eb
Capstone Project: Automated Cake Decorator - Phase II
Originally, the automated cake decorator was designed to allow home bakers to decorate custom cakes without the usual time requirements and tedium. However, as the project developed, it became apparent that the machine would be better used in a commercial environment like a grocery store, where custom cakes have to be created by relatively untrained bakers. The machine can be thought of as a 3D printer for frosting. Unlike a 3D printer, our decorator had to accommodate the size and location of a cake as well as the weight and viscosity of frosting.
Background: Phase I
​
​Our capstone project was considered a Phase II project, meaning that a decorator had already been built by a Phase I team the previous year, and it was our job to improve upon it. This sounds simple, but to incorporate the changes we wanted to make, we found that we needed to basically redesign the entire decorator.
​
Phase I put together an off-the-shelf plotter, altered the dispenser to extrude frosting, and set the controls to create pixelated designs on a cake. They were successfully able to decorate a cake, but there were some key issues that we set out to improve upon. Those problems and our solutions are highlighted in the objective section.
Phase I Decorator
Cake decorated by Phase I - the decorated did not accommodate the size or curvature of a cake, resulting in defects
Objective
​
While many possible improvements were considered, we were finally able to narrow our scope down to four main objectives. We wanted to add another major mechanical component to the project so we decided to make the machine switch colors automatically instead of interrupting the decorating process so that the user can switch out the frosting syringe. This allows the user to walk away from the decorator and come back to a fully decorated cake. The rest of the improvements focus on expanding the decorating ability and the type of cakes that can be decorated. These four objects are highlighted in the graphic below.
Process​
​
Many MANY ideas were considered to achieve the objectives. A few of the ideas that were brainstormed initially are highlighted in the image below.
It became clear very early in the design process, that in order to achieve color switching and make the machine less bulky, a new extrusion method would have to be considered (previously, syringes were used in such a way that the height of the machine had to be doubled). Augurs, hydraulic pumps, and a whole bunch of crazy ideas were brainstormed. For color switching, it was decided that the extra frosting would have to be kept off extruder to reduce the weight and size of the extruding system. For height adjustment, both simple and complex options were considered.
Final Design​
​
Below is an image of the final product. Each of the subassemblies and some of the testing done on them are then shown and explained in detail.
Frame
On the frame, a slotted tray rack permits rough height adjustment to accommodate 1” to 6” cakes. The multi-level tray rack resembles an oven rack, making tray placement easy and familiar. The low-resolution rack eliminates the need for excessive z-motion, allowing the high resolution z-axis to only perform vertical motions required for decoration and adjustment for the cake’s curvature. The frame also supports the frosting cartridge rack where frosting cartridges can be accessed by the extruder and switched automatically.
Crossbar
The crossbar was simply redesigned to replace the large and heavy, off-the-shelf rail system used by phase one. Now the carriage simply rides on two slender, parallel rods, and is driven by a belt.
Height Sensor
To determine the location and topography of the cake’s surface, a time-of-flight infrared sensor maps the surface of the cake prior to decoration. Once the surface and edges are mapped, the bordering and other decorations can be frosted onto the cake in a way that accommodates variations in the cake’s surface and placement. Initial edge and surface mapping testing was conducted by passing a time of flight sensor over a cake frosted with a base layer of frosting while collecting distance readings at forty points. The test showed that the cake’s edges could be reliably located by comparing distance readings at adjacent points. It also revealed that the sensor must be calibrated for the base frosting and that a smaller step size than was used in initial testing is needed for high resolution edge detection. Further testing with an automated process confirmed these results.
Extruder
The extruder mechanism is the key to color switching and reducing the extruder height. The cartridge slides into the front of the extruder, where it is placed concentrically between four locating bosses. A plate that rides on two lead screws is lowered to contact and secure the cartridge. Additional compression of the cartridge via the plate extrudes frosting. Calculations and testing were done to ensure the motor could provide enough force to extrude frosting. as a starting point. The force needed to extrude frosting from the bottle was found experimentally to be <17 lbf using an Admet force/displacement tester. The torque-to-raise equation was used for two ¼-16 ACME lead screws to determine the minimum torque required to compress the bottle, assuming a maximum required force of 17 lbf. A minimum required torque of 1.1 in-lbf was calculated. The NEMA 17 produces 3.80 in-lbf of torque, and so can be used to drive the extruder.
Color Swtiching
The ease of separation of the cartridge from the extruder facilitates color switching. The cartridge can be picked up by the extruder much like a fork-lift and then placed back on the cartridge rack in a similar manner, with magnets used to locate the cartridge on both the extruder and the cartridge rack.
Skills Gained & Lessons Learned:
​
-
How to work towards a tight timeline and to save time for testing. While the machine mechanically worked very well, we did not have enough time to sufficiently characterize several frosting decorations.
-
Work with a team, ensuring that everyones opinions are heard while efficiently delegating work.
-
Tolerancing - both what can be easily achieved with a machine shop and 3D printer, and what is required for a machine to be stable and move effectively. I was able to develop a better intuition for mechanical tolerances.
-
How to communicate with a machine shop so that we get accurate parts in a timely manner. We were also able to get lots of advice on machining requirements from the people in the shop.
-
Brainstorming. While the solution seems obvious now, it was definitely not when we started this process.
-
Considering all objectives to come up with a cohesive design.
