Robotics

FlexGrow: Flexible Growth Chamber (Proof of Concept and Proposal)

Fairchild Tropical Botanic Garden hosted competition in partnership with NASA, Adafruit, Nation of Makers, etc. in order to reinvent systems to grow edible plants on the International Space Station (*heart thumping moment*). The contest seeks to find efficient ways to use three dimensional plant growing space aboard an aircraft, how to maintain plants without human intervention and how to design a fully automatic robotic planting and harvesting system. This is a three year endeavor! Now the first year seeks to address the plant growing challenge.

Phase I of this challenge is to create a proposal that makes effective and inventive use of the available volume on the space craft (50cm x 50cm cube) while also incorporating the necessary features for plant growth such sufficient lighting, irrigation, air circulation, etc. The proposals are sent to Instructables where there are divisions for highschoolers, collegiate and professionals. I entered as a professional since I finish my master’s program in January 2019. Five winners are chosen from each category to continue on to phase II where they need to build the prototype and grow Outredgeous red romaine lettuce.

Now besides the whole challenge with growing in a microgravity environment, the specific constraints of the contest were the following:

  1. Adhere to the volumetric constraints of a 50cm cube.
  2. Contain all necessary features for plant growth
  3. Use 3D space in an inventive and effective manner.

There was nothing really targeting microgravity, so you could choose to design the chamber for use on earth or for use on space. I chose to do a design that could potentially be used on the space station and on earth. A lot of the parameters focuses on using pressure gradients such as wicking for irrigation, using water pressure as structural support and relying on active cooling elements (fan) to distribute the heat gradient within and out of the chamber. Below you can find my Instructables proposals, I formatted it a bit better here. Because I designed, built and wrote the proposal in less than a week and half, some sections are a bit sparse compared to others. You will also find that some areas of the proposal are in an Instructables format (assembly steps). I am currently a finalist in the professional category! Lets see if I get to compete in Phase II!

I. Introduction:

FlexGrow was designed with the following goals:

1. Allow for easy assembly and disassembly via modular and collapsible components to aid in transport.

2. Reduce the overall weight of the growth chamber compared to NASA’s Veggie.

3. Maximize grow space.

4. Utilize only active electronics for airflow, lighting and maintaining the reservoir.

5. Build upon the current plant pillow approach used by NASA.

The overall build of the growth chamber is proposed at a 19.5″ diameter and 19.5″ long chamber. This theoretically allows for the growth of 24 plants simultaneously within 6″ x 6″ grow spaces. The electronics allow for autonomous growing with very minimal maintenance other than assembly/disassembly, seeding, thinning seedlings and harvest. Other maintenance may be done for risk mitigation purposes during operating scenarios where the power is cut, a sensor fails, etc.

In a span of a week and a half, the initial design was formulated, the proof of concept was built and the proposal was written. The proof of concept was created to illustrate certain multi-functional aspects that may have seemed unreasonable with only a CAD file for reference. Due to some problems with the current build, I’ve outlined key improvement areas that could potentially solve the issues I’ve experienced. It is worth noting that the proposal (a bit lengthy) is broken down into the following sections:

– Structure

– Reservoir

– Irrigation

– Plant Pillows

– Light

– HVAC

– Electronics and Software

– Conclusion

In each section you can expect a bill of materials for either the proof of concept or the next proposed build, a tool list, assembly instructions and areas of improvements if applicable.

II. Structure:

The main structural elements within this design are the three 1/2″ ID, 5/8″ OD pex hoops. The pex hoops integrate the reservoir into the support structure of the growth chamber and they allow for the means of irrigation between the plant pods and the reservoir. Not only are the pex tubes light in mass, but they are fairly rigid members that are easy to assemble and maintain.

Materials:

Tool List:

Safety glasses

Bandsaw or Pex Cutter or Utility Knife

Sandpaper or Scissors if using Utility Knife

Tarp

Tape Measure and/or Caliper

1/4″ drill bit

Drill

Flat head screw driver

Sharpie/Marking Tool

Assembly (Proof of Concept Build)

The main purpose for this build was to create a prototype that could demonstrate the feasibility of placing the reservoir within the support structure of the growth chamber.

For simplicity, each hoop has their own label, location, function and materials:

R1: Hoop located at front of chamber; irrigates to 4 plant pillows; 4 T-fittings, 4 pex tube sections, 8 hose clamps.

R2: Hoop located mid chamber; irrigates to 4 plant pillows; 4 T-fittings, 4 pex tube sections, 8 hose clams.

R3: Hoop located at back of chamber; attaches to peristaltic pump for reservoir replenishment; 1 T-fitting, 1 pex tube section, 2 hose clamps.

D1: Hoop attached to R1, used as front door; 1 straight connector, 1 pex tube section

D2 Hoop attached to R3, used as back door; 1 straight connector, 1 pex tube section

1. Assembling Pex Hoops: R1, R2

Enlarge

R1
R1 CAD and Build

Finding the OD of the pex hoops was trial and error. Since my size constraint is 50 cm or 19.685 ” I limited the pex hoop’s OD to a max of 19.5″. The perimeter of a circle with this OD is around 61.25″. I then subtracted the thickness of the pex (5/8″) twice from the OD (19.5″- 1.25″ = 18.25″) to get a perimeter of 57.33″. Effectively this new perimeter gives me a bit of a safety margin due to any sort of play with the T-fittings or hose clamps. Since the T-fitting has an offset around 0.65″ and there are 4 of them, theoretically the length per pex tube (4 per hoop) should be around 13.7″ in length. I would recommend reducing that length to 13.6″ or even 13.5″ sections. It was difficult to create a perfect circle with this assembly and furthermore the pex doesn’t sit completely flush against the lip of the T-fitting when flexing the pex–creating more offsets in our tolerance stack up.

