High Tunnel Gantry System with Transport Cart and Automated Row Cover System to Assist Small Farm Production, Scalability, and Profitability

Objective 1 - Design 

We conducted a rigorous engineering design through two phases - a preliminary design and a critical design. After the design, prototyping, and development stages were completed, Rigorous opted to modify the proposal scope to push forward a second round of design and development to incorporate the lessons learned. Details of the updated design are in the Build & Evaluate section of this report.

As we conducted designs, we reached out for feedback from our target community. We worked with the members of our Farmer Design Review Board, conducted one one-on-one interviews with 21 regional farmers and two greenhouse manufacturers. 

We began designs with the following foundational requirements and expectations. Additional requirements were gathered during the design and development processes.

1. The high tunnel contains truss-style supports and extra tall side walls, largely due to the snow load and tomato trellising.
2. The high tunnel has access to AC power.
3. The high tunnel has access to a WiFi network.
4. The Transport Cart must be less than 55 pounds.
5. The Transport Cart must fit standard harvest bins.
6. The Transport Cart must have an easy on/off connection to the rail.

Preliminary Design - sare_gantry_pdr_r03

At the preliminary design, we presented the concept and drafts of the mechanical, electrical and software designs. We broke the presentation into sections that explained each sub-system in detail such that the Farmer Design Review Board (FDRB) could provide feedback and critical direction.

Trolleys - Trolleys are passive rollers that allow you to hang and move items on the rail.

Rail System - Round tube hung from the high tunnel to allow hanging systems to move along the rail. The rail system is made from SS20 fence toprail. This is an accessible product available throughout the country. The ends are swaged, enabling them to fit together, and are secured by two screws. The rails hang from Railex brand C-bracket that are supported to the high tunnel crossbars with a U-bolt. We chose the 5.5" C-bracket to ensure ample room for the trolleys to pass between the top of the tail and the bottom of the bracket.

Transport Cart - Hanging cart used to move objects. To keep the weight low, we originally modeled the cart out of aluminum. The review board noted that aluminum is challenging to weld and therefore would not be viable for most home fabrication efforts. They recommended we change to steel. Our original design fit a variety of harvest tote sizes, from bulk crates to bread trays. The review board recommended we keep the cart as narrow as possible rather than fit a wide variety of products. In addition, they recommended we change the frame supports such that farmers could easily load the cart from the front or back without reaching around the side.

Toolbar - Generic attachment to connect various implements between trolleys. The toolbar is a generic attachment point that connects the two rails and spans the width of the high tunnel. The toolbar was originally designed to hang from passive trolleys. We presented two styles - a fixed height model and an adjustable height model. To prioritize strength and simplicity, the review board recommended the fixed height. This model is adjustable at the time of the build, so could fit different high tunnels, but not adjustable once it is built.

Trolley Car - Motorized trolley that moves along the rail to move implements. The trolley car is a motorized box that drives the toolbar along the rail. We presented a concept that pushed and pulled the toolbar. The drive system comprises a NMEA 23 stepper motor and motor controller. We have geared this up almost 10x, bringing the total force to 27 newton meters, which we determined to be sufficient to push and pull the toolbar with implements.

Automated Row Cover System (ARCS)  - Implement designed to automatically roll and unroll row cover. It includes spooling motors at both sides of the roll and a frame that attaches the implement to the toolbar. The system also has a motor box and that motor box includes the same components as the trolley car. We estimated that we could spool a 28.5' wide roll of remay in the 30' wide house.

The concept of operations broke down the steps to set up and operate the system components, including through the user interface to the software.

Critical Design - sare_gantry_cdr_r01

In the Critical Design Review, we presented the updated and final designs, including the mechanical drawings. 

One major change was to the Transport Cart. We designed the frame to fit the dimensions requested and changed the material to steel. The review board noted that aluminum would be preferable in some cases to increase potential load and that a modular/configurable cart would be ideal. They also noted that a subsequent iteration could include a support wheel underneath to further increase potential load. They noticed a design flaw where moisture could enter but otherwise approved of the changes.

We lengthened the toolbar to lower the entire system which is best able to maximize the larger width of the high tunnel at lower heights. The final height of the toolbar will be determined during assembly. 

