remember the old adage of “Measure twice, cut once.” Visit the original $300 Food Computer thread and see the bill of materials here: $300 Food Computer.

So my first post is a wall of text. In a day or two I’ll continue to provide more details on the issues I encountered and/or methods I used to build my MVP variant and some thoughts that weren’t addressed in the instructions.

Food Server Documentation (Shipping Container Farms)
$300 Food Computer - MVP
$300 Food Computer - MVP

Now that I’ve mostly introduced my project, I will try to give some lessons learned, and I’ll try to roughly follow the same sequential outline in the original $300 Food Computer project.

There was one additional point I failed to mention in yesterday’s post about why I decided to use a shelving system. I had the thought that it would allow for expansion. You could perhaps call it a “rack system;” but I simply think that you could put two double size MVP builds together. Having completed my build, I think it would be a significant task to take enough of the structure apart to allow me to adjust the shelving for a second unit, but it is still a possibility. Perhaps an idea for someone more adventurous than I am.

Bill of Materials

As I mentioned before, you may face a challenge keeping costs low. If you intend to build a variant of the MVP, for example doubling the space as I have, consider that you will need twice as many of several items. Although the control system can easily support the additional amperage of twice the lights and fans, the additional hardware increases your costs.

Additionally, sometimes going with the cheapest materials doesn’t reduce your costs; you simply waste time when you realize the cheapest option will not work for you and have to backtrack to locate a suitable option.

Also, the materials you order from the BOM may not be the exact materials shown in the instructions. For example, the 12 volt power supplies I received had cords where both the positive and negative wires were contained within a single insulator. I had to carefully remove the outer insulation, and identify the positive and negative wires inside. Luckily they were red and black, but I was prepared to use my voltage meter to determine which was which.

Another example. I’m considering setting up an automatic pumping system to periodically refresh the nutrient from a reservoir. I purchased an aquarium pump, buying the least expensive one I could find. I determined after a test in the kitchen sink that an 80 GPH aquarium pump is not adequate – it’s cheap and ineffective. I now have two rated at 550 GPH and when I am ready to move on to the next phase they should be effective. They easily move water rapidly through a ten foot piece of 1/2 tubing.

A couple of other thoughts to consider about the BOM (I haven’t checked it in a couple of weeks, so this advice may be out of date). Check the Raspberry Pi website Raspberry Pi website for hardware requirements and frequently asked questions. I had several hardware and operating system build questions answered by visiting their site.

One last example about flexibility regarding the BOM, and I’ll move on. Zip ties. The zip ties I received from ordering off the link from the BOM were a great value, but they were too large to fit through the holes in the board of the Raspberry Pi and the switching unit. You need very narrow ties to mount the Raspberry on the brain. When I went to the home improvement store, in addition to smaller ties, I also found some plastic mounts with adhesive feet that allowed the necessary standoff from the enclosure top and were (in my opinion) a slightly more expensive, but better option than zip ties through the wood.

I was planning to include a “prepared parts” discussion in this post, but I’ve already written another wall of text. So I’ll conclude with one additional comment regarding materials. Plastic. It’s not a fire-alarm screaming issue, but consider that plants absorb nutrients and contaminants through their roots. You’re going to have water solution in plastic tubs for months or possibly years. I think it’s worth considering the health safety of the type of plastic your tubs and lids are made from, and any other materials you may use that directly interact with your crop. I think most of the tubs you may find are probably made from the types of plastic considered safe. I found the Smart Plastics Guide to be a good primer on what plastic materials to avoid.

In closing, to keep your cost as low as possible, stop and consider if it will work. Make some notes, drawings, whatever helps you visualize the end result. Think it through in advance, it will reduce your costs and avoid numerous re-orders or extra trips to the home improvement store.


Prepared Parts

I already mentioned the difficulty I encountered with the 12V power supplies. I also had trouble with the fans.

