Spinova was designed using the cloud-based CAD software OnShape, and printed using PLA filament on an Ultimaker 3 Fused Deposition Modeling (FDM) printer using Ultimaker Cura 3D slicing software. Our design took inspiration from CoreElectronics’ Random Positioning Machine, and we started with a top-down approach to parts. In the end, Spinova went through over 5 different iterations and is still accepting and implementing user feedback!

• Drafts of parts are dependent on relevant dimensions and locations of other sections of the design, meaning a single sketch or feature can drive the geometry of several parts in the overall design.
• Each solid part of the clinostat was modeled in its own independent parts studio in OnShape, then combined in an assembly.
• We followed an iterative design process where surveys were sent out to potential end-users and to those knowledgeable about mechanical engineering and design.
• The surveys asked for feedback on each iteration of Spinova, which was then revised in-house and updated following the feedback we received.

To validate the design decisions we made to follow our specified requirements, we conducted a series of functionality tests with the goal of correlating our design requirements to Spinova’s real-life specifications.

Purpose: ensure 3.4:3.8 RPM rotation ratio is maintained for consistent simulation of Martian gravity conditions.

Method: Use a tachometer to assess rotation speed of the sample platform and the inner rotating frame.

Criteria for success: Average speed for sample platform 3.4 +/- .1 RPM, avg speed for inner frame 3.8 +/- .1 RPM, stable over time with >4% variance. By setting a 4% variance limit, we accept a small performance deviation that is within a biologically acceptable range for E. coli. Further testing is an option to validate that E. coli remains biologically stable through growth curves and phenotype monitoring. To have the inner rotating frame rotate at 3.4 RPM, the motor must rotate at 11.6 RPM. To have the sample platform rotate at 3.8 RPM, the motor must rotate at 13.4 RPM. These calculated RPM for the motor come from the gear ratio and speed relationship.

• Gear ratio = Driven teeth/ Driver teeth
• Driven speed = Driver speed x (Driver teeth / Driven teeth)
• For the rotation of the inner rotating frame: Driver speed = 3.4 RPM x (75/22) → 11.6 RPM
• For the rotation of the sample platform: Driver speed = 3.8 RPM x (60/17) → 13.4 RPM

Relevant design choices: Selected the FS90R continuous microservo motor with a range of rotation speeds and used gears to mediate transfer of torque.

Results: When running Spinova with liquid cultures of modified BL21, laser tachometer readings indicated consistent RPM at each 30-minute interval tested. The recorded speed of the inner rotating frame was 11.5 RPM for each time interval and the recorded speed of the sample platform was 13.5 RPM for each time interval. This is within our set tolerance limits.

Purpose: Ensure Spinova meets physical constraints imposed for lab/shuttle use.

Method: Measure weight of all components in CAD and in print to check for deviations in weight, measure actual dimensions to assess deviations in size, place in incubator at 37 C to assess thermal tolerance.

Criteria for Success: Must weigh under 0.375 kg with less than a 10% error margin, fit on a standard lab bench, and maintain structural integrity at biological sample culture temps 37 C. ESA technical constraints for drop tower experiments, which is a relatively small hardware component for an experiment, outlines mass and space allocations for small projects in the setting of a space mission (ESA). These constraints guided our goal weight for Spinova.

Relevant design choices: Used PLA plastic to 3D print plastic components and designed dimensions to fit on a laboratory bench while in use.

Results: Model weight was calculated in OnShape using the “assign material” feature and found to weigh 0.371 Kg, including estimated weight for non-3D printed materials. Actual weight was calculated with a lab scale and found to weigh 0.338 Kg. This resulted in a 9.87% error between the CAD and the physical model. The source of this error was likely due to the CAD model weight not accounting for the infill the print was actually run with. All Spinova 3D printed components were printed at 15% infill. The actual dimensions of Spinova were measured to be 251.00 × 251.00 × 149.48 mm. To test the thermal tolerance of Spinova, it was run continuously for two hours in a 37 C incubator without a sample. Periodic checks revealed no structural deformation. Subsequent wet lab experiments were run with Spinova and found metabolic processes remained optimal and intact when modified BL21 were cultured overnight in Spinova. A graph overlaid with non-Spinova control cultures is displayed.

Purpose: Ensure Spinova can be assembled in a timely manner without specialized knowledge or tools that are not already carried by astronauts for routine spacecraft maintenance.

Method: Time assembly of Spinova, then time assembly of Spinova for someone who did not design it. Assess if tools used for assembly deviate from those described in the user manual.

Criteria for success: Assembly under 5 minutes, only using user manual instructions and tools.

Relevant design choices: Some parts like the table legs are attached using super glue and don’t need to be installed every use. Tolerances were employed to allow close fits while maintaining freedom of rotation where appropriate.

Results: Spinova assembly time was recorded for various lab members, as well as individuals unaffiliated with the iGEM team. Subjects were given a small Philips head screwdriver, zip ties, and rubber hammer to assemble Spinova using the instructions provided in the user manual. An average assembly time of 270 seconds (4 minutes and 30 seconds) was recorded among seven participants. Note, time does not include time to adhere table legs to the outer stationary frame.

The design for the extruder casing was created with a combination of Onshape and TinkerCAD. The casing was modeled around the dimensions to fit the functional parts of the extruder. Parts were sliced with Ultimaker Cura and printed using an Ender3 FDM printer with PLA.

The extruder base was designed as a single piece to hold together the main components of the extruder. The base, in particular, had to be able to hold an auger attached to a Nema 23 motor at a 90 degree angle from the base, so that the auger could go through the metal pipes that fed into the hot end without jamming or getting stuck. An inverted pyramid created a feeding tray for plastic to be fed into the pipe by the rotating auger.