The trolley car is no longer a free-hanging unit, but is now built directly into the toolbar. The weight of the toolbar and implement will provide better traction for the system. We designed a hitch that is made out of off-the-shelf springs and hardware and 3D printed parts. These attach to the Railex rolling trolleys and connect to the motor to create the drive system.

The major change to the ARCS was that we are now using self-aligned bearings. These will allow for sag in the spool. 

Objective 2 - Build & Evaluate

We constructed a High Tunnel Gantry System (HTGS) for Jericho Settlers Farm. The initial (V1) construction process was divided into phases for rail, toolbar, and implement. 

Upon completion of initial system testing, the team decided to redesign key components of the system, test them indoors, and then deploy them to the Jericho Settlers Farm high tunnel for testing.

V1 Rail Phase

We had originally selected a 100 x 30’ Ledgewood High Tunnel in which to build the system as that was the house in which we had built the first prototype. Because the house had underground heat, the farmers preferred to reserve that house for early warm weather crops rather than late winter greens, so we moved to a similarly sized Rimol. We hung the rail in November, but realized after installation that this house was not level, as is common in many high tunnels. The levelness of the house was not a factor we had greatly considered as slowing and braking were not a problem in the initial prototype. We elected to change houses once again to a 200’ x 30’ Rimol. As our budget only allocated 100’ of rail, we could only hang rail in the front half of the house. This house is also not perfectly level, but the grade was amenable to development of the ARCS. 

We made the rails from 1 ⅜” standard fence SS20 round tubing with swaged ends and Railex hangers which are C-brackets with U-bolts.  We used two rails to span the 30’ width of the high tunnel.  The rails operate independently as a monorail or together as a system.  The rails are installed over the high tunnel’s walking paths such that the Transport Cart can be pushed down the rows without impeding the crops.  

The Railex brackets were hung every 8’. We found that distance is sufficient for smooth travel and overall rail load requirements. Optionally you can hang the brackets at 4’ if the cost is not prohibitive and high tunnel includes 4’ spaced supports.

We tested the trolley system to ensure the system can carry designed loads and move smoothly over junctions. We calculated that the trolleys would need to pull with a typical lateral force of less than 10 pounds. While the trolleys passed smoothly over most of the junctions, it is critical that the swaged end fits snugly into the next pipe. Where there is a gap between outer diameter of the two pipes, the trolleys wheels can get hung up during travel.

V1 Toolbar Phase

The toolbar was made from generic strut channel and extends about 28’ across the high tunnel.  The toolbar provides a standard interface for implements.  The toolbar hangs from both rails, hung by two sets of trolleys. The toolbar carries the ARCS but not the hanging cart.

The unistrut is connected by brackets and hardware. After hanging the toolbar, we speculated that the ARCS would hang too low to the ground so we shortened the vertical bars. 

V1 Implement Phase

The Transport Cart was made from welded steel 1” square tube and rail trolleys. We designed the Transport Cart to primarily carry 10-pound and 15-pound tomato flats. The steel cart weighs about 33 pounds so the system has a maximum load capacity of 225 pounds. The cart can carry twenty 10-pound flats or fourteen 15-pound flats. This cart can also be fabricated in aluminum which would allow the tool to remain unpainted. The aluminum cart would weigh 16 pounds. We chose steel to represent a product that others could more easily make or modify themselves as welding aluminum is a more specialized skill.

The trolley cars were originally made from 3D printed parts, motor, motor controller, drive wheels, sprocket, and chain. The final version of the trolley cars were made from fabricated steel and aluminum. The trolley car includes an embedded computer to automatically control the trolley. The computer ran existing Rigorous software to control the system. 

The ARCS was fabricated from standard round tube, self-aligning ball bearings, spooling motors, sprockets, and chains.  The row cover system spans as much of the 30’ width of the greenhouse as possible, determined by the curvature of the hoops at the height of the implement.  

The ARCS includes two motorized trolley cars to move the hanging implement down the rail.  As the trolley cars move the ARCS, the on-board spooling motors roll and unroll the row cover onto a spool.  

V1 Testing

During initial testing, we tested the ARCS’ ability to effectively roll and unroll the row cover, the trolley car’s ability to move the ARCS, the system's ability to synchronize the two trolleys with rolling motors, and the system's ease of use.  

We assembled most of the ARCS sub-assemblies off-site and put the components together at Jericho Settler’s Farm in January 2022. We first tested the system without row cover. The self-aligning ball bearings had enough play to allow the wide pipe in the spool to roll easily even with a slight sag in the pipe. 

