For my Capstone Project at the University of Waterloo, four classmates and I conceptualized, designed and prototyped an essential oil extractor, which allows people to produce their own oils (like lavender and tea tree) conveniently at home. (What is a Capstone Project?)
The extractor produces oils using a method called rosin pressing. No solvent is required, only the application of high pressures and heat to produce oil from plants. It was an invaluable experience in mechanical product design, teamwork, and the general process of bringing a product to market. In addition to engineering my responsibilities crept into industrial design – attention was paid to user experience and aesthetics.
The early stages of the project consisted of research and conceptual design. This involved determining design requirements and constraints, then translating these into specifications. We identified a need for our product, researched its target market, and sketched out its rough form and inner workings. Soon it was time for engineering design and verification. We modeled the extractor in Fusion 360 and NX, and conducted detailed verification of its components using hand calculations and FEA. Finally we sourced components, created drawings, machined and 3D printed parts ourselves, and built a working prototype – complete with a representative enclosure and touch interface.
The “Counter Culture” Countertop Essential-Oil Press received multiple award nominations from the university, including best consistent engineering effort, and came close to a people’s choice award. We received promising feedback from advisors and industry professionals on our design and product outlook. In fact, some of my teammates are continuing with the project as a business, and have been accepted into the Canadian startup incubator Velocity.
My responsibilities in the project included mechanical design of the extractor’s frame and motor power train, as well as contributing significantly to research, conceptual design, project management, and UX/UI considerations.
Currently, there is no affordable and easy to use essential oil extractor, so that people may make their own essential oils at home. They must rely on purchasing expensive commercial extracts, and cannot use herbs that they grow themselves or customize the consistency and makeup of the oil according to their needs.
A need exists to extract essential oils from botanicals conveniently at home. The primary function of the product is to extract essential oils from botanicals, so that customers can use the oil they extract for aromatherapy, disinfectants, massage oils, or in foods and drinks.
The primary constraint of the product is that the extraction must occur conveniently at home. It is clear that many consumers prefer at-home appliances which automate the preparation of consumables that they enjoy often. Some examples include coffee makers, bread machines, rice cookers, and even microwaves. Using these products is preferred over purchasing prepared food from a store due to the lower cost of purchasing unprepared ingredients, increased consumer control and convenient preparation. The constraint of having the product at home also introduces size, noise and safety concerns.
System Block Diagram
As shown above, an at-home essential oil extractor would likely function as follows. First users connect the machine to a source of power. They then place botanicals into an input tray and issue a start command. The machine processes for a period of time during which it displays relevant data such as time remaining. After finishing the machine outputs the extract into a container and displays its potency. Users finally remove plant waste from the machine and add it to their desired food or oil.
Selecting an Extraction Method
Many existing methods for extracting oil from plants use a solvent to wash the plants clean of their oil, which is then distilled or evaporated. These include alcohol distillation, super-critical CO2, and butane extraction. For industrial use these methods work well, however for at-home extraction they are either too expensive, dangerous, or illegal (alcohol distillation requires a special permit in Canada). The rosin pressing method does not use a solvent, so there is no risk of food contamination, and does not require the release of harmful fumes or storage of compressed gases. It simply uses a combination of high temperature and pressure on the input material to squeeze out essential oil in a similar manner to an olive oil press.
The rosin pressing method does not use a solvent, so there is no risk of food contamination, and does not require the release of harmful fumes or storage of compressed gases. It simply uses a combination of high temperature and pressure on the input material to squeeze out essential oil in a similar manner to an olive oil press.
Developing Engineering Specifications
Based on these requirements a set of engineering specifications was developed. These specifications aim to apply concrete, measurable metrics to the above requirements which determine the success of the project. Namely it must:
- Produce 600 psi of pressure to extract the optimal amount of rosin from a plant
- Be capable of heating the pressed plant to 110 C
- Occupy a reasonable amount of space, less than 1 ft^3
- Weigh less than 35 lbs (ISO 1128 safe lifting mass of 99% of people)
- Minimize maintenance, consumables, and cleaning required
- Keep low levels of smell and noise, and good aesthetic design
- Obey laws and regulations for food, electrical and physical safety
Final Design Overview
The four major components and systems of the extractor are the enclosure, the frame, the powertrain, and the electrical system. The role of the enclosure is to provide a pleasing visual aesthetic as well as to facilitate safe interaction with the user. The purpose of the frame is to support the high separation force as a result of the pressing operation as well as to provide a mounting structure for the drivetrain and the electrical system. The drivetrain and control system facilitate the actual extraction process; the powertrain converts electrical power into pressure and temperature and the control system determines the quantity of each and how long and when it should be applied.