Enlarge

R2
R2 CAD and Build

Once I laid out the pex, I cut accordingly. I then marked out the holes in R1 and R2. These holes allow water to flow through the pex hoops as well as the reservoir. I marked three lines around 3.5″ inches on center with the middle hole at the center of the pex tubing (accuracy doesn’t really matter right now). I then bent the tube sections as much as possible and attached both the hose clamps and the T-fittings. I completed the hoop before clamping down the hose clamps and adjusted the bends accordingly. The hoop looks a little wonky at first. I measured out the OD of the hoops–it should be between 19.”-19.5″ OD. If it was larger, I remove material from different sections of the pex. Once satisfied, I positioned the clamps so that their screws are facing the front of the chamber and then tightened.

Enlarge

83626365_2278111525826136_5573931698309038080_n
Avoid the mistake of drilling 3/8' holes and drilling them before assembling the hoop.

R1 will have holes facing R2. At this point, I drilled 1/4″ holes along the marks I made earlier. These are not thru holes! Use the pictures for reference. R2 on the other hand will have 1/4″ thru holes. I drilled accordingly. You may notice that the holes I made in R1 were originally 3/8″, ignore that (magnets were intended to attach the doors, but that didn’t go so swell).

2. Assembling Pex Hoops: R3

Enlarge

10
R3 CAD and Build

Using the same logic and methodology above, the pex section should be around 56.5-56.75″. I would reduce this to around 56-56.25″ to get an OD between 19-19.5″. Again, trial and error. I marked out the holes similar to how I did for R1 and R2–around 3.5″ on center per hole with around 15 holes. I repeated the same process outlined above. R3’s holes need to be drilled out exactly like R1’s–no thru holes.

3. Assembling Pex Hoops: D1, D2

The straight fitting has an offset of around 0.125″, so the pex section should be around 57-57.25″. No holes or clamps are needed for the doors since they do not act as irrigation. I cut and attached the sections accordingly.

Enlarge

doors
Doors to Unit

Areas of Improvement:

There were problems with irrigation and the reservoir seals. Please see those sections for potential material changes for the next prototype.

Enlarge

84148873_870935063336742_8797507135895240704_n
Problems of using a utility knife to cut pex. Use a pex cutter instead for straight cuts! Or a bandsaw.

III. Reservoir

The reservoir is the innovative feature of this growth chamber. Once it is fully filled, it acts as structural support along with the pex hoops. The main material of the reservoir is 4mm thick PVC in the form of sheets. PVC in pipe form is commonly used in hydroponics whereas in sheet form it is commonly used for IV bags. Since it is flexible, it creates a collapsible structure. While the proof of concept remains clear, the reservoir in the final build should eventually be covered in mylar internally (reflect light) and externally or with another opaque material externally. This will block any light from entering the reservoir, but will reflect light internally to the plants.

One sheet of PVC wraps around one edge of the pex hoops, whereas a secondary sheet wraps around the underside of the pex hoops. This means that the reservoir has a thickness equal to the diameter of the pex tubing: 5/8″. If the reservoir came out to be around 18″ in length and without factoring any additional space from potential elastic strain, the reservoir could hold a volume of maybe 11,000 cm^3 which equates to 11,000 ml. This is more than enough water for 24 plants for a period of 4 days (100 ml/plant constraint). 11000 cm^3 of water has a mass of 11 kg or almost 25 lbm. Doing a quick axial and hoop stress calculation assuming a thin walled cylinder and a PVC tensile strength of around 18.47 MPa or 2680 psi, shows that the PVC sheet can support this weight on earth without tearing.

The major question is the seal. Using JB weld seemed like a good fit for prototyping. It has a shear strength of around 13.7 MPa or 2000 psi–no rating for peel strength. However I did run into issues with waterproofing the seal since the geometry is fairly difficult to work with. If you have a facility with proper ventilation I recommend using something like vinyl cement (thermal based adhesive). McMaster has a 1099 or 4475 option. A heat sealing approach may be a better method, but PVC does off-gas during heat treatment. This is not feasible to perform safely at home during prototyping. Instead, this could be a method used for the final product.

Materials:

Tool List:

Safety glasses

Gloves

Card or Roller

Measuring Tape and/or Caliper

Tarp

Tape

Assembly (Proof of Concept Build)

1. Inner Reservoir:

Enlarge

83914447_614986145993395_452004252311617536_n
First sheet attached to R1.

Enlarge

83984167_2480079902239816_3354549524679884800_n
R1's holes should be facing towards R2 to allow for water flow.

The linear dimensions of the first PVC sheet are 62″ x 20″. This gives a few inches on either end to adhere the sheet to the pex hoops. I started out adhering the PVC sheet to R1’s inner diameter (remember, R1’s holes should be facing the position where R2 is placed). I did this by taping the PVC sheet to the hoop while applying tension to the sheet in order to remove potential air bubbles or air gaps. I ignored the fittings and the hose clamps during this process.

Enlarge

83888652_793944374442628_8479890619274100736_n
Ignore fittings for now.