The trolley cars, as originally designed, were not consistently effective in pulling the implement up and down the rail. The motor was mounted to the top piece of the toolbar. A chain connected the sprocket on the motor to a sprocket on the trolley car. When there was any motion in the toolbar with respect to the trolley car, for example, a jiggle as it moves over a rail junction, the distance between the sprockets would contract and expand. This activity would pop the chain from the sprockets. We decided to redesign the trolley car such that the motor was attached to the trolley rather than the toolbar.

Once we redesigned and rebuilt the trolley car, we returned to swap assemblies. We used extra Railex trolleys to make cable trolleys for the extension cord. This allowed us to test running the unloaded system up and down the length of the rail. In addition, we coordinated the spooling motion of the ARCS motors with the trolley car motors. 

We controlled the system through the browser-based app that we developed for the system. The application runs off of the RGS software library with a specific instance and User Interface (UI) for this project. The UI allows the operator to turn the machine on in both automatic and manual modes. As we were testing basic functionality, we used manual mode.

In manual mode, an operator can direct the system to go forward, backward, spool roll, spool unroll, or a combination of linear direction and spin which is required to roll an unroll row cover. The buttons that controlled all of the motors in tandem read “cover” and “uncover”. There is a stop button on every page.

Once the trolley cars and ARCS motors were working effectively, we loaded the ARCS with row cover for the first time.  We found that the added weight of the row cover over the 28’ span of caused a significant bend in the spool.  The bend in the spool caused a large moment on the spool, making it too powerful to rotate with the spooling motors.

We determined that the width of the tunnel was too long for the ARCS to work as expected. We discussed possible solutions and determined it would be possible to operate with a center support dividing two sections of row cover.  To test, we folded the row cover over on itself and rewrapped it around the bar such that it covered about two-thirds of the bar. We then used a strap to hold up the center of the spool bar. The strap allowed the bar to still spin freely while supporting the weight. This worked well and we were able to use the ARCS to roll and unroll row cover for the first time.

We designed a center support system. This required us to cut the row cover into two pieces. The high tunnel had five beds when planted, so we cut the fabric to fit over 3 beds and 2 beds respectively.

We reset the row cover on the ARCS. We were able to roll and unroll the row cover along the full length of the rail. There was a gap between the two sections of row cover. Yet, when the ARCS finished unrolling, it often pulled the row cover taut, closing the gap. We ran the system multiple times and documented the progress with photos and videos.

We then broke the system down so the farmers could till the high tunnel to plant early-season potatoes. We removed the ARCS from the toolbar. The implement is easily disconnected with the removal of 4 bolts. 

Once the beds were ready, we reassembled the system. By this time, the daytime temperatures in the high tunnel were consistently in the 60s and 70s. We believe the temperature and humidity were showing us new system problems that were not apparent during winter temperatures.

The trolley car glides up and down the rail with a rubber wheel. In the warmer and more humid environment, the wheel had trouble gripping the rail and would frequently spin out. This would cause the ARCS to get stuck and not complete the traverse down the high tunnel.

We came up with potential solutions to the rail grip problem. We decided to purchase adhesive sandpaper, hoping the grittiness would be sticky for the rubber.

Once the tape was adhered to the rail, we tried running the system again. Unfortunately, the tape made the grip issue worse. When the motors lost grip and began to spin out, the sand was ground off the sandpaper.  

After further testing, we determined that we could not rely upon the ARCS to automatically cover and uncover the row cover without human intervention. Overall, we made significant progress on identifying the system specifications, designing a solution to meet the requirements, and testing the design in the field.  We found the ARCS design did not meet the required reliability, but through this process, we have identified remaining technical challenges requiring a solution.

We decided to modify the scope of the grant to allow for another round of design and development. We then proceeded to break the system down and try to fix the trolleys off-site in our shop, shooting for a reinstallation in the 2022/2023 winter season.

Transport Cart

We finished the design of the transport cart by incorporating the farmers’ feedback from the critical design review. We sent the design off to a fabrication shop and painted the steel for rust protection. We installed it upon completion.

