The Secret in Building 26: The Untold Story of America's Ultra War Against the U-boat Enigma Codes

Building the Perfect Machine

March 1943-Dayton, Ohio

In a secure meeting room inside NCR's Building 26, while shotgun-toting Marines stood guard outside, chief engineer Joe Desch grew increasingly impatient as he listened to one staff member after another report on continuing glitches with the two prototypes of the U.S. Bombe, Adam and Eve. After enough bad news, Desch resorted to what was becoming an all-too-familiar motivational technique among his hard-pressed group of seventeen engineers and technicians. He jumped out of his seat and onto the meeting-room table and began pounding his fist into his hand with every word he shouted. "No more excuses! We've got to work harder, faster, smarter! Everybody's ass is on the line!"

What Desch couldn't tell his staff, and what had been pointed out to him repeatedly by his own Navy supervisors, was that too many ships were going down, too many men were dying at sea, while the team failed to produce a working codebusting machine that had been promised for delivery to the Navy three months before.

Because of the project's ironclad security, Desch's staff was not permitted to utter even among themselves the words "Enigma" or "Bombe" or the seemingly innocuous name for the top secret operation, "U.S. Naval Computing Machine Laboratory." The project was self-contained within NCR's former night-school building, constructed seven years before on a large, open tract that had once served as the city dump. Behind the building, on a lonely spur of railroad track, sat an empty baggage car with an overdue delivery date to Washington, D.C.-the Navy's not very subtle way of reminding the project's managers that the top brass was impatient for results.

But OP20G, the Navy unit in charge of analyzing and decoding enemy radio communications, may have been asking for the impossible. As late as August of 1942, the Americans still had high hopes that an all-electronic decoding machine-at least one hundred times faster than anything built before-would be able to crunch through more than four hundred thousand possible Enigma solutions in the unheard-of time of fifty-five seconds.

From those wildly optimistic expectations, the American team plummeted two months later into a misinformed pessimism. Desch then thought his best possible Bombe might take hours to complete a run of all the Enigma possibilities, not just a few seconds, and that the Navy would need 336 of the sophisticated machines to get the job done. A big part of the problem was that the Americans had still not mastered the information the British were supplying about all the challenges in the Shark system, nor did they know all of Bletchley Park's clever methods in attacking them.