I ordered a four-pack of computer fans; they were both economical and the link in the BOM at the time sent me to a fan that had LED lights on it, which was not what I wanted. I cut the connector off my first fan, and thought I had identified the positive and negative wires based on how they were marked. I stripped the ends and hooked it to a power supply to test it, and nothing happened. I tried several variations and could not get the fan to work. Finally, I took another one and started applying power to the pins on the computer connectors until I located the proper two (you probably won’t burn out a fan if you only connect power improperly for a second or two). My internet research revealed that most modern computer fans are designed for variable speed, and that it’s not always easy to determine which wire is the positive wire. I basically understand how the variable speed is done, but the point is that you may have to identify the proper wires on your fan. I recommend cutting the connector farthest from the fan first and try to identify the correct wiring it it’s not detailed in the documentation or on the box. Also, consider the distance between the fan and where it will connect to the power supply before you start cutting the wires.

When I viewed the instructions for cutting the holes in the bus tub lids my first thought was how long it would take and how rough the holes would be. And then I remembered I had the perfect tool; I had a two inch circular bit in my toolbox, it costs about ten dollars at home improvement stores.

The heat of the drill will melt a little bit of the plastic…

And you’ll probably want to use a file or sandpaper to clean it off…

But it’s a lot faster and uses less arm power than cutting several holes with a keyhole saw. Just be careful not to drill down into whatever you have under the lid, and make sure it’s secured (or hold it firmly) so the lid doesn’t try to spin around.

On the subject of power tools, I needed to cut the ends off of eight of the LED lights. I did the first two with a small hacksaw, but it was a lot of work. Also in my toolbox is a Dremel tool. I also had some small cutting attachments for it, and made quick work of the light covers. I strongly suggest that you wear eye protection if you will cut with a Dremel tool; I did. I had one of the cutting “blades” (it may have been made of something similar to bakelite) shatter while using the tool, and although it didn’t fly toward my face, I was glad I was wearing eye protection. Use the appropriate safety measures whenever you use power tools.

What remains on this topic? I’ll briefly mention the Mylar insulation. I bought a roll that was 24 inches by 25 feet. Not only was it the only size my local home improvement store had in stock, but it was perfect for my double size build. Fortunately, I carefully considered what I would need before I started cutting it, and it worked out well. The tape used to connect the pieces is a pain to break loose a corner to get the backing off of, but I’m sure you’ll get the hang of it.

Let’s discuss placement and drilling holes on the plywood top, the “brain” of this device. When I placed my Pi and the relay, I turned the Pi 180 degrees, thinking it would be easier to access the SD card from the other side. That resulted in a somewhat tight connection for the power jumper (red wire) to the relay, and it’s a little tight to plug in USB connections (the relay is in the way for one of the USB ports). Before you attach either of those boards to the brain, take a moment and consider if you’ve allowed adequate room (or made them close enough) to attach connections to them.

What’s left on this topic? I won’t go into measuring and cutting PVC, as I used a shelving unit that was ready for assembly. I’ve used PVC pipe for several different projects, and I think it works well if you want to go that route.

Last on this topic; soldering the pins on the Temp/Humidity Sensor. In the documentation it shows the gentleman using a small device sometimes marketed as “helping hands.” The small stand with clips and a magnifying glass. I have one I purchased from Harbor Freight for about five bucks. If you don’t have something like that, you probably can lay it on your workbench after the pins are inserted and soldier it, but it would be a little more difficult. You don’t need an expensive, high temperature soldering iron for this step, an inexpensive one will work just fine. When you soldier the pins to the sensor, don’t remove the plastic sections from the pins. Simply put the (remaining) four pins (the short side) into the four holes to the right from the top of the chip. Soldier the pins from the bottom of the chip. Hold the iron to the pin and touch the soldier to it near where the iron is on it. It doesn’t take a lot of soldier; as soon as the soldier flows into the gap for the pin, you’re done and can move on to the next one. The pin will heat up quickly; don’t hold the soldiering iron on the pin for a long period of time or you may damage the circuits on the chip.

Another wall of text, but I believe that covers all of my observations and suggestion for the prepared parts stage. One last note - take photos! Consider photo documentation early. You might want to share your experiences with others who are involved with this project, or you may simply want to show your friends what you have accomplished. If you’re trying something unique, others may certainly benefit from your experience, and photos help explain it.

