History: from a paper cheque to a chip

1967 — the first cash machine. On 27 June in London, at the Barclays branch in Enfield, a machine by De La Rue, invented by John Shepherd-Barron, started working. There were no cards yet: a customer bought paper cheques at the bank, impregnated with weakly radioactive carbon-14. The machine recognised the cheque by its radiation, checked it against a numeric code typed on the keypad and gave out a fixed bundle — 10 pounds. The cheque stayed in the machine and went to the bank to take the money off the account.

cheque1967paper with C-14,£10PIN1966–67a number onlythe owner knowsstripe1969–71a machine readsthe numbernetwork1970sthe balance isin the bankchip1990sa secret key,challenge-responsea stripe can be copied, a chip cannot
Sixty years of ATMs: from a paper cheque the machine recognised by a radioactive mark to a chip with a secret key. Every step closed a hole left by the one before.

1966–1967 — the PIN. The Scottish engineer James Goodfellow patented the idea: a card with a code that a machine reads, and a personal number that only the owner knows. A match means the owner holds the card. That is how the PIN appeared.

1969–1971 — the magnetic stripe. A card with a magnetic stripe, like on passes, makes cheques unnecessary: a machine reads the number, and the card stays with the customer. But ATMs still work on their own: they do not know the balance, and stolen cards are caught by a list the bank loads into the machine.

1970s — the network. ATMs are connected to the bank computer. Now the machine asks the bank: is there money, is the card stolen, is the PIN right. The balance lives in the bank, and the ATM only asks and pays out.

The trouble with the magnetic stripe

The stripe stores the card number — and that is all. It can be read and written onto a blank card in seconds: the copy cannot be told from the original. Hence skimmers — overlays on the ATM slot — and spying on the PIN with a camera. Against a copy you need a card that cannot be copied: one with a secret that never leaves it.

The chip and challenge-response

1974–1984 — France. The inventor Roland Moreno patents a card with its own memory (1974), the company Bull makes the CP8 card with a built-in microprocessor (1979), and in 1984 French banks trial chip cards in Blois, Caen and Lyon. ATMs of that time look like ours: a glowing segment display, a keypad with digits and coloured keys, a card slot, and inside — a small processor and logic around it. By 1992 every French bank card has a chip.

DAB · 24/24INTRODUISEZVOTRE CARTE1234567890ANNULATIONCORRECTIONVALIDATIONa segment displayours: a segment LCDa slot for a chip cardours: a slot for two cardsdigits + cancel / enterours: 12 keysthe cash shutterours: a number on the screen
An ATM of the early 1980s — as it was seen in the French towns where chip cards were trialled: a glowing segment display, a keypad with three coloured keys — red "cancel", yellow "correct", green "confirm" — a card slot and a cash shutter. Our ATM is built the same way: a segment display, a keypad, a slot and a processor inside. A composite drawing, no brand.

1990s — EMV. Banks agree on chip cards. The chip is a computer with a secret key. The ATM gives it a random challenge, the chip answers with a cipher — and the bank checks the answer with its copy of the key. The number and the PIN can be stolen, the key cannot: it cannot be read even from the card itself. A stripe can be copied, a chip cannot.

That is exactly what we will build: an ATM that knows neither the PIN nor the key, yet tells a genuine card from a fake.