Tesselating magnetic lights


Modular tiling lights created for ME325: Making Multiples

Tesselight began as an idea in ME318: Computer Aided Product Creation, but was ultimately scrapped for that class in favor of a different project. The idea of making modular lights eventually resurfaced again when I took Making Multiples, due to the inherent ability of injection molding to quickly make a lot of something. The initial idea of snapping together magnetic lights proved too expensive and overscoped for an eight week project so the design was ultimately simplified, but the high level concept of tiling lights remained the same.


The design consisted of three main components, the housing, the diffuser and the PCB. The PCB has a switch, two magnets and a battery on the back and a warm white LED on the front. When placed on a magnetic surface such as a blackboard, the magnets on the PCB pull the body onto the surface, depressing the switch and turning on the light. This light then shines through a diffuser to provide a nice, even color without any aggressive hotspots. The PCB is attached to the body with thread forming screws, and the diffuser is heat staked through the body onto the front.

Cutting Tooling

The first tooling to be cut was for the body of the light. It was cut on a HAAS OM2, and the toolpaths were done in Autodesk HSMWorks. I made core and cavity inserts for a MUD base in the Arburg vertical mold press in the PRL. Cutting my own tooling allowed me to put the molds back on the CNC and make adjustments as needed, such as opening up the gate for better flow, or adding ejector pins to make production easier. This proved critical when I decided to heatstake instead of overmold the diffuser, and had to add new shutoff surfaces to make holes through the body of the light.

Initial Shot

The initial molding of the housing worked well, and molded with very little sink or knit lines. Unfortunately, when the part shrank during cooling it grabbed aggressively to the core side of the mold. This made it almost impossible to remove by hand, and confirmed my suspicion that ejector pins would be required to remove the part cleanly from the mold. In the interest of protecting both my part and my mold, I removed my insert and began work at adding ejector pins.

Ejectors In Action

With ejectors in place, the part was easily stripped from the mold. This change facilitated production of over one hundred housings in the span of two hours, and was critical to the success of the project.

Diffuser Molding

The diffuser for each light was made from the same polypropylene that the main body was molded from. This helped to mitigate any issues with different shrinkage rates between materials, ensuring that the heat stake posts would line up with their matching holes. Two millimeter wall thickness proved sufficient to eliminate hotspots from the LED and provide a nice diffuse glow. Due to the shallow and small nature of the part, it was done in a multiple cavity mold to reduce production time.

PCB Assembly

After getting a small run of PCBs done in China, I began assembling each of the boards. The first thing done was mount the magnets that pull the light to a magnetic surface. After fixturing the magnets on and letting them glue overnight, I soldered on the remaining surface mount components by hand, and then attached the PCB to the housing with thread forming screws into bosses in the underside. Each assembly was then tested, to ensure that all of the lights would function when assembled.


With all my lights assembled, I decided to design some packaging that would complete the product. I chose ten per package so that I would have a good number of individual packages to give away, and so that anyone with a package would have enough to make a compelling arrangement. The packages were cut from cardboard on a laser cutter, folded and then closed with a large rubber band. Each package showcases the light from the front, and has features that prevent the lights from being activated while in storage.


The final outcome of this project was the closest I have gotten to going from concept to product in the span of one class at Stanford. It was delighting to watch people play with my lights, and take joy in their simple and intuitive interaction along with the sheer number of lights that I had created. This experience helped solidify my understanding of mold design, and resulted in a product that was a fun and educational progression from ideation through development to production.