In a day or two I’ll move on to the next topic. If you’re just starting this experiment, I wish you the best.


Build Enclosure

I already mentioned that I used a commercially manufactured metal shelving unit. I’ve used this same type of shelf for years, so I’m familiar with it’s assembly.

Basically, follow the assembly instructions. You probably will want to omit a shelf, but this unit has two vertical sections, so there must be a shelving bracket between the sections. It uses gravity to hold it together; a rubber mallet comes in handy for ensuring the sections are properly joined, but no other tools are required for assembly.

Where the two vertical sections join, I reinforced them with zip ties; probably not necessary, but if I have to move the assembled unit, I don’t want it coming loose at the weakest part of the structure.

As I’m using a wooden shelf for the bottom, I stapled the mylar to the shelf.

You’ll have to make your own measurements for the mylar walls. Considering that your zip ties are going over the shelving brackets, turn them so the spot where the end of the tie connects is not near the shelving channel. She shelves generally aren’t perfectly tight, so you shouldn’t have any trouble with the zip ties interfering.

And cut your hole for the exhaust fan. I thought about installing two fans, but one seems adequate as I didn’t do a perfect job when I taped and the doors are not airtight, so I have increased airflow.

I quickly realized I would not be able to reach the far side from the single door so I decided to place a door at either end.

I don’t remember any guidance in the original instructions on securing the door. I decided to use short bungee cords.

There is an issue with the cords gradually pulling through the mylar of the doors where they are poked through. Reinforce that area with tape. I’m still searching for a better way to secure the doors. Velcro maybe?

Edit: Thought just occurred to me as I was proofreading; perhaps wrapping a couple of layers of tape around the hook would decrease the tearing, by increasing the surface area it’s contacting. Haven’t tried this yet. Suggestions are welcome!

The only other advice I will offer for this section is think twice about where you will build this. The original MVP probably doesn’t weigh very much; once built this shelving unit will require at least two people to move it as assembled. I left a space near the wall for access to the back, and I carefully considered locations of electrical outlets.

In my next installment I’ll discuss the Brain preparation and assembly, and then will conclude with what little advice I can offer regarding the software and software build.


This is incredible feedback. I’m going to steal your advice regarding soldering for our documentation - eventually we want to find a pre-soldered sensor (SI7021 is just so cheap). With regards to tools (hole saw, dremmel, saw) we assumed the person had nothing to start. I think it would be helpful to provide tiers of tool options. I’ll be honest, I would never use the keyhole saw instead of a hole saw on my own build, that being said, it works and is the cheapest option.

@melanieshimano @jimbell @ferguman @drewthomas89 @webbhm

@bennis found that reusable zip ties work great for the door. I think Velcro would also be a solid option. I know that this is a missing component - I just haven’t found the right answer yet but will share soon as I do.

Thanks again, @RadioDave, keep up the great work, I can’t wait to see your plants!


Brain Preparation and Assembly

Although preparation is included in “prepared parts” in the original instructions for the $300 Food Computer, there really isn’t a reason it can’t be prepared and assembled at the same time.

Considering I was constructing a variation of the original design, I created a rough sketch and tried to get an idea of where to place the components for my build.

It made it easier for me to make changes to the sketch than drilling holes in the wrong places. I started with a few scrap pieces of wood underneath thinking I’d drill on the floor, then realized I had the necessary tools to do it more properly.

You may notice I changed the location for the camera. I was concerned that if I placed it too close to the light, I might have trouble getting the angle I needed.

You may also notice I’ve added two holes, two inches in diameter near the edge on both sides, one slightly out of alignment. As this shelf will sit on top of metal rails, I didn’t want the opportunity for the metal to cut into my wiring. The hole to the far right of the photo above is intended to be large enough to pass a standard 120V electrical plug, while the slightly smaller on toward the left is intended more for sensor and low voltage wiring. Later, I cut slits into the insulation on the underside; the wires pass through and the insulation mostly closes the hole (not illustrated).