The powertrain itself consists of 5 major subassemblies which can be seen below. The motor, the motor shaft, the lead screw, the press, and the heaters. Each of these subassemblies work in conjunction with one another to convert torque from the motor into pressure as well as electrical energy into heat that is applied to the plant in order to carry out the extraction process.
Detailed Design and Verification by Subsystem
In order to size the mechanical systems the loading conditions were first identified. A plate size was determined based on the amount of plant substance that was required to be pressed. Based on the area of the plates a compressive force was calculated. With the pressure requirement of 4.5 MPa this force worked out to be 11.2 kN. For the detailed design outlined in this report a safety factor of 1.5 was applied to this load resulting in a design load of 16.8 kN. Because of the imprecise nature of the data used to identify the pressure requirement for the sake of simplicity 17
kN and 11 kN are used in this report as the design load and required load respectively. This design load was used for the sizing of all mechanical components meaning that the safety factor used is a global safety factor.
The lead screw, lead nut, and frame were designed to allow for approximately 50 mm of opening in the press plates. The final travel as determined from a cross section of the extractor in CAD is 43 mm. This includes 30 mm of thread engagement between the lead nut and lead screw at all times, as well as 5 mm or greater clearances between components at maximum and minimum extension.
Lead Screw Assembly
The purpose of the lead screw assembly is to increase the torque from the motor and convert it into linear motion. This is accomplished using a worm gear mesh with a 40:1 torque multiplication and a 5/8-8 ACME lead screw and nut. A tapered roller bearing in conjunction with a needle roller thrust bearing is also used to facilitate smooth rotation under load. A cross section of the assembly can be seen below.
In terms of the layout there are two main sections to the lead screw assembly; the bearing block, and the lead screw. The bearing block uses a tapered roller bearing that is rated to handle the required load. The loading conditions on the lead screw assembly are summarized below.
In order to determine the gear forces and the torque requirement it was important to consider losses from the lead screw/lead nut interface. These were calculated using equations for ACME threads from Shigley’s Mechanical Engineering Design, with an assumed friction coefficient of 0.16. This resulted in a calculated torque to raise requirement of 33.6 Nm. The reaction forces were also calculated based on the torque requirement and the gear geometry. This resulted in the loading conditions shown above.
There are two main sections to the lead screw assembly; the bearing block, and the lead screw. The bearing block uses a tapered roller bearing that is rated to take the 17 kN press load as well as the radial loads simultaneously. Since the lead screw is also supported radially by the lead nut through the platen mount this in conjunction with the tapered roller bearing is sufficient to support the lead screw. Also, because the 17 kN press load is only applied in one direction the tapered roller bearing is designed to handle this load while the needle roller bearing is only designed to retain the lead screw by preventing it from being pulled out of the housing.
The lead screw consists of six different pieces as shown above. An off the shelf ACME threaded rod and collar are used to bolt the screw onto a custom shaft. Torque is transmitted from the worm gear to the shaft via a 5×5 key and the axial press load is transmitted to the tapered roller bearing through a bearing adapter located at the end of the shaft. This configuration was selected based on cost, machining capability, and time constraints. Firstly, the lead screw could have been simplified by consolidating the threaded rod, collar, shaft, and bearing adapter into a single component. When assessing this design the problem was being able to manufacture it. ACME threads require precise tolerances which necessitated finding a vendor with the capabilities required. A vendor was found however they were unwilling to work with us because the quantity was so low, i.e. 1. Even if they were a one-off outsourced component would have been prohibitively expensive. Instead the design was broken up into the pieces shown. Welding was also looked at for joining the threaded rod to the collar. This would have allowed the rod to be joined directly to the shaft without the collar. However as will be discussed there was some doubt as to whether the lead screw would be strong enough so it was important to be able to replace the lead screw, which is comparatively inexpensive to the shaft, in the event of a problem.
