Making RAM at Home [video]
A viral YouTube project shows an engineer fabricating dynamic RAM chips in a self-built clean room shed, complete with modified microscopes and homemade plasma etching gear. Commenters are both impressed by how far hobbyist semiconductor fabrication has come and clear about the limits: the process is expensive, labor‑intensive, dangerous due to chemicals like HF and phosphine, and nowhere near commercially viable yields. The thread also touches on rising RAM prices, the role of niche technical content on platforms like YouTube and Patreon, and broader hopes and skepticism about “backyard” manufacturing as a path to computing independence.
Overall reaction to “homemade RAM in a shed”
- Many commenters are amazed by the sophistication of a backyard clean room and RAM fabrication effort, calling it “incredible” and “exactly the sort of thing” they come to see.
- Some stress that this is far from casual DIY: estimate of ~$20k in materials, continuous air circulation, daily cleaning, and deionized water systems.
- Others note that the creator is a highly skilled semiconductor engineer; this is not something “regular people” can realistically reproduce to a reliable production standard.
Clean room, tools, and process
- People are impressed by the modified microscope acting as a photolithography tool and the homemade plasma etcher; the furnace itself is noted as an off‑the‑shelf Chinese unit.
- Positive-pressure backyard clean room is described as almost “mystical” to achieve.
- Some compare the feat favorably to early semiconductor fabs, noting how far individual tinkerers can now go.
How DRAM actually works
- Multiple replies explain DRAM operation for confused readers:
- Tiny capacitors store charge that leaks quickly and must be refreshed.
- Reading intentionally disturbs the charge and uses sense amplifiers to detect a tiny voltage change, then restore it.
- The demo chip only has a small array of cells and no full sense/refresh circuitry yet.
Safety and chemistry concerns
- Strong warnings about dangerous chemicals, especially phosphine and hydrofluoric acid, which are common in real fabs and can be lethal even when “empty” tanks are handled.
- Several argue this is a major reason home semiconductor work shouldn’t be widely replicated, contrasting it with much safer electronics tinkering (e.g., 7400 logic, PCBs, 3D printing).
Economics, shortages, and “artisanal RAM”
- Discussion of RAM price spikes driven by AI demand; some in infrastructure/healthcare say they’re already seeing large cost increases.
- Skeptics argue small-scale fabs can’t compete: yields would be poor, per‑unit cost far above industrial RAM, and the market is highly commoditized.
- Jokes and semi-serious ideas about “homegrown” or “small-batch” RAM as a niche, but consensus is that scaling this into a viable business is unrealistic.
Meta: relevance, video format, and future tech
- Some worry videos don’t do well on Hacker News; others say this kind of deep technical tinkering is precisely on-topic.
- Side discussions touch on core memory, neuromorphic computing, open hardware fab projects (like HackerFab), and whether advanced (e.g., 7 nm) home processes might ever be possible—most treating that as far-future or unrealistic.
- A few frame home fabrication as important for long-term “computing freedom,” especially if large corporations and states dominate chip production.