I extended the leads from the sensor using female/male/female. I wrapped the connections and sections of the wire with electrical tape to help keep them joined.

When I went to the home improvement store to purchase smaller zip ties for the two electronic boards, I found these plastic mounting bases. You could still use zip ties through the wood, but these allow securing the circuit board to the wood, and act as stand off insulators. I have been reluctant to place electrical tape on the circuit boards. It can get gummy and sticky over time, and if the circuit boards need to be modified or re-purposed, I didn’t want to ruin them.

Below is what it looked like after I had stapled the mylar to the underside of the brain, installed the lights, and mostly completed the connections. Due to the size of some of the power supplies, one power strip would not have been adequate so I added a second.

When you begin connecting the jumpers between your Raspberry PI and the relay, you might want to go to the brain sub-assembly in the original plan, and print the wiring diagram so that you have it close. If you conduct your final tests and something isn’t working, the first thing you should check is that your jumpers are connected properly.

I’d also like to note that I did not wrap the relay with electrical tape. I was careful to cut the leads on the 120V extension cord so that no bare wiring would be exposed, and I secured them solidly in the terminals. I also wrapped tape near the ends, and used zip ties to help prevent movement of the cord. I considered covering the entire relay assembly with an electrical box, but as I have no children or pets who might be at risk, I didn’t pursue that thought.

Once everything was assembled, I did a full power on test before I installed the light arrays.

You can see where the sensor is hanging toward the left of the photo. I wanted it lower toward the middle of the chamber to get a more average measurement of conditions.

I actually tested both of the light fixtures first, then cut the cord and assembled the electronics later. Each time I tested a new fixture, I heard a “pop.” I suspect the foil underneath contacted or nearly contacted the live connections, and the electricity arced and deflated that section from contact. Edit: This could create a potential hazard, you may want to consider cutting circular sections from the bottom of the brain insulation to install the fixtures directly to the wood. You might also consider using a plastic electrical fixture mount to install the light fixture.

Important Edit: The light fixture mounting is under redesign and a discussion is underway regarding the $300 MVP design for the lights. When I have completed my modification, I’ll replace the above paragraph.

Everything seemed to be working, at least from a purely electrical perspective. I was pleasantly startled when the fan powered on about 30 seconds after startup. I still had a few issues with the software control, which I’ll address later.

As I started to wrap up the final stages, I realized the cords on the air pumps would not reach the power strip. I brought in an extension cord through the larger access hole and zip tied it to the unused shelf support. I wanted to ensure the 120V outlet was well clear of any possible liquid that might be spilled in the chamber, and I used a three prong grounded cord.

There is one additional suggestion that I probably should have addressed during the materials preparation discussion, but I didn’t perform it until this stage. It may go without saying, but wash your grow tubs and lids with a little bit of soapy water. Mine were covered with some sort of dust or powder, and I did not want my plants absorbing some unknown contaminant.

If you’ve managed to stick with me this long, I hope you have also familiarized yourself with the original $300 Food Computer MVP build, and have been able to see the differences between those plans and what I’ve done, as I’m not attempting to create a stand-alone set of instructions. I’ll complete this series within a day or two with my observations regarding the software build and operation.


@Webb.Peter This is a safety hazard. I didn’t realize until looking at your documentation yesterday that the light fixture is situated such that exposed 120V wiring is mounted against the reflectix. My understanding is that stuff is basically aluminized mylar. I’m not sure if the surface layer is conductive, but presumably if you poke it with a strand of copper or a sharp screw, it would be easy to make an electrical connection with the aluminum layer.

The popping noise sounds like a portion of RadioDave’s reflectix was acting as a fuse and vaporized. Depending on how much of the reflectix melted, one side of the wiring could still be making contact with the aluminum layer. With the potential for humidity and condensation, that could lead to a shock hazard, high voltage getting onto the I2C sensor wiring, etc.

Adding a plastic junction box under the light fixture would probably solve the problem.