Once I did this through the entirety of the hoop, I moved on to taping the inner reservoir sheet around R2’s inner diameter. R2 should be around 8 or 9 inches on center from R1 with its pipe clamp screws facing R1. I repeated this for R3 (remember, R3’s holes should be facing R2). Afterwards I taped the extra PVC sheet along the longitudinal axis while still applying stress to the sheet to reduce air gaps.

Enlarge

84466064_526237284661574_9210222204204089344_n
Inner reservoir sheet pre-adhering.

Enlarge

84182048_211717283319534_8499761989843681280_n
Inner reservoir sheet pre-adhering.

Once happy with the shape, I moved on to gluing. The JB weld has a 1:1 mixing ratio, so I squeezed that into a styrofoam cup, mixed it together and began applying the glue between R1 and the inner reservoir sheet. I did this after removing a few pieces of tape at a time around the hoop. Using a card to smooth out the air bubbles/gaps, I then re-taped down the sheet. The JB weld has a 15 min working time. After waiting an hour or so for that joint to harden, I moved on to gluing the inner reservoir sheet to R2’s pex sections in order to keep it in place. I then repeated the same process I did to R1 for R3.

Enlarge

84048622_485800815416134_7394480739225436160_n
I used chairs to help hold the inner reservoir sheet in tension while applying adhesive.

Enlarge

83653532_476527616320918_7559784628765065216_n
I used chairs to help hold the inner reservoir sheet in tension while applying adhesive.

2. Outer Reservoir

The linear dimensions of the second PVC sheet are 64″ and 20″. Since this is wrapping around the outer diameter of the pex hoops, it’s important that there is enough extra PVC sheet available in case bunching occurs. I repeated the same taping process I did for R1. I did not tape R2 or R3.

The next stage is to glue the PVC fabric together around R2’s fittings. This will then allow access to the T-fitting by cutting a hole around the fittings along the inner reservoir’s sheet. Once I was sure of the outer reservoir’s sheet placement, I applied glue around the fittings and the adjacent pex. I then pressed the outer reservoir’s sheet to the inner’s. See the pictures for more detail. Once I did this for all of the fittings and allowed the adhesive to set, I moved on to taping R3. The outer reservoir’s sheet should wrap over the inner reservoir’s sheet thus allowing for a PVC to PVC bond. At this point I can repeat gluing the fabric around the fittings for R3. Once the glue has set, the outer reservoir sheet can be adhered to the inner reservoir sheet over the pex tubing. It is important to not seal any of the sheets between the holes that were previously drilled.

Enlarge

84338953_124942315447152_3679435754485841920_n
Some air bubbles were present in this build.

This same procedure was done for R1. Once the glue dried, I poked a hole in the plastic directly under the nozzle of the barbed hose fitting (inline) and stretched the plastic so that only the barbed hose fitting protruded out. I made the mistake of cutting out the hose clamps as well, which were difficult to fully seal. If worried about the seals around the fittings, extra patches can be made and adhered to the inner reservoir where these holes are located.

Enlarge

84705739_184136725981221_3811299771688157184_n
Reservoir fully sealed.

Enlarge

83846224_1400845140089537_6500592324307845120_n
Compresses as expected.

3. D1 and D2

Attach a PVC sheet to D1 and D2. Allow a little 1″ linear lip form over the pex tube so as to help seal the growth chamber. This will serve as the doors of the unit. Specific cut outs will be made on D1 to attach the vents and the display whereas metal plates will be attached to D2 to allow for the fan, pump and any other electrical equipment such as relays and power supply to be attached. Doors will be attached to the chamber via velcro.

Areas of Improvement:

Since I had a relatively hard time getting a proper seal where the reservoir interfaces with the pex, I’d recommend that for the next prototype, marine grade sealant is placed directly next to the seams after the glue adheres the reservoir sheets. For the inner reservoir, the sealant should be placed directly along the inner diameter of the pex hoops so that is it compressed by the inner reservoir’s sheet. The same can be done around the outer diameter of the pex hoops directly where the outer reservoir sheet compresses the pex hoops. Another layer of glue after the sealant may be necessary so as to keep that keep that seal compressed once the reservoir is properly filled up and potentially pressurized. This method may also be necessary in the areas where the fittings are sealed away from the reservoir. Before testing the seal, the irrigation materials should be added.

Enlarge

84119795_2398938757085412_7562378728882307072_n
Removed holes for fittings too early.

Furthermore, the pipe clamps were a bit large and pointy. The ends of the clamp jut and out could be hazardous. Using 1/2″ pex crimp rings could help in reducing pointy objects in the reservoir as well as allowing for a nicer/easier seal between the the PVC sheets around the fittings. A crimping tool will also be needed.

Enlarge

83755539_263216064643201_7196838795373707264_n
Pipe clamp edges jut out.

Furthermore, due to the glue I’ve used, I am hesitant to eat anything that is grown in the proof of concept. The sealant as well as the adhering method/material will need to be chosen (for the final product) so that it is food safe.

III. Irrigation

The irrigation system attaches the reservoir to the plant pillows. This also where the materials used in the proof of concept diverges from the proposal. I will only be attaching the bill of materials for the next proposed build. It is at this point that the materials proposed in the next build get a bit expensive due to the unique size constraints I have.