Development Round 2

In November, 2022, we disassembled the ARCS at Jericho Settlers Farm. We moved the ARCS to the Rigorous warehouse in Williston. We designed and built attachment points for the rail in order to build and test the ARCS indoors before reinstalling it at Jericho Settlers Farm. We did not need to modify the V1 rail, toolbar, or transport cart.

V2 Implement 

We determined that the trolleys were the critical component to redesign. Not only are the trolleys critical for the ARCS, but they would be critical for any motorized or automated implement on the High Tunnel Gantry System.

We hung sections of rail from the Rigorous warehouse ceiling rafters. We then redesigned the trolleys with the following major modifications.

Drive Wheels

There are two drive wheels on the trolley rather than one drive wheel and two guide wheels. In the V1 trolley, the drive wheel was a Urethane roller attached to a gear, belt, and motor drive. The Railex trolley served as the guide wheels and the attachment point to the rail. In the V2 trolley, both drive wheels are Urethane rollers with an inward taper for maximum traction over the bend of the pipe.

Motor Placement

In the V1 trolley, the motor was attached to the toolbar, below the rail. The motor powered the drive wheel, which sat on top of the rail, by a drive shaft, chain and sprocket. The V1 trolley chain would pop off of the sprocket when the rail or toolbar was jostled, as was found to be common when moving over rail attachment points. The V2 trolley design incorporated the motor, using it as both the drive mechanism and also a counterbalance to the larger trolley bracket system. The motor is still connected to a chain and sprocket system, but there are 3 sprockets rather than 2, with one sprocket controlling each drive wheel. 

V2 Testing

We rebuilt the ARCS in the Rigorous warehouse with the new trolleys. We tested the trolleys alone and then with the full weight of the toolbar and ARCS. They performed significantly better than the V1 trolleys. 

We tested the ARCS in automatic mode at the Rigorous warehouse for two weeks. In automatic mode, the software controlled the ARCS and commanded it to “uncover” at 8am and “cover” at 5pm each day. After successful testing in the warehouse, we reinstalled the V2 ARCS in a high tunnel at Jericho Settlers Farm in March 2023. 

After installation, the team began to troubleshoot some failures of the servo motors on the Trolleys. If either motor skipped, the trolley would lose traction on one side, the assembly would become crooked on the two rails. Said another way, as there is one Trolley on each rail, when one Trolley moves and the other does not, even for a moment, one side passes the other. When this happened, the Trolley would at times lose guidance and fall off of the rails. The passive safety rollers prevent the toolbar from falling off of the rail, but the drive mechanism no longer worked without manual intervention.

The engineers determined that the servo motors were failing while passing over the unsupported areas of the rail. The rail is supported every 8’ by C brackets. We chose 8’ increments rather than 4’ increments to reduce the cost of the rail. In the center of the 8’ span, we measured a 1” sag in the rail. When the trolleys were climbing up the slope, the servos were likely to skip steps. We tried different speed and accelerations of the trolley motors and learned that the motor is more likely to skip with higher speeds and less likely to skip with lower speeds. 

The following day, we installed the row cover on the ARCS and we reduced the speed of the ARCS. Even with the reduced speed, the servo motors continued to skip steps when driving up the slope. We determined that a gearbox, different bearings, and/or different motors would be the next design improvement recommendations.

We then tested the row cover spooling. The row cover was traveling effectively for about ¾ of the 100’ span. The row cover was attached to the ARCS in two sections with a center support bracket dividing them. Because the high tunnel has 5 beds, we cut the row cover into 2 pieces that fit over 3 beds and 2 beds respectively. As the ARCS rolls the row cover onto the bar, small variations in the row cover caused the two sides to roll at different rates due to the larger and smaller diameter sizes. 

We uninstalled the trolleys from the ARCS to make design modifications. We installed gearboxes to add torque to the stepper motors. In late March 2023, we reinstalled the trolleys in the high tunnel. With the gearbox and very slow speed, we were able to get the trolleys to run down the length of the greenhouse consistently.

We observed that the ARCS spooling chains, which were on each side of the 28’ span, were at times out of alignment and thus competing with each other. We removed one of the chains on the ARCS and found that the spooling system had plenty of torque with just one servo motor.

After testing the system manually, we turned on automatic mode and set the timer to automatically uncover at 7am each morning and cover at 5pm each evening. We watched the system through a monitoring camera to ensure that it acted as expected. We tested the ARCS in automatic mode at Jericho Settlers Farm for one month.