By Bob Atha, the technician who assembled, tested, maintained and ultimate destroyed the bombe.

From late 1943 through the end of World War II, I was a maintenance watch chief in charge of a crew servicing bombes installed at the Naval Security Station, 3801 Nebraska Avenue, N.W., Washington, D. C. I helped assemble and check them out at the National Cash Register (NCR) Company, Dayton, Ohio, helped install them when we received them from NCR, helped maintain them for the duration of the war, and helped destroy them when they were declared obsolete at the end of the war.

How did I come to be involved with OP-20G and the bombes? Simple! I received orders and followed them! I initially served as a Naval Reserve radio operator on board several ships including USS BROOKLYN from which I was transferred about a month before Pearl Harbor. I was subsequently assigned to duties as a radio intercept and D/F operator at Cheltenham, Maryland; Chatham, Massachusetts; and Muirkirk, Maryland. Before joining the Navy, I had received a BS in Electrical Engineering at New Mexico State University and worked as a junior engineer for the Gulf States Utilities Company, Beaumont, Texas. While in high school, I became interested in amateur radio and received my ham operator’s license that I hold yet to this day.

I was assigned to the bombe program about a month after I married my childhood sweetheart. We had settled in our new apartment in Congress Heights on the southeast side of Washington, D. C. when I received orders to Dayton, Ohio. We sublet the apartment and headed west for a “second honeymoon”. Some honeymoon! Starting in August 1943, I worked on the bombes from about 0800 to 2200-2300 at the National Cash Register Company. Under the direction of a civilian technician, I helped assemble and test the machines and, in so doing, learned by the process of osmosis how to trouble-shoot, repair and maintain them. By the end of the six weeks, I could do just about anything on the bombe except a few things like relay adjustment and some chassis repair.

After this “training” period at NCR, we returned to Washington, D. C. I never received a briefing on the purpose of the bombe or its theory of operation, and I don’t remember ever seeing a schematic for the machine. However, I had apparently learned enough. Upon my return to Washington, D. C., I was assigned to 3801 Nebraska Avenue, as chief of a maintenance crew and stood watches until the end of the bombe program. Of course, I just didn’t walk in and start standing watches. First, we had to receive and install the bombes we had helped assemble at Dayton, and get them working. Once operational, a bombe never stopped running except for setup time and maintenance until the end of the war.

I had never seen or heard of 3801 Nebraska Avenue before reporting for duty. Upon arrival, I saw the buildings of a former girls’ school, which had been taken over by the Navy, enclosed by two tall chain-link fences guarded by Marines. It was extremely secure. Across the street were barracks for the WAVE personnel who operated the bombes. The bombes themselves were installed in a new two-story building that had been specifically designed for them. In the basement were rooms for tools, supplies, lockers and a dressing room with showers for the men, DC generating equipment, an air compressor, and a loading platform. A large elevator with a capacity of one bombe ran from the basement to the upper two floors. Installed on each floor were 48 regular bombes plus two additional machines in a separate room. These two machines had been modified so they were joined both mechanically and electrically. The regular maintenance crews were not called to service them. Counting these special machines as two machines, there were 50 bombes on each floor for a total of 100 bombes overall in the building. In addition to a lot of noise, the various electrical relays, motors, and vacuum tubes inside the bombes generated a lot of heat. This required installation of large, wall-mounted, commercial Carrier air conditioners, one between each outside window in both machine rooms. These air conditioners ran continuously winter and summer because neither we nor the machines could have endured and functioned without them. We were comfortable working on the bombes during both winter and summer. However, we sweltered in the hot, humid Washington summer when we went home.

The exact function of the bombe equipment was not explained to me. Because of the strict need-to-know practice imposed, this total knowledge was probably known to only a few analysts and design engineers. Of course the basic purpose of the machine was to do the exact opposite of the enciphering machines, without full knowledge of the crypt-variances involved. It was done at high speed, for that time. Although I never heard the word enigma mentioned, while working on the bombe program, I knew what an enigma message was, because I had copied many of them while I was an intercept and D/F operator listening to German submarine traffic.