Materials

Tools

Scissors

CAD Assembly

Enlarge

irrigation_cad
CAD of irrigation

At this point I’d recommend moving to a T-fitting with a 1/4″ inline reducer. It may be tricky fitting a 1/4″ ID hose over the inline reducer; so the tubing may need to be sealed into place and then crimped on. Using 8 push to connect on/off valves (1/sleeve) provides a cleaner attachment between the pillows and the reservoir and allows for manual interfacing. Tubing that has been bonded with the plant sleeve (potentially a heat seal) can be attached to the on/off valve (attachment not shown in CAD). Within the plant sleeve’s tube is a large wick. This method seems like the best method for the current reservoir configuration because once the reservoir is fully filled, the water will want to travel from high pressure to low pressure. The wick will help regulate the flow of water to the plant pillows via a passive approach.

Enlarge

irrigation_cad_2
CAD of irrigation

R3’s T-fitting does not have an on/off valve. Instead it attaches directly to the peristaltic pump via a reducer (1/4″ ID to 1/8″ ID). Along R3 will also be a pressure transducer used to regulate the water pressure within the reservoir.

Assembly (Proof of Concept Build)

Enlarge

83980553_208819460243888_8727153978153893888_n
Proof of concept with 4 out of 8 valves attached

Enlarge

83356546_490066351707726_7657491561336274944_n
Proof of concept with 4 out of 8 valves attached

I had a bit of difficulty with the attachments in the proof of concept. The hose did not sit well on the pex fitting. I ended up attaching a 3/8″ ID hose to a 1/2″ ID hose and via elbow grease and an adhesive. I applied a silicone sealant over the T-fitting and then attached the 3/8″ hose over the T-fitting. Afterwards I attached a 1/2″ to 1/4″ ID reducer to the 1/2″ ID hose. I used 1/4″ OD Lemoy ball valves since they were cheap and had push to connect ends. At this point it was obvious that I should have purchased a 3/8″ to 1/8″ ID reducer since the Lemoy valves required a 1/4″ OD tube. Luckily I had a 1/4″ OD, 1/8″ ID tube with an A scale durometer. Straining one end of the tube created a tight, but shallow seal over the 1/4″ ID reducer. Another 1/4″ OD tube was fitted into the other side of the Lemoy ball valve and this ran down to the plant sleeve where a 1/8″ barbed hose fitting was attached. Another 1/4″ OD hose was attached to the other end of the fitting and this is where the wick was located.

Enlarge

83489319_983127765471652_7469607929209946112_n
Poor selection of fittings.

R3 does not have an on/off valve. It was directly attached to the peristaltic pump, which is electrically controlled by the growth chamber’s logic.

Once the irrigation is attached and the valves are shut off, the reservoir should be tested. I tested my reservoir in my bathtub. I could only fill it with about 2L of water before I noticed leaks. Once the leaks are secured, the reservoir should be fill tested manually a few times to allow the PVC sheet to strain a bit, thus checking the integrity and compression of the seals.

Enlarge

83645092_2543484095864673_1506282242303852544_n
Irrigation testing

Enlarge

83468671_2781604235284151_590911972929699840_n
Filled with water!

Areas of Improvement

While the proof of concept’s irrigation worked, it was difficult to fill the reservoir due to the bad seals. Using less irrigation equipment is crucial for proper maintenance. While I struggled with the fittings I had chosen, I was able to fix the problems in CAD. However, my proposed build solution for irrigation is expensive. Therefore more prototyping and research may be necessary to find cheaper solutions for the next build. This could include changing the size of the pex to utilize smaller and cheaper quick connect valves like the Lemoy ball valves I used in my proof of concept.

IV. Plant Pillows

I really like the original plant pillows developed by NASA. I think they are necessary for the way I’ve designed the growth chamber. The plant pillows are not only collapsible, but they are light in weight and allow for expansion when the roots grow larger or when the grow medium is wet. Essentially 24 plants are grown in a single cycle towards the center of the chamber where the light source is located. It is important to note that when the plants reach full height (15 cm), each plant will be only a couple of centimeters radially away from the light source in the current configuration I have. This is assuming that the plant pillows swell to no more than 2 inches in height.

In order to reduce the amount of irrigation equipment I needed, I decided to place three plant pillows into one sleeve. There are about 8 sleeves in total and each sleeve is around 18″ x 6″ allowing the pillows to be at largest, almost 6″ x 6″. Total height of the pillows and sleeves can range between 0.25-0.5″ under full compression (no solid media) and up to 2″ at full expansion (with solid media). This allows for a grow footprint slightly over the 15 cm x 15 cm per pillow, but limits the plant height to the estimated 15 cm due to the light placement. As mentioned in the irrigation section, each plant pillow has a 1/8″ ID tube that has been bonded to it’s inner material (PVC–same as reservoir). Within each tube is a wick, which travels under the open mesh end of the pillows.

Enlarge

84332794_187108595730025_8668934133088518144_n
One out of eight plant pillows on reservoir.

It is possible to incorporate the necessary irrigation per plant pillow by adding additional pex hoops or more fittings per hoop (thereby eliminating the sleeve). However my current proof of concept and my next build will focus on incorporating 3 pillows per sleeve.