Due to the cost and lead time associated with the lead screw assembly it was important to conduct a thorough analysis of the design to ensure that nothing was going to fail when tested. To do this a combination of FEA and hand calculations were used. FEA was used primarily on the shaft, bearing adapter, and bearing block although it was also used to assess the keyway in the worm gear (not shown in this report). Stress calculations from Shigley’s were used for checking the thread stress in the threaded rod and collar. FEA was not used on the thread due to the complexity of the geometry and time constraints. The most critical components with regard to strength were the shaft and the thread.
The results of the shaft stress analysis are shown above. As shown the press load and the torque transmission force were the only forces considered. The radial loads were ignored since they were orders of magnitude smaller. A fixed constraint at the bearing adapter interface was used to represent this interaction. The results of the analysis showed that the limiting geometry was the keyseat radius. Mesh refinement was used and the keyseat length was increased along with the corner radius until a satisfactory safety factor was obtained. Note that this analysis and all subsequent analysis already has the global safety factor of 1.5 in the design loads. Also this optimization was done for C1144 which was selected among the materials available at the UW Machine Shop where this was intended to be manufactured.
The results of the thread stress analysis are shown above. As can be seen from the calculations the safety factor for the lead nut is less then 1. Since having a vendor machine an ACME thread was infeasible a lead nut had to be either borrowed from another piece of equipment or bought off the shelf. It ended up being impossible to find a lead nut strong enough in the time that was available so an iron lead nut, which was the strongest found, was purchased. Even though the calculations showed that the lead nut did not meet the design specification it was reasoned that it could still work for the following three reasons. The first was that the design loads used the 1.5 safety factor. This meant that the actual safety factor for the lead nut was 1.5 times higher bringing it to 1. The second was that the equations used to calculate the thread stress were for a square thread which is somewhat weaker than an ACME thread due to the shape. Finally, a scissor jack was analyzed and it appeared to be handling loads far higher than these calculations deemed was possible. Therefore, for these three reasons it was felt that this design was acceptable.
Ultimately during testing up to the required press loads it was observed that there were no structural failures or excessive wear in any of the lead screw assembly components. In addition to achieving the load requirements the assembly also spun freely with lubrication.
A diagram of the press assembly labeling its sub components is shown above. In summary, the lead nut converts the rotational motion of the lead screw to linear motion via ACME threading and being fixed rotationally by the guide rails. The aluminum platen mount then slides along the guide rails until the two platens meet and provide the pressing surface. They are made of polished 304 stainless steel, for food safety and corrosion resistance. The platens rest on steel dowel pins in the platen mount and heating plate so that they can be removed for easy washing. The heating plate is also able to slide on the guide rails. This allows the springs behind the heating plate to deflect during operation and account for any backlash and creep in the system when the motor turns off. Aluminum was chosen as the material for the platen mount and heating plate due to its very high thermal conductivity, since these parts are part of the thermal pathway of the heating system. The heating plate thickness had to be minimized to have the fastest heat transfer from the heater to the plant material but also had to be strong enough to handle 17 kN of load during operation.
The heating plate houses the heaters in a cutout in its back side. The initial thickness of the heating plate was 12 mm in front of the element. According to the early 1-D, steady-state thermal model this was acceptable. However, preliminary thermal testing showed that the transient time to reach the desired temperature was very long. So, the part’s thickness needed to be reduced. Minimizing the thickness led to mechanical failure in the FEA simulations due to a stress concentration. To remedy the issue and reduce the thickness as much as possible the cutout was reduced in width and the heating elements shaved down so that the stainless-steel platens spread the load to the thicker part of the heating plate. The figure below is a simulation after the redesign, the safety factor at some point contacts in the mesh is only 0.8, however, the real case has springs that redistribute that load, so the worst expected case is one with slight local yielding. The rest of the part has a safety factor of at least 3 and the stress concentration issue has been resolved.
