Youtuber Successfully Manufactures Working LEDs in Backyard Cleanroom Fab
- Matthew Hartensveld has expanded his backyard fabrication project by successfully manufacturing working gallium nitride LEDs in a class-100 cleanroom built inside a garden shed, according to recent technical...
- Constructing functional computer memory requires attaching finished silicon dies to circuit boards, a process complicated by the fact that silicon is opaque to visible light.
- To attach the newly minted diodes to an ENIG-finished circuit board, Hartensveld utilized indium bumps.
Semiconductor engineer Dr. Matthew Hartensveld has expanded his backyard fabrication project by successfully manufacturing working gallium nitride LEDs in a class-100 cleanroom built inside a garden shed, according to recent technical updates and interviews. The project serves as a crucial packaging test bed for his ongoing efforts to build functional homebrew DRAM chips outside the control of major multinational manufacturers.
Why Gallium Nitride LEDs Matter for Backyard RAM Fabrication
Constructing functional computer memory requires attaching finished silicon dies to circuit boards, a process complicated by the fact that silicon is opaque to visible light. This opacity makes it difficult to align tiny contacts with corresponding pads on a circuit board, prompting Hartensveld to experiment with gallium nitride, or GaN. Because GaN is transparent, it provides an ideal test bed for home packaging experimentation while also allowing the engineer to fabricate basic blue-emitting diodes.
Etching GaN traditionally relies on hazardous chlorine gases, which Hartensveld noted in his video documentation are both impractical and unsafe for home use. Instead, the engineer deployed a 355-nanometer ultraviolet laser etcher that the GaN material absorbs. After verifying the etching process, he applied a standard production flow involving photoresist, metal deposition for p-type and n-type contacts, and the addition of indium to form working quantum wells that emit blue light under power.
Packaging and Color-Shifting Techniques in the Shed Fab
To attach the newly minted diodes to an ENIG-finished circuit board, Hartensveld utilized indium bumps. Because the LED material is transparent, footage captured the exact moment the indium contacts melted and created a solid bond. Furthermore, the engineer successfully altered the light output to a warmer white by applying cerium-doped yttrium aluminum garnet, the same yellowish material found in commercial lightbulbs.
These packaging milestones directly support the creator’s long-term goal of producing homebrew computer memory. Working DRAM cells represent an early step, but transforming those cells into functional DDR5 chips requires sophisticated micro-packaging techniques that avoid the high costs associated with industrial foundries.
Next Steps for Open-Source Semiconductor Manufacturing
Hartensveld intends to apply the newly refined packaging methods to RAM modules after completing additional development work. According to correspondence with The Register, the engineer’s immediate priority involves shrinking both transistors and capacitors to increase density and mimic modern commercial DRAM techniques.

My upcoming objective involves reducing the size of the capacitor and the transistor — with the former being especially crucial — allowing me to raise the density and begin testing various methods applied in contemporary commercial DRAM.
Dr. Matthew Hartensveld
The project is documented openly on the semiconductor.diy website, which includes bills of materials and instructional guides for both the cleanroom shed and the LED manufacturing process. The initiative receives support from Vensa, a nonprofit backed by Ethereum cofounder Vitalik Buterin that promotes open-source silicon development to improve computer safety and technological sovereignty.
Opening up semiconductor production and fostering a localized, grassroots chip-making network is certainly a key component of the overarching goal.
Dr. Matthew Hartensveld
While the creator does not expect a garden shed to replace billion-dollar industrial plants, the work demonstrates that hobbyists, researchers, and startups can execute complex microfabrication tasks using simpler, lower-cost equipment.