@wsnook, @Webb.Peter

Cutting a circular section out of the Reflectix and mounting the fixture directly to the wood would probably also solve the problem, as I suggested. My thought was if it was a continuous short, it would have popped the power strip, or my home breaker. But you make a good point about humidity and condensation leading to a potential hazard. Over the next couple of days, I’ll shut off the power and follow my own suggestion and remount the fixtures directly to the wood.


Enclosure Build Door Modification

I think I’ve found a solution to securing the doors that works for me. I will stick with the idea of the bungee cords until I see something better, as I can easily attach them to the holes in the vertical supports and basically make them as tight as I want.

Here is my solution:

I bought this kit for about ten dollars at the home improvement store. Extra grommet part packs cost three or four bucks, if you would have a future need to use this tool. I saw something similar on Amazon for a little less than seven dollars, but I wanted to try it today, and I’m pleased with the result.

I reinforced the area where the previous hole was with tape, then installed the half-inch grommets. They seem to be firmly seated, and will prevent the hook from cutting into the mylar material any further.

It took about 30 minutes to install four grommets. It was a slight challenge working alone, as I had to hold a wood backstop and the back part of the tool in place on the already installed vertical door while I hit the front side with a hammer. This modification probably would not work well with an assembled MVP with a PVC frame, but the metal frame and wooden shelving at both the top and bottom near my doors made it reasonable. If this was done during the materials phase of the MVP build it would not be a problem.

I’m not suggesting this is the only solution, but one that seems to work for me. Obviously it’s one more item to purchase, so there may be another, lower cost solution.


The aluminum layer in that stuff is very thin, so it probably can’t handle much current before it melts.

When you remove the light fixtures, I’m guessing you’ll find a few charred holes where the Reflectix vaporized back far enough to disconnect the shorts–similar to a low amperage, fast-acting fuse.


@ferguman This is something to consider about mylar. @drewthomas89 We will want to make this alteration to the design so that they cut out a circular piece of mylar.

@RadioDave Please include pictures when you make the fix if you do notice these charred holes. Also, the door modification looks great.


@Webb.Peter, @wsnook, @ferguman, @drewthomas89


I’m not sure my earlier suggestion is the solution. This evening I did what I suggested; I lowered the fixtures and cut a circle in the insulation. You’re probably aware my build has two lights. The second light was not an issue, but I think I didn’t completely cut the hole to fit and it only has one set of wires going to it. It fit snugly into place, but I think the edges of the insulation may have kept it from completely contacting the wood.

My first fixture was problematic. I thought I’d “do it right” and ensure it was mounted snugly against the wood. It refused to cooperate. I think there are two reasons, The first is that I ran my second light in parallel of the first, using the second set of posts on the fixture, so I had two sets of wires. But this type of fixture is not designed to be flush mounted; I think the insulation was acting as a stand off to allow room for the wiring. Here’s what the bottom looks like in case you haven’t looked at one recently. There’s almost no space for the wires.

The circular piece to the left of the fixture is part of the (broken) fixture. It came apart when I went to reinstall the lights in it. I guess the ceramic fixture is not made any better than the plastic. Luckily I had an extra left over from a previous project. I have it working, but the install is not permanent (no photo). But this isn’t the issue at hand. I had to replace the fixture, and that’s when I realized the problem. There simply wasn’t enough room for the wires under the fixture.

I’m thinking that wsnook’s suggestion might be the preferred option. A plastic fixture box costs less than a dollar at home improvement stores, and would require only slight modification to use it for this purpose.

Photo from lowes.com

If builders don’t buy the right type of box, the holes won’t align with the screws on the fixture. I’m planning to modify my MVP with this soon.

As far as my concerns that started this discussion, I believe my original idea was correct. There’s no charred or melted material on the sections I removed. This insulation is basically bubble wrap with a foil cover. I think the current arced across the foil, popping the bubble underneath, which then deflated it away from the terminals. The only evidence I saw is what appear to be small pinprick holes in one of the sections. (The color appears copper, it’s not. It’s reflected from the wood cabinet near where I took the photo).

