TURN A BACKYARD SHED INTO A RAM CHIP FAB
A shed, a furnace, a homemade stepper, and twenty working memory cells that charge in nanoseconds.
by Dr. Semiconductor (Matthew Hartensveld)
- difficulty
- ●●●●●
- time
- Build time not stated by the creator
- license
- license not specified
- repo
- repo FINISHED
●●●●● · Build time not stated by the creator · license not specified · repo FINISHED
WHAT YOU’LL NEED
COMPAREE VERDICT
Matthew Hartensveld turned a backyard shed into a class 100 clean room and fabricated working DRAM cells with tools he built himself. The design is a 5 by 4 array, one transistor and one capacitor per cell, with a gate a bit under a micron. Silicon oxidises in a furnace at 1,100 °C, gets patterned with photoresist under a microscope stepper running custom software, doped with phosphorus, and finished with aluminium contacts from a homemade sputtering machine. The capacitors measured 12.3 picofarads, and a cell charged to 3 volts in a couple hundred nanoseconds. The retention time is just over 2 milliseconds, versus more than 64 in commercial DRAM. It is twenty memory cells, not a memory module, and it cannot go in a PC yet. This is not a how-to and not a home recipe: there is no repository, no parts list, no drawings of the furnace or the stepper, and the process uses hydrofluoric acid and piranha solution, industrial chemicals that can kill. If you want to understand semiconductor fabrication at a human scale, this is a clear window into it. If you want to build RAM for a computer, this is not the project.
IN THE REPO
GOOD TO KNOW
- —No repository — this is documentation of a working proof-of-concept shown on YouTube.
- —No CAD files, no bill of materials, no step-by-step guide to replicate the tools.
- —The video shows the process in detail but does not provide schematics or measurements for the custom equipment.
- —The process uses industrial chemicals, including hydrofluoric acid and piranha solution, that are lethal and not consumer products.
- —No licence declared — this is a demonstration project, not a released design.
Can I build this?
Build at your own risk. Projects involve tools, electronics and sometimes mains voltage — follow the creator’s safety notes.
Videos
Homemade RAM fabrication walkthrough
Full process from design and oxidation to testing the finished cells.
More builds like this
All projectsGallery
Start here
Navigation into the creator’s own docs — we don’t rewrite the guide, we route you to the source.
- 1.Watch the full video (The video is the primary documentation — no written guide exists.)
- 2.Understand why this is not a home recipe(The process uses hydrofluoric acid and piranha solution, lethal industrial chemicals, plus a 1,100 °C furnace. It is not a home recipe.)
- 3.Follow the next step on the channel (The creator's next step is stitching the cells into a much larger array and hooking it up to a PC.)
KNOWN ISSUES
- The video shows the result, not a kit or a parts list — you cannot buy what you need and follow along.
- The process uses hydrofluoric acid and piranha solution. Both are lethal industrial chemicals; this is not a home recipe and must not be attempted from the video.
- The cells hold their charge for a little over 2 milliseconds versus more than 64 in commercial DRAM, and there are only twenty of them — they cannot go in a PC yet.
- The custom stepper software and the sputtering machine design are not published.
- This is a demonstration of what one person did with a custom-built lab, not a weekend build.
Can I build this at home?
No. This is not a home recipe. Matthew Hartensveld built every fabrication tool himself and the process uses hydrofluoric acid and piranha solution, industrial chemicals that can kill. The video shows what is possible, not something to copy.
Where do I get the chemicals?
This page does not cover sourcing or handling chemicals. The process in the video uses hydrofluoric acid and piranha solution, which are lethal industrial chemicals. It is not a home recipe.
How much does it cost?
The video does not give a cost. For scale, Micron's new memory fab in Idaho is a 15 billion dollar project; this shed lab is a one-person build with custom-made tools.
Can these cells go in a PC?
Not yet. There are only twenty cells, and a cell holds its charge for a little over 2 milliseconds versus more than 64 in commercial DRAM, so it needs refreshing much more often. The creator says the next step is stitching the cells into a much larger array, then hooking it up to a PC.
Community builds
No community builds yet — be the first, we feature the best ones.
Discussion1
FROM THE COMPAREE TEAM
Twenty working cells, a 12.3 picofarad capacitor, and a charge time in the hundreds of nanoseconds — all from a backyard shed. What would you want to see first when these cells become a larger array?
Dr. Semiconductor (Matthew Hartensveld)
Matthew Hartensveld (Dr. Semiconductor) turned a shed behind his house into a class 100 clean room, built his own semiconductor fabrication tools from scratch, and fabricated working DRAM cells. He calls it the first time ever RAM has been made at home. The work is supported by his Patreon subscribers.
DISCLAIMER
- Comparee is not the author of the projects featured here. All rights to each project belong to its creator — every page links to the original source, and we never host creators’ files.
- Information is provided without warranty and may become outdated as projects evolve. Prices are indicative bands only — always check the creator’s parts list for current costs.
- Building and operating any project is at your own responsibility. Protective equipment, safe workshop practice and compliance with local regulations are the builder’s responsibility.
CompareeTEAM21d ago
Practical notes from our verification: this is not a replicable build in the usual sense — there is no repository, no schematic, no parts list, and the process relies on hydrofluoric acid and piranha solution, which are lethal. The video documents one person fabricating working DRAM cells in a shed with tools he built himself. The biggest surprise is how close the measured capacitance (12.3 pF) came to the design, and how fast the cells charge (a couple hundred nanoseconds). The biggest limitation is retention time: a little over 2 milliseconds versus more than 64 in commercial DRAM, so the cells need refreshing much more often. It is twenty cells, not a memory module. The YouTube channel is the primary home for the work.