The design of this base was both too large and too inconvenient, both in wasted filament on supports and on time, to be printed on the lab’s Ender3, as it simply did not fit on the print bed. A decision was made to break it up into modular pieces that could be snapped together like legos and redesigned independently.

The design was broken down into nine separate pieces that would be designed to snap together. The largest and most complex component, the collection tray, was printed as a single part. To determine the optimal connection method, we conducted small-scale printing tests comparing pegs with slot-based fittings. The slot design failed due to printing inaccuracies from large overhangs, resulting in poor fits. In contrast, the peg design was easier to print reliably and allowed for simple post-processing. When the printer's offset—a difference of around 3 mm between digital and physical parts—caused fit issues, the pegs could be easily shaped with a Dremel tool to ensure a proper snap-in connection.

The second extruder housing was modular and could be snapped together using the peg connections. The separated parts could be individually modified to account for errors during the design process. These parts were printed in separate batches with PLA filament using an Ender3 printer with 20% infill.

Parts were tested for fit after printing. Measurement errors, such as misplaced pilot holes for screws on the plate holding the metal flange, were fixed immediately after.

Fit testing with all our printed parts showed that the front plate with the hot end of the extruder had difficulty snapping onto the rest of the casing. The screw heads on the other side of the flanges were running into the plastic of the collection tray, preventing it from snapping all the way. Additionally, the body housing the auger to motor coupler was too short to fully house the shank of the auger. The plate attaching the NEMA 23 was also misaligned in its screw holes. We learned from this version that our previous measurements, taken from online documentation, were not precise enough for our purposes. We manually took measurements from our ordered parts using calipers and rulers for designing the next iteration.

Version 3 of the extruder casing used the improved measurement information from V2. It had sharper angles for the collection tray to move the walls out of the way of the screws. The sharper angle also helped in causing plastic in the tray to fall and be guided towards the rotating auger. An additional section of the casing housing the auger was added to allow for the length of the auger and to further stabilize it. The plate holding the NEMA 23 was also redesigned in accordance with the physical screw placement of the motor.

Version 3 was assembled and tested following the complete assembly of the extruder. It was found that the weight of the auger caused the angle to be off from 90 degrees, where it was tilting down and pressing against the tube lightly. Due to this, when the motor was turned on, the front half the casing would lightly oscillate from the movement of the motor. The weight of the nema was also an issue since the extruder would frequently buck during oscillations and fall over from the housing collapsing due to the removable pegs. Additionally, the front plate holding the extruder did not snap in tightly due to the fact that screwing the plate to the flange was mildly warping the plastic, causing it to not be completely straight. This led to large gaps through which plastic could fall out of and a collection tray that would easily fall apart with movement. The modularity and pegs themselves also became issues as they were too loose to hold together the entire extruder. We learned that the modular, snap-fit design was not suited for the mechanical load, and the system needed to be anchored to a rigid external frame.

Through testing Version 3, we determined that a fully 3D-printed housing for the filament extruder was not sturdy enough. We changed to eliminating the housing for the motor and everything except the collection tray. The collection tray was also redesigned to print in a single part, eliminating the use for pegs, to eliminate any gaps and better hold the tray together under strain. A base for the tray was also printed with holes that would fit the legs of the collection tray, so it can be easily removed, and had screw holes that could anchor it down to a base. Metal bushings were added to the holes that the auger would pass through to eliminate the auger cutting into plastic and bucking around with the purchase. This helped stabilize the movement of the auger. The housing was replaced with a base made of 2’’ thick pine wood, cut with a hacksaw, both to pad against the heat generated by the extruder and to create a solid surface it could be anchored onto. Blocks of wood were screwed onto the wooden base with metal brackets to brace the extruder against horizontal movement and to keep it easily removable for fixes. This version was tested and it successfully extruded filament without human intervention, as we did not have to hold together the extruder parts to prevent them from moving.

The filament winder was based off a design from shoaibkhan786 for a PET bottle recycler on Thingiverse. The gears were modeled in Onshape and the rest was modeled in TinkerCAD. The winder is powered off a NEMA17 motor and slowed down through 5 gears to slowly turn a collection reel. Parts were printed in batches and in PLA plastic. 6 19 mm OD 7 mm ID ball bearings were inserted into each side (3 per side) with 7 mm printed rods going through the center to allow for free rotation of the gears on the rods. Gears were superglued down on the rods to prevent movement and shifting. The NEMA 17 is mounted to the side with a block placed under to better distribute the weight and hold up the motor. 4 mm 65 mm screws and nuts were inserted in each corner and the bottom block to hold together the winder. Version 1 was tested and to have been found functional with no major flaws, confirming that the gear train was able to rotate the collection reel at a much slower rate successfully from the NEMA 17 motor.

To test the initial performance of the extruder, we used PLA filament readily available in the lab. We chopped the PLA filament reel into small pieces to accurately mimic the size of PHA granules typically extracted from microbes. This input allowed us to successfully produce filament using the extruder; however, the auger drill occasionally became stuck when unevenly cut pieces lodged between the drill and the hot end.

To prevent the auger drill from getting stuck, we decided to use a more powder-like material resembling the plastic granules expected after microbial extraction and purification. For this, we used waste or byproducts from previous 3D printing projects and blended them into a fine powder. We then took this a step further by applying a polymer purification and recovery method (can find the exact procedure in our Wet Lab Protocol section) to purify and precipitate plastics such as PLA and PHA, forming a soft, fluffy, and airy substance that extrudes more efficiently, with minimal interference between the drill and the hot end.