Several designs were considered for the lead nut. The calculations from Shigley’s for thread strength determined that a minimum yield strength of 400 MPa was required for the lead nut material. The material also could not be steel to avoid galling. C863 bearing bronze was chosen with a yield strength of 415 MPa. C863 bronze is an expensive material, to reduce cost option 1 in the figure below was explored, as well as, a second more standard option 2.
It is difficult to simulate threads in FEA so the loads were approximated by a shear stress and a torque over the internal diameter of the nut and the hand calculations for thread stress were used. The FEA shows that both designs are acceptable however, the threads could not be machined by a local machine shop in time and within budget. An iron lead nut with a yield strength of 300 MPa was chosen that could be bought off the shelf. Based on back calculating the load on a car jack with a similar size lead nut it was determined that the iron would be acceptable, and that the hand calculations were overly conservative.
Based on the screw torque and distances to the guide rails the transverse load on the rails was determined to be around 820 N on each rail, which is very low for a tool steel rod of the required length. Therefore, the 8 mm diameter was chosen based on the available bushings. The polished, hardened tool steel material was chosen to provide a smooth surface for the bushings to slide along that would be wear resistant. Adding the friction coefficient of the tool steel, bushing combination allowed the calculation of axial loads on the guide rail system. This load was only around 130 N, which was not an issue for the aluminum guide rail anchor or the steel back plate where the rails are mounted.
The platens were designed to be easy to make, they consist of a 304 stainless steel 5 mm thick plate and a 1.5 mm thick aluminum that is screwed on by M2 fasteners. The aluminum has holes machined in to go over pegs in the platen mount and heat plate. These were dimensioned carefully so that no load ever passes through the thin aluminum and into the dowel pins. Rather, it goes directly through the steel.
Beville spring washers were chosen to replace the initially designed load cells. It was determined that precise knowledge of the pressure was not required if there was a way to determine whether a minimum pressure of 4.5 MPa was achieved between the platens. To do this, spring washers that deflect at a load of 5000 N were selected. With two springs in parallel, if deflection occurs the machine has hit the required load. Additionally, these spring washers provide some travel which prevents a loss of pressure on the plant material due to any slack in the press gearing system.
All the components mentioned above were tested during assembly level tests. None were damaged during the testing, this is an indication that the validation and hand calculations were accurate or at least conservative enough to avoid failures. The springs were tested using a load cell and an arbor press to ensure that they met the specifications. This test showed that the maximum load output of the machine was around 11500 N which is just above the designed output of 11 kN.
The frame and motor shaft subsystems are shown above. The purpose of the frame is to resist the large axial force generated by the press, and to locate components accurately. It consists of a main, upper section constructed from mild steel and a foot-pedestal made from 6061 aluminum. The upper section consists of three 1/2” mild steel plates and two 3/8” steel tie rods, and the lower section four 1/2” aluminum standoffs and a thin foot plate.
One could also connect two side plates using four tie rods or two base plates, however these options are problematic. Using four tie rods makes locating components difficult. This is especially true for the press guide rails which must be carefully aligned to prevent binding. On the other hand, using two base plates was considered unnecessarily heavy and expensive.
The front, back and base plates were the most difficult components of the frame to manufacture. Laser or waterjet cutting were originally considered due to their accuracy and speed for positioning the many mounting holes in the plates. Since the edge quality of laser cut steel is low, critical features such as the motor shaft bearing hole would be undersized and then finished with a reamer. An agreement was made with a local shop to laser cut the steel plates, however it fell through. In the absence of this the plates were machined using a mill.
In terms of analysis, an engineering decision was made to trade design effort for weight. Optimizing the frame was not a primary concern for a first prototype, however building one which did not fail and could be manufactured and designed quickly was. The frame plates, tie rods, and fasteners were generously oversized as “cheap insurance”. FEA was however conducted on the front and back plates to determine verify their thickness (below). The analysis was conducted in Autodesk Fusion 360.