In my proof of concept, each plant pillow and its sleeves are made from 4mm thick PVC sheets. Each of these pillows are sealed with an adhesive and have a mesh backing in order for the water from the wicks within the sleeve to reach the growing media. Each pillow has a smaller mesh surface and a nozzle to allow for placement of and protection for the seeds. This configuration means that the seeds will sit in the middle of each pillow and can be capped off fully or partially from a light source during germination. Partial light block off may be preferred due to the lack of gravitropism on the ISS; this is assuming that the root growth will orient away from the partially exposed light. Caps can later be removed during the growth stage. No sensors are equipped within the plant pillows.

I have built one sleeve and three pillows as the proof of concept. In the final build there should be eight sleeves and 24 pillows for a single crop cycle. It is expected that the crew who is maintaining the chamber participate in seeding and thinning the plants. It has not been determined if the prototypes will be reusable for another crop rotation, but the final product should allow for reuse.

Enlarge

83729158_627028141460318_5498837330464604160_n
Prototypes!

Material:

Tools:

Scissors

Needle

Tape

Cup for mixing solid media

Assembly (Proof of Concept Build)

1. Building Sleeves:

Enlarge

84078451_2730479807029164_7139806195233062912_n
Making the sleeve.

The sleeves weren’t too hard to build. The overall area of the sleeves are 18″ x 6.25″. I built the sleeves from a single sheet of PVC. The cut out area is around 18.5″ x 12.5″. Once the sheet was cut out, I folded the sheet in half so that it was now 18.5″ x 6.25″. I then laid out a 1/4″ on each edge (excluding the bend) in order to get a rectangle with an area of 18″ x 6″. I glued down each 1/4″ offset perpendicular to the bend (see picture for reference) making sure that the glue did not run past the marks. Once sealed, I bent the glued edges inwards and glued them down again. I then applied clear masking tape to keep the bends in place while the glue cured. There should be one open side remaining.

2. Building Pillows

Enlarge

84330184_113599270094698_3765612701322051584_n
Front of plant pillow with fine mesh attached. Seed will be placed on mesh.

Each pillow has an area of approximately 6″ x 6″.They were made similar to the sleeves, with a single sheet of PVC that was approximately 12.5″ x 6.5″. I repeated the same process above with the 1/4″ offsets and gluing until I had a 6″ x 6″ pillow (with one open side).

Enlarge

83877250_605025127004233_429154987642191872_n
Back of plant pillow. Large mesh allows for solid media to be irrigated by wicks found in the sleeve.

Once cured, I laid out 0.75″ from each side. I cut out the inner rectangle and placed a 6″ x 6″ nylon mesh over the opening. I glued down the mesh and then taped down the edges. Flipping the pillow over, I laid out a 2″ x 2″ square at the center of the pillow. I cut this out and then placed a 2.5″ x 2.5″ fine mesh over the opening and glued/taped it down into place. I adhered velcro around the fine mesh’s perimeter. I also adhered velcro along the open side of the pillow along the 1/4″ offset.

Enlarge

83923528_817151438748895_1582231381654634496_n
Velcro used to seal plant pillow. Allows use to re-access pillow if necessary.

3. Interfacing Pillows with Sleeve:

Enlarge

83944451_175702120368673_2531351194803109888_n
Back of sleeves with pillows integrated (pre-wick addition).

Since there is a bit of a tolerance stack up, the pillows may fit snugly into the sleeve. It’s important to mark the pillows to know which one goes in first versus last. Once I found the centers of the pillows and laid them out along the sleeve, I made an incision with a needle. Once I stretched this hole slightly, I wedged part of a 1″ diameter nozzle through the inside of the hole and taped it in place (this nozzle was taken from almond milk containers).

Enlarge

82388680_513535166265432_1824062891100209152_n
Nozzle attachment.

After tracing out the lines of the velcro located on the pillows, I could attach similar length pieces of velcro to the inside of the sleeve so as to secure the pillows in place. This also makes sure that any seed placed on the fine mesh is not lost within the pillow.

Enlarge

83372689_1087897811553383_1805305274269958144_n
Nozzles placed on front of sleeve with velcro additions.

4. Apply Mylar to Sleeve:

Enlarge

83618809_537005473688703_1202626087381630976_n
Attaching mylar to sleeves.

Once the pillows have been interfaced with the sleeves, mylar can be taped in place on the sleeve. This makes sure that no light enters the plant’s reservoir and all the light being reflected onto the pillows is reflected back to the plants. I only placed mylar on the tops and sides of the sleeve. Leaving the bottom exposed gives me the opportunity to check the plant pillows if a problem arises during germination/growing. This is important since there are no sensors in the pillows that allow for feedback based on moisture control, pH, electroconductivity, etc. Once the mylar sleeve is taped on, the velcro can be adhered to the mouth of the sleeve.

Enlarge

83738357_2532146150366564_2761682899248873472_n
Nozzle exposed on mylar surface.

Enlarge

84142911_644673056305952_4597599723992383488_n
Velcro attached to mouth of sleeve.

5. Testing the Fine Mesh:

Enlarge

83920809_2609154679197925_5021941865771958272_n
Day 1: Chia seeds laid on fine mesh to test if roots can grow through the mesh.

It was important to see whether or not the mesh I chose would allow for the seed’s roots to pass through. I put together a dish of solid media (coconut coir and vermicompost). I placed the mesh on top of the solid media and uses a tablespoon of chia seeds as my test group (they germinate very fast). I sprayed the area with water every day and then placed a lid over the area during germination. Within 2-3 days, the roots sprouted through the mesh. I could flip the mesh over and the plants stayed in place. This mesh proved perfect for germination and will make it easier to germinated the seeds in an environment with gravity.