The holes don’t go through to the other side. The scratches are from when I rotated the fixture into place, and then rotated it again to remove it. I’m not an electrical expert, but I again have doubts this is a significant safety issue. But, I know enough to know the current instructions do not even approximate national electrical code, and with people of unknown skill levels constructing the MVP, it could become an issue, causing potential liability for creators of the instructions.

I’ve set myself a suspense to make this modification sometime over the next week; I’ll take photos and make notes and add them here. You can incorporate them into the MVP instructions, add them as an addendum, or disregard them as you see fit.

$300 Food Computer - MVP

@RadioDave From your picture and description, that seems like a mild and fortunately less-spectacular version of what I’d imagined. That’s good!

@Webb.Peter I agree that the liability angle is worth careful thought because you’re expecting these designs to be used in schools. Speaking from personal experience, working with 120V circuits is risky–small mistakes can be dangerous.

If I sound like a paranoid scaredy pants, it’s because I’ve had a couple near misses that did indeed scare me. Since then, I’ve been a lot more careful and have had no more problems. These days I’m an enthusiastic fan of precautions like using non contact voltage testers, working one-handed, and making sure high voltage wiring is well-enclosed where it can’t get wet and where nothing remotely conductive can accidentally touch it.

One way to steer clear of a whole category of possible trouble would be to just use unmodified UL or CE certified timers and grow lights. Another option would be to use solid state relays with covered screw terminals (e.g. Crydom CL Series) combined with boxes and fixtures comparable to what you’d use for a kitchen or bathroom. [edit: This IoT relay by DLI looks neat too: amazon product page and same thing at Adafruit]


Light Fixture Modification

Please be aware I have no qualifications to give advice regarding electrical code or safety. These are my opinions, and although it may be close to electrical code and safer than other methods, you should consult a qualified expert to ensure your safety and compliance with any codes or regulations at your location.

Tonight I removed and replaced one of the fixtures. I considered a couple of different options, and decided to go with a fixture mounting box designed for installation in an existing structure. One of my main requirements was that it not lower the lights into the grow chamber any farther. I was also concerned about difficulty of installation, but that was a secondary consideration.

The box I selected is designed to be inserted into a hole without access to the structures above, and uses a metal bracket to secure and tighten it into place. I believe this is referred to as an “existing installation” box, as in install into an existing structure.

It has “knock out” holes, so that you can create the openings for wires that are applicable to your application. It also is designed to hold the wiring in place once inserted.

I cut a rough hole in the brain using my two inch hole bit, then used a keyhole saw to expand it to the proper size. The fixture has a lip that is approximately 7/16 of an inch, so you have nearly half an inch to play with and the hole in the brain will still be covered. I traced the outline using a sharpie and cut the rough hole.

Then finished with the keyhole saw…

Until it fit snugly into the hole.

I then mounted it from the bottom and pulled the wires through.

There should be five or six inches of wires protruding from the holes.

Above is what it looks like from the top once installed, prior to inserting the wires. If you’re only going to have one light, only use one of the knock out sections.

I put crimp connectors…

… on the ends of the wires, because stranded wire is difficult to properly connect to screw terminals. If you use connectors, cut the stripped ends short enough that not much of the exposed wire extends beyond the crimped point. Crimp them securely, but not so hard you break the insulation. If they are crimped correctly, you can tug on them with moderate force, and they won’t feel lose or come apart.

If you look at the cord, one of the wires has ribs on the side, and the other side will be smooth. The ribbed side is the ‘neutral’ and the side of the wire without the ribs is ‘hot.’ You’ll notice that the terminals on your fixture are different colors. The ‘hot’ wire should be attached to the terminal that looks like brass, and the ‘neutral’ should be attached to the terminal that is more silver colored.

My understanding is that by connecting the proper polarity with a polarized plug, even a circuit that is not grounded will afford some protection against shorts or electrical shock. I’m reasonably sure I have this information correct; the circuit will work if you connect it differently, but if you want to be sure you have it wired correctly, you should confirm my suggestion with your own research or consult a qualified electrician.

You’ll notice that I have two sets of wires. That’s because I have a second light that is powered from a parallel connection from this fixture. If you are installing a single light, you will only have two wires.