It can be seen the simulation geometry is different to the final geometry of the front plate. The plate above features four tie rod holes, rather than two tie rods and a base plate. The results of this analysis were assumed to apply to the final geometry, provided the safety factor was generous enough. A fixed constraint was specified at each of the four tie rod holes above, and the design force (17 kN) to an equivalent footprint of the press. Built in material properties of mild steel were used – a Young’s Modulus of 2.24 GPa, Poisson’s ratio of 0.38 and yield strength of 300 MPa. The plate was meshed using tet elements, and locally refined at the tie rod holes with one element through its thickness. From this analysis, a 1/2” plate was found to be sufficient with a safety factor of 2. Since washers were applied at the tie rod holes the simulation also likely over predicts the local stress.
Hand calculations for the frame plates were conducted to provide confidence in the FEA results. The frame plates were assumed as simply supported Kirchoff-Love plates with a 28 mm diameter circular load applied in the middle. Material properties, plate dimensions, and the 17 kN load were entered into an online plate bending calculator published by MIT to find that maximum bending stress in the plate was 164 MPa. This corresponds closely to the stresses seen in FEA and suggests the safety factor in the plates is in the order of 2.
The tie rods were sized with a back-of-the-envelope calculation according to:
where n the safety factor against yielding was generously chosen as 5, A is the cross-sectional area of the tie rod, σ the approximate yield strength of mild steel (300 MPa), and F an applied force of 4250 N. The force was determined by simply assuming resists approximately a quarter of the total axial force of the press (11 kN). It was determined a mild steel tie rod of 8 mm diameter or greater would be more than sufficient. Fasteners were chosen based on convenience and standardization across other subassemblies. The horizontal and vertical plates are held together with five class 5 M4 fasteners, with a proof load of 7100 N each. The safety factor on a tensile failure of the fasteners was very high, even assuming a worst-case scenario where only two of the five fasteners support the axial load.
It was not deemed necessary to conduct FEA or hand calculations for the bottom plate. This is simply because the tie rods, front, and back plates were all more than sufficient in strength while under worse theoretical loading conditions. The finished frame performed its purpose of locating subassemblies accurately, and did not fail under repeated full-load tests. The frame is the heaviest component of the extractor however and in terms of specifications brought the prototype slightly above the desired weight spec of 15 kg.
The motor subsystem consists of a 23 Nm stepper motor, a worm gear and worm gear shaft, and various mounts. It is designed so that the motor can be mounted accurately and transmit torque with low friction. The motor and motor shaft are shown in red, while the worm gear shaft is shown in blue.
Main challenges in the design of the motor assembly were determining a mounting strategy for the motor, and determining a method to isolate heavy reaction forces on the worm shaft from the motor shaft. The motor shaft is not rated for significant axial or bending loads, yet based on previous gear calculations it was determined the worm receives approximately 790 N axial force and 280 N lateral force at its contact with the worm gear, at design load. Existing worm gear sets were researched to determine how the worm shaft could be mounted, and it was found most worm shafts are supported at either end with bearings. Rather than purchasing a compound bearing, it was cheaper to purchase separate radial and thrust bearings. Upper and lower bearing mounts were designed to hold one radial and one thrust bearing each.
The shaft of the specified motor is ¼”, and inserts into the ½” shaft specified for the worm. This “slip interface” is shown in the motor subsystem figure above. Two strategies were investigated for how to mate the motor to the worm shaft. One was to machine a slot into the worm shaft with an end mill, and two flats into the motor shaft. This would allow the shafts to slide axially relative to each other, and avoid axial load sharing. However, machining a slot into a tool steel shaft was deemed too difficult. The chosen option was to drill a hole into larger shaft, machine a flat into the smaller, and secure them with a set screw. Theoretically axial load could be transferred through the set screw, but testing showed the motor was not damaged due to this.
The upper bearing mount was machined from ¾” thick 6061 aluminum, due to its availability. The same material is used for the press guide rail mount. Its geometry was kept as simple as possible for ease of machining. The shaft of the motor is only 15 mm in length, meaning the worm shaft needed to be mounted as close to the motor as possible for sufficient insertion length. As a result, the lower bearing mount was machined directly into the base plate. It consists of a hole machined through the base plate to house the bearing set, and two thin 1/8” mild steel plates to interface with the motor and serve as a backstop for the bearings. The bearings were not rested directly on the motor casing, because no specification could be found as to the casing’s strength. The motor is secured onto the bottom of the base plate using four M5 fasteners and nuts.