Enlarge

83943706_2510441455838053_5809653470747164672_n
Day 3: Chia seeds germinate.

Enlarge

83829671_796681854168328_7685899205176459264_n
Day 3: Chia seed roots can be seen growing through the mesh.

Enlarge

83824555_2550819098522925_7677151967407243264_n
Day 3: Chia seed have anchored to the fine mesh.

Enlarge

84671693_185283069226347_1804184880511188992_n
Day 4, chia seeds are growing nicely on mesh.

For the plant pillows that are upside down, it may be important to use cotton balls (good wicking material) to hold the seed in place while the pillow is upside down as well as gather water from the solid media. The caps on the sleeves will help hold the seed in place as well as block out the light fully or partially until it fully sprouts.

While my experiment worked, it is important to note that I sprayed the mesh daily. A different approach may have been to place a wick directly under the grow medium to see if water passed through the grow medium and reached the seeds. In case that fails, an extra wick may need to be routed within the sleeve to the surface of each pillow’s fine mesh. This will allow for direct water transport to the seeds.

7. Interfacing Wicks and Irrigation:

Enlarge

s12
Wicks added to sleeve.

In my proof of concept, I used a 1/8″ barbed hose fitting to attach the hose to the pillows. Attached to this fitting was another 2″ hose that had the primary wick centrally located within. This wick spread throughout the length of the sleeve. Under each pillow was an additional wick that attached to the primary wick. I do not plan on using the 1/8″ barbed hose fitting in my next build.

8. Applying Solid Media in Plant Pillows

Enlarge

1-50
Example of solid media recipe.

The recipe for the solid media will most likely be the following (subject to change):

1. 1/2 part red clay

2. 1 part peat moss

3. 1 part water

4. 1/2 part perlite

5. 1/4 part vermicompost (fertilizer)

This is a simple recipe that can easily be formed and solidified together after mixing and a day of drying (depending on the climate). I plan to form the recipe into a 5.75″ x 5.75″ square with a 1/4-1/2″ thickness. It is a modified seed balls recipe. The red clay helps hold the mixture together for easier placement within the pillow (as well as during the transport of the pillow) whereas the peat moss (main growing material) will expand to allow for more root growth. The perlite will help keep the media nicely aerated (avoiding wet feet problems) and the vermicompost will act as the nutrient layer.

I’ve had a bit of success with this recipe (subbed in coconut coir for peat moss) in germinating and growing chia seeds and sunflowers.

7. Attaching Sleeves to Reservoir

Enlarge

83947880_644215316420901_3807757463051042816_n
Attaching plant pillows to reservoir.

The sleeves are adhered to the reservoir via velcro. This method will only secure the sleeves along 2 axes. If more restraints are needed, straps can be added to secure the sleeves into place (particularly for the sleeves along the top of the chamber).

Areas of Improvement:

The sleeves restrict the movement of 2 out of 3 of the pillows. If something goes wrong with a pillow on the opposite end of the sleeve’s access port, then the user needs to remove the first two pillows to access the third. This is not possible once the plants are growing. Furthermore because the sleeve’s mouth is velcro, it’s not fully sealed. Water could ingress out, in case too much water is absorbed by the wicks and the media. One solution to this is to seal all three pillows in the sleeve, but this reduces all access to these pillows. It also needs to be seen whether or not 100 ml of water will be made available to the plants per day with the current wick set up. Again, this will come down to prototyping and experimenting.

V. Lights

The lights are one of the main challenges I faced during this design. In order to feed the plants between 300-400 𝜇mol/M2/s within PAR (400-700nm), I had to find ways to convert the manufacturer’s illuminance (lux) rating to PPFD. I found a website that could convert the LED’s illuminance to PPFD, but I cannot confirm the validity of my results without a PAR meter (which I do not have). So to the best of my ability, here is how I sourced and designed the lights for my enclosure.

Materials:

Tools:

Drill

Drill bit set

Tap

LED Choice:

LED Supply offers 24 Volt LED Strip Light Engines in 12″ x 1″ strips. The LED strip has 48 Nicha 757 diodes placed on a rigid, aluminum strip that can be cut into three inch pieces allowing for the lights to be daisy chained together. It requires only 24 VDC since is has drivers already on the board and it works with PWM dimming and has reverse polarity protection. I have attached the specifications as a photo and the data sheet is located here.

In order to estimate the amount of PPFD I can output, I chose the following configuration:

1. 3x 12″ strips of red light (assuming 650nm wavelengths)

2. 1x of 12″ strips of blue light (assuming 450nm wavelength)

3. 1x of 6″ green light (assuming 520-535nm wavelengths).

At a 24V input, each 12″ strip will run at 7.68 Watts. At this wattage LED supply predicts that the strip will output 870 lumens. If the surface area of our inner cylinder (where plants are located) is around 1164 in^2 or 0.7509m^2 (assuming our worse case radius is 9.75″ and our worse case length is 19″), then our illuminance for one strip becomes 1160lx (870 lm/0.7509m^2). For five strips, our illuminance becomes almost 5800 lx.