Edit: Looking at the photo above, I am not sure I connected the wires that lead to my second light with the proper polarity. When I add the second box, I’ll take this light down and double check it.

This fixture also has “knock outs.” Basically they are thin areas that can be easily removed for multiple applications. To mount your fixture, depending on the hardware, you may need to knock out the inner set of holes.

Once you have the wires attached to the proper terminals on the fixture, fold them back a little so they will fit into the box, and attach the fixture to the box. If things have gone perfectly, you should be able to simply place the fixture over the screws already inserted into the box, and rotate it to lock it into place (and then gently tighten the screws until firmly fastened).

120V electricity can cause injury or death. Do not work on circuits unless you are sure there is no power attached. And if you have any indication of a problem (unusual noises, burning smell, failure to light), disconnect the power and check your work, and/or consult with a qualified electrician.


@wsnook, @Webb.Peter

I’ve posted my recommendation that I believe mostly covers the lighting safety issue. I believe at some point, the MVP plan should be reviewed by a certified or licensed professional. An electrical engineer, someone with an electrical contractors license, or some other recognized credentials. I doubt any school or other public institution will undertake the MVP project without some type of professional certification of safety. I did some research and spoke with an electrician about my modification, but it’s not adequate for widespread adoption.

Perhaps MIT can conduct the review.


Software Build

I’ve put off this last post for a couple of days because I’ve tried to consider what I can offer. I am not a coder (software developer), although I do have some very basic experience with coding in Basic and C+, and learned computers when everything was command line. I have an Associates degree in Information Technology and have taught Microsoft Office and introductory courses for computer hardware and operating systems at the undergraduate level, so I have a very basic understanding of the issues. But if your a coder and familiar with any of the software or operating systems used by the Raspberry Pi for this project, you probably don’t need to read any more.

The biggest problem as I see it, is the Raspberry Pi, its operating system, and the code used for the MVP is open source, intended for experimentation and software geeks, and is not designed for those who simply want to drop in a disk and have the software install after clicking “agree” on the Terms Of Service.

The software build instructions and video are good, but there are different versions of the operating system, which is under continued development. Some application code that works with one version might not work with another, or there may be unresolved issues with a particular version. I ran into this issue. For example, the operating system I installed originally did not work with the database (I think this has been resolved), and it also would not allow me to connect to my home network using WiFi. There are fixes available, but they require some coding knowledge to implement.

My first suggestion is before you do anything with the Raspberry Pi, is to visit the Raspberry Pi website and look over the hardware and setup information. There are a couple different ways to install the operating system.

You may not have to go the route of installing the operating system and then installing the MVP software. The process called “imaging,” is where someone basically saves a complete image of their hard drive (in this case, the Micro SD card), and you simply install that full image onto your device. It works best when installing to hardware identical to the source, so from Raspberry Pi to Raspberry Pi, the imaging process should not be a problem. The software that you use on a PC to format and image the Micro SD card are reasonably straightforward, and the instructions for imaging the card are good. If everything works as planned, you should be able to image the card, plug it into the Raspberry, and nearly everything will already be ready to go. If you do have to go through the software build steps, you’ll need to do it near a computer with an internet connection. The video was extremely valuable to me; I did an install step and paused the video, and then moved to the next step.

I’ll share a few additional lessons learned before I close.

When you format the card, I recommend the slower overwrite format. Depending on your PC, it could take almost an hour. Imaging the card also may take nearly an hour.

I attempted to use an old VGA monitor with the Raspberry, using an HDMI to VGA converter. It failed to work repeatedly until I purchased a monitor with HDMI input. I was told this is an issue that is known to happen.