FEA was conducted in Fusion 360 to verify that the upper and lower bearing mounts would be sufficiently strong to resist the applied loads, with minimal deflection. Excess deflection could cause the worm and worm gear to become misaligned. The analysis (below) showed that maximum deflection was less than 0.1 mm, and the mounts both had safety factors of greater than 3.
Electrical and Software
The electronics used to control the device are summarized graphically below in Figure 17.
As seen in the figure above, the electrical system is split into three parts: thermals, powertrain and UI/safety. The thermals system is comprised of a relay card which takes in digital inputs from the microcontroller (ON/OFF) to control temperature based on the thermistor’s feedback signals. The powertrain consists of a driver, which takes in a clock, direction and enable digital input from the controller to spin the stepper motor. Limit switches are used to determine the start and stop positions of the motor. Finally, the UI/safety for symposium consists of a buzzer, LED and buttons for user interaction with the device. CSA certification was avoided by power supplies for thermal and powertrain subassemblies.
As the motor’s main functionality is to convert torque at the shaft to axial force at the press, a series of calculations were performed. These calculations use the required pressure at the platens as a starting point and first determined the torque to raise the lead screw mechanism then through the gearset based on the measured efficiency. Due to timing constraints, the motor was slightly undersized from its 1.8 Nm specification and high vibrations were observed when “full stepping” (i.e. 200 steps/revolution). In order to reduce these unwanted vibrations, the number of steps per revolution were increased which transforms the motor’s current waveform from square to sinusoidal.
Consequently, this reduces instantaneous torque since current is directly proportional. Normally, this optimal operating point (number of steps and speed for max torque) is clearly identifiable on the supplier’s torque speed curve. Unfortunately, the specification sheet for the selected motor was limited therefore, this optimal point was determined by trial and error. Best results were observed by 1/8th stepping (i.e. 1600 steps) @ 200 RPM. Since no optical encoder could be obtained during the time of the validation, speed was validated based on the measured microcontroller’s clock signal frequency. A clock pulse of approximately 5kHz is observed. This pulse is sent to motor driver which converts this frequency to motor speed.
An overview the of the heating system’s thermal design is seen below.
This thermal circuit was used as an approximation to determine the power requirement for the heaters. Initial heat transfer calculations concluded that the purchased (56W) heaters were more than adequate for this application. This large factor of safety was confirmed when the heaters were put to the test with a test plant material, beer hops. During this test, a small amount of hops were placed in a bag and exposed to rated pressure with a target temperature of 110 C, as shown below.
From above, in approximately 14 ½ minutes, the beer hops were heated up to the target temperature. With the data gathered from the test , the following correlation between measured temperature at the “heated platen”. This was determined using MATLAB matrix divide to determine gain temperatures.
This correlation was then coded into the firmware.
Although the enclosure was not the focus of this project, a somewhat visually representative enclosure was designed and manufactured for use at the design symposium in order to help convey the nature of this project as a consumer product. Although the enclosure was not necessarily the most elegant it added significant value to the display. The enclosure was constructed using a roll of 1/64” aluminum sheet, a chair, cardboard, an epoxy adhesive, and black spray paint. The enclosure was made by cutting and the aluminum sheet with a pair of scissors and bending it using the legs of the chair. The cardboard and epoxy were used to stiffen the flat sections of the enclosure and to keep them strait. Three coats of self-etching primer and three coats of satin black primer were applied in order to give the enclosure a finished appearance as well as to hide blemished and panel gap. Additionally, a small wooden stand was constructed from wood scrap and the adhesive to create a stand for a phone that was used during symposium to run a demonstration user interface. The enclosure can be seen in the figures below.
Branding and User Interface
In addition to the mechanical engineering above, careful attention was also paid to the industrial design and branding of the extractor.
Due to the cost of the gears required for the extractor’s press, the final sale cost of the product is likely to place it in the premium, luxury category, similar to high end coffee and cappuccino machines. This was kept in mind when designing the branding and aesthetics of the extractor; to attempt to convey a premium look and feel.