In order to find the PPFD of my red LEDs, I plugged in 3475lx with a Monochromatic Red LED 650nm spectrum to get a PPFD of 267.06 𝜇mol/M2/s. To find the PPFD of my blue LEDs, I plugged in 1160lx with a Monochromatic Blue LED 450nm spectrum to get a PPFD of 134.04 𝜇mol/M2/s. To find the PPFD of my green LED, I plugged in 580lx with a Red + Blue + White LED 450+650nm+3500K spectrum (there was no green LED option available), this resulted in 15 𝜇mol/M2/s. If all lights are run at full capacity, then I can expect a PPFD of around 415 𝜇mol/M2/s, a bit over the prescribed amount. This could potentially be adjusted by dimming the LEDs through adjusting the PWM.

I chose to incorporate more red LEDs due to the ratio of lights used in NASA’s Veggie (ratio of 12 red: 3 blue: 1 green). Further research also showed that more red light is necessary for plant growth. Advanced Led Lights mentions that red light is important for stem and leaf growth and can also regulate any flowering, germination or dormancy periods. The website also mentions that blue light affects chlorophyll production and leaf thickness, but needs to be regulated in case of stunting certain plant species. Therefore less blue LEDs are incorporated throughout the LED built. Green light is important in order to help support the leaves in the lower canopy since its wavelength can penetrate through the upper leaf canopies, therefore I tried to incorporate a bit of green light in the light set up.

Since my plant pillows are wrapped in mylar, the emitted light will be reflected off the surfaces of mylar, thereby increasing the efficiency of my lighting.

CAD Assembly

Enlarge

12
CAD of light within chamber--no guy wires shown.

Enlarge

aluminum_light
Rough render of the light build

My lights will be daisy chained together and attached onto a hollow 1″ x 1″ square, aluminum tube that is around 18″ in length. The LED strips provide thru holes. After drilling and tapping these holes along the aluminum tube, I can attach the LED strips in three inch sections to the hollow tube via 1/4-20 screws. See the CAD photos for colored light placement (ie. red, blue, red, green, blue). Since I expect to have about 15″ of light strips on each side of the aluminum tube, it can be expected that I will have a max of 1.5″ extra space on each end of the tube for mounting holes (assuming no spaces between LED strips).

Enlarge

light
CAD of light shown with attachments in chamber.

In order to mount the light at the center of the chamber, four eyebolts with shoulders will be placed on either end of the chamber–attaching directly through R1 and R3 (see CAD for details). Since these holes are accessing the reservoir, they will need to be sealed with something like marine grade sealant. Additional holes in the rectangular tube will need to be made so that lifting wires can be strung through, thus holding the light in tension at 4 ends on either side (almost like guying the LED light fixture into place–think of a spiderweb). This configuration allows for the chains and light to be removed when it’s time to disassemble the chamber and collapse the structure. The power leads of the light will lead to the back door (D1) where the power supply and relays will be supported.

VI. HVAC

Other than using the fan to cool the chamber, I don’t plan on adding any external heating, cooling devices or misting devices. I do plan on adding a few humidity and temperature sensors throughout the chamber and outside the inlet vents in order to provide a closed loop cooling system.

Materials:

Enlarge

hvac
Fan shown next to electrical box.

Tools:

Chop saw or some sort of cutter

Screw Driver

Scissors

Tape/Glue

Drill

Drill bit set

Fan Choice:

Outredgeous lettuce is a cool weather plant and grows optimally at around 60-65 deg F weather. It was estimated that the temperatures in the ISS range from around 65-80 deg F. If we estimate the following:

1. Volume of air within the growth chamber: 5680 in^3 or 3.287 ft^3 (worse case diameter: 19.5″ and worse case length: 19″)

2. Exchange rate of air: Every 3 mins

Then the baseline airflow we need in a fan is around 1.1 CFMs.

Now we consider all of the wattage produced by the lights, pumps, microcontrollers, carbon filters, etc. If this wattage were converted directly to heat, then I need a certain airflow rating to cool my equipment (see next step for equipment ratings) to maintain my chamber’s temperature. Again we can estimate the following (all wattages pulled from data sheets):

1. Lights: 40W (8W/ft, 5ft of lights)

2. Pump: 3.6W

3. RPI (worse case): 12.5W

4. Fan: 5W

5. Display+ Decoder: 2.5W

This sums up to a rate of 63.6W worse case scenario. Round this up to 65-70W to account for any heat released by the additional sensors.

If I want my chamber to stay around 60-65 deg F, and I accept a temperature rise of no more than 5 deg F (ie. inlet air temperature at 70 deg F), then by using the following heat dissipation equation ((3.16* Internal Heat Dissipation)/Allowable temperature rise in deg F), I can expect my necessary airflow to be upwards of 41.08 CFM–depending on how crucial my allowable temperature rise is. This equation will need to be modified for use on the ISS since it incorporates conversion factors, specific heat and density for sea level air.

A filter impede fan airflow, therefore it needs to be factored in. Hydroexperts recommends that using exhaust filters should factor in an additional 20% of airflow. If one is using CO2 replenishment, an additional 5% needs to be factored in. 25% of 1.1 CFM is around 0.275 CFM. I can expect to use a fan that provides at minimum 43 CFMs.

I chose a fan that can supply up to 63 CFM for the time being–was chosen for airflow, size requirements (mainly thickness) and cost. Ideally I would want a fan that can output an airflow maybe 1.5-2.5x times my baseline use (depending on how one defines their safety margins and their budget–some would argue this safety margin needs to be multiplied by one’s worst use case). While the final build should have a carbon filter and the usual extension it provides, I will instead adhere a mesh to the front of the fan in the interest of fitting my size constraints.

