The "bombe" was a significant development in the evolution of the Allied cryptanalytic attack on the German Enigma cipher machine. From 1928 to 1938, Polish cryptanalysts, aided by the French, solved both the wiring of the three rotors used at the time and the indicator system, and through the use of a small electrical machine using Enigma rotors, created a card file which could anticipate the 6 by 17,576 possible rotor positions used in setting up the Enigma. In 1938, however, the Germans changed this entire indicator system.

To counter this move, the Poles created a hand system with six sets of perforated sheets (each with 26 sheets) which recorded about 1,000 possibilities on each sheet. When suitable combinations of the sheets were superimposed, they enabled the Poles (and later Bletchley Park) to determine the daily settings for Enigma. At the same time, they devised a method of recovery consisting of six electrically powered sets of Enigma wheels that rotated to search for repetitions in the "message key" (the starting position of the rotors for a specific message)."

However, in December 1938 the Germans added two more rotors to the three-wheel Enigma. Instead of six ways of arranging the rotors, there were now sixty. Therefore, recovery required sixty perforated sets (each containing 26 sheets) instead of six sets.  Although the Poles soon solved the wiring, they simply did not have the resources to advance further on the bombe. At this point, the British assumed leadership of the cryptanalytic attack on Enigma. Using all of the Polish data, and that supplied by the French, the British decided on an entirely different approach - a general method of recovering the wheel settings that did not depend on solving the indicator system. Alan Turing and other British analysts solved the problem, and the Turing prototype bombe appeared in May 1940. The first British high-speed bombe was produced in April 1943. It was a four-rotor bombe but Bletchley Park got no real value from it until June [1].

In August of the same year, the Americans produced their first bombe. Thereafter it was a matter of refining the technical operation of the bombe and applying as many resources to the problem as it took to maintain the production of intelligence. Production of the USN bombes was pretty constant. Eighty bombes had arrived in Washington DC by December 1943, 90 by June, and 115 by December 1944. The Germans made very few changes to the operation of M-4 naval  Enigma, which was Op-20-G's primary target.

Early in WWII, President Roosevelt signed the bill establishing the Women's Army Auxiliary Corps on May 15, 1942. A few months later he signed a similar bill for the US Navy which created the Women Accepted for Voluntary Emergency Service, more commonly known as the WAVES. Only those women meeting higher qualifications were admitted into cryptologic work and many operated the bombe.

The Navy also told their WAVES as little as possible in order to maintain secrecy. They sent some 600 newly inducted WAVES, along with 200 men, to Dayton, Ohio, to help build and train on the cryptanalytic Bombes manufactured by the National Cash Register Company . Each Bombe required sixty-four rotors to be wired to match the rotors actually used in the Germans' Enigma. WAVES performed this task. Each woman was given a wiring diagram for one side of the two-sided rotor. She spent her eight-hour shift soldering wires to rotors. Another WAVE soldered the other side. This arrangement maintained the secrecy of the rotor wiring. No WAVE would have knowledge of both sides of any rotor.

Close-up views of a bombe rotor. The top photo shows the rotor wiring. In the photo below it, details of the cover (left) and commutator (right) can be seen. These components are on display at the NCM. (Photos by Jerry Proc)

All above: Navy WAVEs operating  bombes at the Naval Communications Annex.

All photos in this table courtesy NSA.

 

M-9 bomb checker. (Photo by Jerry Proc)  Since the Allies did not know the German's daily rotor selections, several Bombes worked on the same message. Each Bombe tested a different set of wheel orders. The Bombes usually found two sets of possible rotor settings on each run , but only one solution on one Bombe was the correct wheel order and rotor position used by the Germans for that day.  After the Bombe completed a run, a WAVE supervisor checked the printed results on the M-9 Bombe checker. She checked each result looking for the correct one.  Once she found the results , she used the M-9 to fill in any missing plugboard positions. The Bombes could only find a portion of the Stecker positions because the menus were between thirteen and sixteen letters long which was too short to find all the plugboard connections.  Having found the correct wheel order, rotor position and Steckers, the supervisor then sent the results back to the library where WAVES and cryptananalysts used an analog and decrypted the message. Short messages could be decrypted directly on the M-9 to work against messages that had other problems such as garbles.

Eventually, the Navy built 121 Bombes and sent them to the Naval Communications Annex. Here, women began operating the machines. Three shifts of women ran them twenty-four hours a day. The work was noisy and hot, but the Navy had impressed upon them the importance of their work, without ever explaining how their work fit into the whole cryptanalytic effort. Commander Gilman McDonnell, a supervisor on the Bombe deck, recalled an incident in which a rotor solution, known as a jackpot, was accidentally thrown out. The resulting delay made it apparent to the chief of the Bombe operations, John Howard, that some kind of explanation was necessary.

Eventually we got the answer and then they realized that that wheel order had already been run and should've been a jackpot. Well, it had obviously been thrown in the burnbag by mistake. So John got permission from his superiors to tell the girls something to give them some sense of how important the work was. We shut down operations for about ten minutes and he got up on a chair in the middle of the operations area. He didn't tell them specifically what they were doing, but said they were attacking an  enemy cipher and it was a very important job....and that's all they knew.

Front view. The NCR Bombe as it is displayed in the National Cryptologic Museum. (Photo by Jerry Proc)

Rear view of Bombe. (Photo by Jerry Proc)

Above: Various closer views of the Bombe's dial indicators.

All photos in this table by Ralph Simpson

DEVICE ASPECTS:

Rate At Which Machine Tries Assumptions: 20,280 tries per second.
Running Time: 50 seconds for a 3 rotor Enigma run.
                         20 minutes for a  4 rotor Enigma run.
Rotor speed: 1725 R.P.M.
Number of bombes produced :121
Type of Detection: Prints only correct "stories" which satisfy all the restrictions of the input menu.
Operations Personnel: 700
Maintenance Personnel: 135
Dimensions: 8 feet wide, 7 feet high and 2 feet deep.
Weight : 5,000 pounds
Cost for 121 bombes: $6 million

The Op-20-G bombes were constructed at the U.S. Naval Computing Machine Laboratory. Each bombe was the equivalent of 16 Enigma machines and consisted of a set of 4 spindles and brush holders on which cross wired wheels were loaded by hand.

MORE ON THE BOMBE

In February 2001, The Dayton Daily News published an 8 part story featuring Joe Desch, the man who led the effort to manufacture a practical, working machine. Select this link to read the story titled "Dayton's Codebreakers".

The University of Dayton plans to demolish the building where the bombe was developed. Efforts to save the buildings are told in this New York Times story dated April 1/2007.

FOOTNOTES:

[1] C. H. O'D. Alexander, Cryptographic History of Work on the German Naval Enigma, p 53 (National Archives, Public Record Office, Kew, Surrey, HW 25/1)

---

Contributors and Credits:

1) The Bombe: Prelude to Modern Cryptanalysis. NSA publication.
2) Sharing the Burden: Women in Cryptology in WWII. NSA publication.
3) Solving The Enigma: History of the Cryptanalytic Bombe. NSA publication.
4) Ralph Simpson <ralphenator(at)gmail.com>
5) NCM Bombe placard

Aug 1/23