The user interface of the product is a touchscreen that allows the user to customize their settings to experiment and optimize their own process for the specific results that they want. A touchscreen was selected due to its flexibility, easy cleaning, and low relative expense compared to other internal components. It makes it easy even for those with little experience to perform simple extractions based on their input and desired output.
The figure above shows some of the early stage design flows and mood boards used to determine the overall flow of the interface and the style. The flow diagrams and wireframes were tested by having potential users do cognitive walkthroughs to find where they were struggling and how the design could be improved. The mood boards explored different styles, color pallets, typography and icon designs.
Two competing styles emerged from the early work. One that was more photography oriented and one that had a dark, utilitarian, high-tech theme. These two are shown in the figure below. The latter was chosen because for a home appliance the challenge is not to stand out. The dark theme is more subtle and allows the photography that is functional to standout. The entire design appears less busy and fits better with the industrial design of the product.
Additional images illustrating sample panels and the viewing angle of the screen are shown below. The touch screen was angled so that it could be viewed easily while sitting on a counter.
Safety and Regulatory Considerations
The design and extraction method have inherent safety risks and can be summarized as follows:
- Motor, heating element and other components need to be powered
- Mitigation: kept electrical components under 100VA
- Temperature of 110 C reached by platen
- Mitigation: enclosure to avoid users touching any hot components
- Press force: 11 kN
- Mitigation: robust mechanical design, safety stop button trigger for software
- Food quality: avoid adding metallic components to output substance
- Moving parts: pinching force of press
Throughout the course of the project, CSA certification (i.e. > 100 VA or 30V) was a big question. To avoid this complicated process, power the thermals and motor were separated. Heaters and fan use 12V @ 5A (60W) while the motor uses 24V @ 3A (72W), both of which have appropriate CSA mark.
We learned a lot about how to work as team, and balance project management overhead with engineering work. We attempted to keep our overhead light, only using PM tools which actually served a needed purpose.
We tracked our progress by first setting overall milestones with deadlines in a Gantt chart, then breaking these goals down into tasks, and subtasks using an online kanban board. Individual responsibility was assigned to each task so it was clear what each person needed to do. Although, we understood the group was in the same boat and helped with items we were not directly responsible for frequently. We kept version control of CAD and important documents such as specifications and our bill of materials using Fusion 360’s online services and OneDrive. Finally an online group chat, and regular weekly meetings between team members and our advisors were used for clear communication.
Ease of Assembly and Detailed CAD
Considerations for ease of assembly are important, even for a one-time prototype, and even under time constraints. Having nuts on both sides of the frame plates resulted in the entire frame plates having to be removed to make adjustments. In hindsight nuts could have only been added to the outside of the frame.
Including everything in CAD down to fasteners and washers is also important, to avoid any possible interferences. At one point, an out of place washer resulted in two components interfering.
Cost vs Time Decision Making
As a group, there were times when cost vs time was not properly taken into account. For instance, our group was offered free machining, and free thermal paste. However, these free services came to us at the cost of us time. Early on, much time was wasted waiting for these services, and we would have had more success early on had we understood that being frugal is not the best way when it comes at the cost of time.
Better Understanding of Blockers
When budgeting time, not much thought was given to potential blockers, and time management was done with perfect cases in mind. Early on, we had a lot of difficulty following our schedule. Building a product is a very messy path, and this needs to be accounted for during planning.
Choosing an Ideal Project Management Methodology
Initially, we had started with the Waterfall approach to project management, as things were seen as one independent part that follows one flow, with one final build test. However, after a while, it was determined that an agile methodology would be more successful for our team as it follows an iterative approach in which testing and building happen at the same time.
Importance of Appearance
Under time constraints it is easy to let the appearance of a prototype go to the wayside. However, as an advisor suggested, not paying enough attention to appearance can undermine the hard engineering work that went into a product by making it seem haphazard. We took this advice in mind and built an enclosure and symposium display to communicate the aptitude of our engineering work. Ultimately, it was a success as on symposium day the crowd was drawn to our display.
This post was written with the permission of the University of Waterloo and my friends at Counter Culture.