Hebern rotor machine was one of the first cryptographic rotor machines that allowed to encrypt messages automatically and effectively, and was supposed to provide more complex cryptographic algorithms than the ciphers used manually.
Edward Hugh Hebern created his first rotor machine in 1917 and patented it one year later. By that time, he had already invented several electric machines that were supposed to be used for message encryption and decryption. His cryptographic rotor machines had never become popular, due to some lacks in design, which made the US Army not to purchase more than a few copies.
Similar to other cryptographic rotor machines, the Hebern machine used a disk with electrical wires to encode and decode characters. Each rotor contained 26 electrical contacts on either side. While rotating, the contacts changed the connections to the wires on both sides of the disc. The wires on both sides of the disc were connected to input and output characters, so the rotating disc acted as a simple substitution cipher.
The rotor installed in the Hebern machine rotated a gear each time a key was pressed. The secret key in this case might be presented as an internal wiring of the rotor. Because the rotor had 26 connections, the key settings would be reused after 26 characters. Attacking such a cipher would be a relatively easy task, and the amount of work would be comparable to attacking old polyalphabetic substitution ciphers.
Over time, to make the key size longer, Hebern added additional rotors to the machine. All input letters were passing through all rotors, which means that every letter was changed several times, before it was delivered as the output ciphertext character.
The first rotor moved after a keystroke, while each next rotor turned once after the previous one rotated a full turn.
The machine was proved to be unsecure by an American cryptographer, William Frederick Friedman. He proved that due to the fact that the rotors moved only when the previous disc had rotated a full turn, the whole algorithm might be divided into a number of single substitution ciphers, each one with 26-letter long texts. This mean that the encryption was easy to break by using common frequency analysis methods.