Classical and machine ciphers, from Caesar’s shift and the ACA’s 62 types to Enigma and Lorenz, and text as notes: enciphered with a key, chained, many broken without one.
SIGABA
Simulates SIGABA, the ECM Mark II (CSP-889, and the Navy’s CSP-2900), as its patent, the Navy’s CSP 1100 (1944) and the Army’s M-134-C instructions describe it, with the message setting from an indicator and the 26–30 check. The index rotors are the USS Pampanito’s; the cipher and control rotors are Pekelney’s, invented. Known plaintext gives the message setting by Lasry’s meet in the middle (2019), or the index setting.
About this tool
History
SIGABA, the Navy’s ECM Mark II, was the U.S. Army’s and Navy’s high-grade rotor machine of the Second World War; Laurance Safford and Donald Seiler filed its patent in December 1944, granted only in 2001. The Navy’s operating instructions, CSP 1100 of May 1944, and the Army’s for the Converter M-134-C give its procedure. No wartime wiring of a cipher or control rotor has been published: Richard Pekelney, who restored the USS Pampanito’s machine, made a set for his emulator in 1996, and every simulator and paper since uses it.
How it works
Fifteen rotors in three banks. The letter passes five cipher rotors. Four live contacts pass five control rotors, the middle three stepping as an odometer; their 26 outputs are tied into nine leads, which five index rotors, set by hand and never moving, carry to five magnets. So one to four cipher rotors step each letter, in thirty patterns. A rotor turns backward through the alphabet, a reversed one forward. On the CSP-2900 six contacts are live and rotors 2 and 4 turn the other way.
The operator zeroized the rotors and turned each control rotor by its key to the indicator; the 26–30 check enciphered A thirty times from zero. Z was keyed as X, a space as Z.
Without the key, Lasry’s meet in the middle runs two cipher rotors forward from the plaintext and three back from the ciphertext; each start that fits eight letters is followed, and every control start looked up.
How it was checked
Eighteen examples published on Pekelney’s rotors (Lasry’s, Sullivan’s, CyberChef’s, Dunn’s) come out letter for letter. The stepping maze of CSP 1100’s stepping table, recorded on a real machine in 1944, comes out at all 56 printed steps to 10,000. The meet in the middle finds the one true key from twelve known letters, keeping about 0.107 stepping sequences a start, as Lasry measured.
Limits
The cipher and control rotors are invented, so no wartime message reads, and the sample key list’s check groups do not come out. That a reversed rotor carries at O, and that the magnets are set before a control rotor moves when the indicator is set, are Pekelney’s readings. The CSP-2900 rests on one machine. The searches are reductions for the CSP-889: with the whole key unknown, about 2^95.6 keys, Lasry needed some 2^60 operations.
- L. F. Safford and D. W. Seiler, Control circuits for electric coding machines, U.S. Patent 6,175,625 (filed 1944, granted 2001)
- Navy Department, ECM Operating Instructions, CSP 1100(C) (1944), transcribed by R. Pekelney
- T. J. Mucklow, The SIGABA/ECM II Cipher Machine: “A Beautiful Idea” (NSA Center for Cryptologic History, 2015)
- R. Pekelney, ECM Mark II emulator and curatorial report, USS Pampanito
- G. Lasry, “A Practical Meet-in-the-Middle Attack on SIGABA”, HistoCrypt 2019, pp. 41–49