Mechanical Operations

Actually the bombe accomplished its purpose by rotating electrically connected commutator wheels by means of a gear box driven by a motor, and sensing, electronically for possible electrical pathways through the wheels and Switchbank settings, then recording them on a printer, when the machine was at rest.

The bottom two rows of wheels on a machine were rotating too fast to allow the scanning relays and the printer to record the desired information. So, a memory system was incorporated which recorded the position of the two lowest rows of wheels on a special machine at the instant the pathway was sensed. This was called the “hit” signal. Then the machine would stop its forward motion and rewind the wheels to the “hit” position, scan and print out information on the printer, after which the wheels would resume their forward rotation until the end of the run.

The bombe was an electro-mechanical device designed with a number of removable chassis’ (or modules) to reduce maintenance downtime. This worked well to combat electrical problems and procedures were developed for isolating malfunctions to a specific chassis. After a problem was isolated, the defective chassis would be replaced with a spare, and sent to the maintenance laboratory for repair at a later time. This included the relay chassis, which probably included the large hit thyratron tube, and the Memory chassis, which contained about 50 miniature thyratron tubes. These miniature tubes received their firing voltage through fixed segments on the bottom two wheel levels of the machine and were activated by the “hit” signal.

Located on top of the machine’s main frame were the Switchbank chassis’, one for the inputs and outputs of the top sets of brush holders in each row on each side of the machine. We didn’t have much trouble with these. They were removable and replaceable but with difficulty because they were heavy and hard to grasp. All inputs and outputs of the two Switchbanks were connected to one end of the Diagonal Board. The Diagonal Board was the largest of the removable chassis’ and was located in a chassis cabinet just above the Switchbanks nearest the printer. On another side of the Diagonal Board were rows of female Jones sockets. Cables from the Printer were plugged into selected outputs of these sockets. There were also connections to the “Hit” Thyratron chassis, the Relay Scan chassis and perhaps others. The Diagonal Board contained a large number of special diode tubes, each the size of miniature glass receiving tubes containing one cathode and four separate anodes. I don’t know many tubes were in this chassis — perhaps over 600 or even more. Through its connections to the Switchbanks and wheels, the Diagonal Board sensed the “hit” and sent the signal on to the “hit” thyratron and memory thyratrons in other chassis’.

I was told little or nothing about the Primary and Secondary Scanning Relay chassis. I got into trouble once by trying to fix one of these relays while it was on the machine. I was later told to keep my hands off but wasn’t told why. I got the thing running again but probably incorrectly, from their point of view. There was also a button on the front of this chassis that, if pressed when the machine was at rest, would cause these relays to begin scanning. Their operation was monitored by observing two rows of neon lights that glowed sequentially during a scan cycle. If the sequence stopped, either the Printer would start operating or something was wrong elsewhere, I can’t recall. This procedure might have been part of the pre- and post-operation checks. The only other chassis I recall contained time delay relays for the application of voltages.

Mechanical repairs or adjustments had to be performed on the machine; there were no quick mechanical substitutions. The mechanical parts of the machine included a three-phase drive motor, a brake to slow down and stop the mechanical motion of the machine, a clutch to engage the rewind motor to the drive system, a slow speed rewind motor, an indexing pin system to lock the gear train in the correct segment position during the print cycle, a disengagement mechanism for the top row of wheels to prevent their movement and allow them to be set at any selected position, and the gear train which provided rotation to the wheels so a full rotation of one set of wheels caused the wheels in the next highest row to rotate one segment.

One of the most annoying mechanical problems was broken carbon commutator brushes. There were 6,656 brushes per machine. To replace a brush, you had to loosen a brush-holder and turn it so you could work from the back to install a new brush. While installing the new brush, you had to be careful not to break any of the other exposed brushes. As a result, we had more mechanical-type than electronic-type personnel assigned to maintenance crews.

