← Back to Research
Research Note · Computer Systems

From Vacuum Tubes to Modern Chips

Published: Jul 2026•6 min read•Author: Chetraj Jaishi
Computer ArchitectureVacuum TubesHardware HistoryBinary Logic
Abstract: How does a computer actually compute? Tracing computing history from 1804 punch cards to 1940s vacuum triodes, relays, and early binary switching logic.

✦HOW DOES A COMPUTER COMPUTE?

Isn't it kind of wild if you think about it? A device that can store, compute, calculate. At first these machines could only do basic math, but now a single chip runs your entire laptop. How did we even get here?
I had the same question for a while and just brushed it off with "it just works." But it kept bugging me - like how does a tiny memory card hold GBs, even TBs of space? It's not a fridge or a storage room where you physically stack things. The data exists digitally... but how? Could I actually build something like that myself? So I went digging into the history of computing to understand it properly, and here's what I found.

✦1. ON/OFF PUNCH CARDS (1804)

Hole = on (1), no hole = off (0). Pure physical logic - a hole was literally the symbol for a bit.

✦2. VACUUM TUBES / RELAYS (1940s)

Picture a tube kind of like a short, bulky thermometer that needs electric current to work. Genuinely amazed people were engineering something this precise back then.
Diagram Placeholderin this area - schematic of a vacuum triode tube, showing filament, grid, plate, glass envelope, tube terminals, and A/B/C-battery connections
Research Schematic
Figure: in this area - schematic of a vacuum triode tube, showing filament, grid, plate, glass envelope, tube terminals, and A/B/C-battery connections
Parts of the tube: grid, glass envelope, filament, plate, tube terminals, and three batteries (A, B, C) powering it.
Function: these acted as electric switches - they'd hold their last output state and could be flipped between 0 and 1.
How it actually works: current flows from the cathode, and electrons jump across to the anode. The grid sits in between and acts as the gatekeeper - it's supplied its own voltage and decides whether the electrons are allowed to jump or not. More voltage on the grid lets electrons through (flow = 1), less/no voltage blocks them (no flow = 0). This on/off electron gate is basically the same logic all the basic gates (AND, OR, NAND, NOR) are built on top of.
Diagram Placeholderin this area - diagram of electron flow between cathode, grid, and anode
Research Schematic
Figure: in this area - diagram of electron flow between cathode, grid, and anode

✦DOWNSIDES OF VACUUM TUBES

•
Needed a constant, continuous current supply to keep working
•
Bulky, not portable, and slow
•
Fell apart trying to handle large/complex data
Eventually (about 4-6 years later) this got replaced by magnetic-core memory. Worth noting though - even with that upgrade, we were still working in plain binary, 0s and 1s.
DID YOU KNOW?
ENIAC used over 17,000 vacuum tubes, 70,000 resistors, and ran on decimal arithmetic. It could do 5,000 additions per second, weighed 30 tons, and pulled 150KW of electricity to run.
---
to be continued

Key Research Takeaways

  • Physical Representation of Bits: Computing began with physical holes in 1804 punch cards (hole = 1, no hole = 0).
  • Thermionic Switching: Vacuum triodes functioned as voltage-controlled switches by using a grid to throttle electron flow between cathode and anode.
  • Foundation of Logic Gates: The binary on/off electron gate provided the physical substrate for elementary logic gates (AND, OR, NAND, NOR).
  • Early Engineering Constraints: Early machines like ENIAC required massive power (150KW) and thousands of tubes before solid-state transistors took over.