CAD Assembly:

Enlarge

front_HVAC
10x 1 inch vents shown around display screen.

The fan will be attached to the back door (D2). Two metal bars run along the face of D2 and are attached to the pex hoop via bolting. A cutout will be made in the PVC sheet to allow for placement of the exhaust fan on these metal bars thereby positioning the fan directly at the center of the chamber. You will notice the black box of electronics next to the exhaust fan. It is important that these electronics are located on the outside of the back chamber rather than the front of the chamber. This is done so that the air being brought in the chamber is not affected by the electronics’ temperature.

Ten 1 inch diameter louvre type vents will be located on the front door (D1). They can be adhered to the PVC sheet (metal to plastic bond). The exhaust fan will essentially pull air through and out the chamber. In the final product a carbon filter should be attached to the fan so as to filter out unwanted debris. No filters will be attached to the vents. Additional attachments can be made to D1 if one wants to inlet a supply of carbon dioxide. In this build I will not be attaching carbon dioxide inlets because I am assuming that the air being pulled through the chamber will have the necessary amount of carbon dioxide for the plants.

Enlarge

Back_Hvac
Fan shown next to electrical box.

One temperature sensor is located on D1 to measure the inlet air. Another temperature/humidity sensor is located mid-chamber and the third temperature sensor is located on the inside of D2 (back of the chamber). These three sensors give me a spectrum of temperature throughout and outside the chamber. The middle temperature sensor will be the only sensor configured to give humidity readings.

VII. Electronics and Software

Materials:

1. Wiring:

As mentioned before a majority of electronics will be located on D2 (back door). The black box in my CAD model represents the electronics enclosure. It is important that all sensors/active devices have quick attachment connectors such as JST connectors. This will allow for some sensors/active devices (ie. lights) to be removed from the chamber easily without having to reconfigure the wiring after assembly or before disassembly. The wire routes to the various temperature sensors and LED lights will be harnessed together in heat shrink and then integrated along the inner reservoir sheet via velcro. D2 should remain attached to the chamber due to the wiring configuration.

It is important to note that the inlet temperature sensor and the display screen are located on D1 (front door). Their wires will be routed along the outside of the chamber and then within the pex tube positioned near the outside bottom right of the growth chamber (see CAD for details). Precaution will need to be taken when opening D1 so that the wires from these electronics are not stressed or displaced. The tubing that allows for the wires to pass through is adhered to the outer reservoir sheet via velcro and can be removed upon disassembly. Unfortunately I will not be able to provide a circuit diagram with this submission.

2. Power Supply:

Enlarge

electronics_config
List of electronics and other characteristics.

I have attached a list of my sensors and active electronics with various ratings. Since my lights have a 24V rating, I will need a power supply than can supply this voltage rating. The rest of the components such as the fan and the RPI will need DC power converters. I can use this DC power converter between the Power supply and the RPI (24Vin-5Vout). I can also use this DC power converter for the fan and the peristaltic pump. Since the temperature sensors, display screen and the pressure transducer attach directly to the RPI, no power converters are necessary there. The power supply I will be using provides more than twice my power load. I am leaning towards using this brand of power supply because I’ve used a 24V 350W version in a previous hydroponics build and it seemed to work great.

3. Relays:

I will use relays to control my fan, pump and lights (see bill of materials). The relays I chose allow for a 5VDC input and has a maximum output of 30V. I also chose non-latching relays that are quite popular among RPI and Arduino users.

4. Software, Miscellaneous:

By using relays, the RPI could then operate a timing circuit to turn the lights on and off every 12 hours so. The fan will turn on and off depending on my inlet temperature and current temperature of my growth chamber (taking the average of the two temperature sensors within the growth chamber). Airflow could also help manage humidity, but I am not introducing extra steps to maintain humidity. It is expected that the duration of peak airflow will increase when the plants are near to harvest. I expect that the chamber will become more hot/humid when the plants are at their largest since they will become obstacles in the path of convection.

The water in the reservoir will be regulated by the pressure transducer and the peristaltic pump. This closed loop circuit is necessary so that the peristaltic pump does not over-pressurize the reservoir or under-pressurize the reservoir thereby risking the integrity of the structure.

The display screen is used as a means to access any recorded data (water level pressure, temperature, humidity, etc. It should also allow for manual control of the various active electronics such as the pump, lights and fan via a basic GUI. The growth chamber will have four main states: OFF, ON, MANUAL. The on state is when the machine is growing the plants. This is when the lights are running on a cycle, the sensors are continuously pumping out data and the active elements are responding accordingly. The manual state allows the user to shut off and disconnect the sensors/active elements that they see fit. The main elements that should be activated in this state are the display screen, peristaltic pump and the pressure transducer. The manual state allows for the reservoir to be filled or emptied depending on the mission: assembly or disassembly. The software and electronics are subject to change during the prototyping of the next build.

VIII. Conclusion

In conclusion, I believe that my FlexGrow is relatively innovative and could be a potential design for growing plants in microgravity. I believe that I have succeeded in creating a collapsible structure that can be easily assembled and disassembled (once fully built) and that can grow a large number of plants within the size constraints given. While there are still risks and questions with certain components and interfaces in this design, I believe that it offers a lot of potential as to what the next ISS grow chamber could look like. Thank you for taking the time to read my submission. Mahalo!

Author

smundon@bu.edu

Leave a Reply

Your email address will not be published. Required fields are marked *