Bombes had other problems which routine machine checks wouldn’t reveal. One was undetectable except by reviewing a machine record of “hits” during runs. A “hit” represented a desired piece of continuity information on a machine during a “run”. When there was a history of no “hits” during several sequential runs, the operators would submit a trouble report. Invariably the culprit was the large “hit” thyratron tube, probably due to reduced cathode emission reducing sensitivity to the duration and/or amplitude of the triggering voltage of the “hit” signal. Repair was simple — replace the thyratron tube!

Sometimes a machine would register a “hit”, stop, start to rewind and continue rewinding until an operator stopped it. This was mainly due to a malfunction in the Memory chassis, probably a small memory thyratron tube that didn’t fire. There was one tube for each segment of a wheel position on the bottom two rows of wheels on the machine. The solution was to replace the Memory chassis. Occasionally a machine would stop normally but the Printer would begin printing out rows of numbers without stopping until an operator pressed the stop button. The solution was to replace the Printer. There were many, many other things that could and sometimes did go wrong. The machines kept us busy.

The machine used four sets of commutator wheels, 16 wheels per set, at a time. A fifth set of wheels was stored in a cabinet when not in use. There were four concentric rings of commutator segments on a wheel. The wiring to the commutator segments, located on one side of a wheel, was identical for each wheel in a set. The other side of the wheel had a black cover with an identifying color spot also identical for each wheel in the set. The wheels were probably the electrical equivalent, but not the mechanical design, of those in the German enigma machines.

As stated, when loaded the machine contained four horizontal rows of wheels on each side. The bottom row contained the high-speed wheels. When the wheels in one row completed one full revolution, the wheels in the row immediately above rotated one segment. This sequence was routinely followed in rows one through three, as well as row four if they were engaged. When not engaged, the top row of wheels was pre-set to a designated segment and a “run” took about 10-12 minutes to complete. A long run, with the top row of wheels engaged, took considerably longer.

************

The bombes were operated by the WAVES who lived in the barracks across the street from the Naval Security Station. I would estimate there were some 300 of them plus their officers working on the bombes. A team of two WAVES operated four bombes. Each team received its crypto-assignment, essentially instructions on the wheels, wheel sets, their location, wheel off-set, connections from the Diagonal Board to the Printer, switch settings on the Switchbanks, and type of run – long or short, from its supervisor. After the “run”, the results recorded on the Printer located at the end of the machine were checked and turned over to the team’s supervisor. It took the two WAVE operators, working together, to do the machine checks that were performed before and after each “run”. If a problem was indicated by the machine check or during a “run,” an operator would make out a trouble report and bring it to the Maintenance Watch Chief of which I was one.

While working on the bombe program, my wife and I kept our apartment in Congress Heights and I commuted diagonally across the District of Columbia to Nebraska Avenue via bus and streetcar. A weekly pass allowing travel anywhere anytime on the system only cost $1.25. Commuting was generally uneventful but two events stick out in my mind. I once saw President Harry S. Truman, my favorite President, taking his morning walk on Pennsylvania Avenue near the White House, his Secret Service bodyguards trailing about a half block or more behind. The other occasion was V-E Day. I was on my way home from work and had to change to a bus near the FBI Building. It was between 0000 and 0100. All of the lights in downtown Washington were turned on. Crowds of people were everywhere, including in the streets. Everyone was shouting and cheering. Spotlights were focused on the flag atop the Capitol Building. But the next night, and until the end of the war with Japan, we returned to black-out conditions in Washington.

I helped install the bombes and I helped destroy them. When the war was over the bombes had served their purpose and were now obsolete. I was in charge of the group of maintenance personnel who were ordered to destroy them beyond recognition. Using screwdrivers, sidecutters, pliers, wrenches, crowbars, sledge hammers and a big, hot fire for the combustibles, we did just that. At the time I believed we had done a complete destruction job. However, I was told recently that one or two survived. If true, I don’t know how they got away.

**Featured image is the Bombe that “got away” located at the National Cryptologic Museum,