Personal Computers contain a small battery that maintains certain data when they are turned off, such as the time. The Raspberry doesn’t do this. It’s supposed to obtain the current time when it starts from the internet connection, mine doesn’t, and it may be due to the WiFi connection. I searched online and quickly located the way to update the clock. From a command line prompt, the syntax is:

sudo date -s ‘yyyy-mm-dd hh:mm:ss’

If you don’t update the clock, your lights will operate based on whatever time the Raspberry defaults to, and the dates on all of your logs will be incorrect. On a similar note, be advised that although you should set your location and time zone in the Raspberry’s operating system, the MVP software will log your data in UTC (Coordinated Universal Time or Greenwich Mean Time ) which is the worldwide time standard used for aviation and scientific coordination. It was intentionally designed to log this way.

A couple of last quick points. The Raspberry isn’t really considered “plug and play.” Peripherals might work if you plug them in hot, but you probably should connect your monitor, keyboard, and mouse prior to applying power.

Also, be prepared to have your peripherals connected to the Raspberry at your MVP. I had most of the software build completed before I started connecting sensors and controls, but you can’t run the validation script without those devices connected, and without the validation script, it won’t set up the link to make your data accessible over the network. And every time I shut it down, I have to again manually update the date and time. If you’re lucky, your Raspberry will update the date and time automatically.

The software was the most challenging part of this project for me. If you run into problems, don’t give up. Check the internet, there are forums dedicated to the Raspberry, and you can also post your issue in the appropriate thread on the openag forum, and someone will likely assist you in solving it. Also there will soon probably be a FAQ on the openag forum; your issue may have already been answered.

Timestamp of data in Couch DB

One Last Thought

I’m not one of the main contributors to the $300 MVP Food Computer project, but I am willing to help when I can. If you have a question, feel free to post it here, or send me a private message through this forum. In matters of software I probably won’t be of much help, although I’ll help if I can. If your question relates more to the physical build of the chamber, or soldiering, or hooking up circuits, I probably can help more with that. I’m especially willing to help if you have questions about the version of the MVP that I have built; hopefully my posts have been clear, but I’ll gladly clarify anything that may not be clear to you.

I hope you learn as much and feel as much of the sense of accomplishment in your project as I have in mine. Best wishes to you on your journey of lifelong learning.


@RadioDave I will continue to edit this post with feedback, so far I’ve repsonded to your first two posts :joy:.

We noticed this issue and have gone back to the original 2 wire 12-volt power supplies on the BOM.

The link for Zip Ties in the BOM now is a variety pack with several sizes.

This is something we’ve thought about a decent amount. Honestly, it’s very difficult to find plastics intended for hydroponic usage. The commercial grade (2’x2’, 2’x4’, 4’x8’) reservoirs sold at hydro stores are ABS plastic with a UV resistant coating. I also have read that HDPE tends to be the safest/best for these applications (often used for residential water lines). It is important to note that “Food Safe” and “Food Grade” are also not the same thing as explained in this article and neither of these are actually standards related to hydroponics. According to the Hefty website the plastic we are using is PolyPropylene (couldn’t get more detail beyond that).
According to the article you provided there were no known issues from PP, in this article I’ve referenced previously it does mention a few, but they rate it as “relatively safe”. This is another article on plastic types for hydroponics you may find helpful, they also say that PP is acceptable

Interestingly from the two high-schools that are building MVP’s the LEDs have been a hit! I had a similar reaction and previously had gone with a similar 4 fan pack on Amazon. I actually had to change the documentation 2-3 times because every single computer fan we used was different. That being said I think you’re correct that an improper connection isn’t the end of the world.

Good notes about tools and equipment. I agree with you completely, we went with the keyhole saw as a recommendation to try to make it as accessible as possible.

Our next iteration will have an enclosure for all of the “Brain” & power supply/wiring.


Today I harvested enough leaves from my first grow to allow my wife and I both have a salad. Here’s what my first grow of lettuce (at about 28 days) looked like today before I harvested part of it:

I still have a little work to do, but my MVP variant is mostly complete. It’s not a cost effective way to grow food, but it gives me control, and it may help contribute to how we get our food in the future. And I like to build things. I think as they progress, the Food Computer and the MVP will contribute to the body of knowledge that will eventually change how we grow and obtain our food.


The 2” circular bit made the job of opening up the lids so much easier. Would highly recommend it. Plus the end result is really clean aesthetically.