The Torpedo Data Computer (TDC), Mark III aboard USS PAMPANITO in San Francisco has been successfully restored to operating condition. The TDC is the electromechanical analog computer that solved the torpedo targeting problem in the fleet submarines during World War II. The restoration project took over 18 months to complete, and was done with the support of Russell Booth, director of the USS PAMP ANITO museum. We believe that restoring this historically significant device to an operating condition is the best means of preservation. The TDC Mark III computer is one of the two remaining examples of the TDC Mark III still installed in a museum fleet submarine.
How It Worked
The TDC was unique in World Warn. It was the computational part of the first submerged integrated fire control system that could track a target and continuous) y aim torpedoes by setting their gyro angles. The TDC Mark III gave the U.S. fleet submarine the ability to fire torpedoes without first estimating a future firing position, changing the ship’s course, or steering to that position. Instead of hoping that nothing in the setup changed, a fleet submarine with the TDC could fire at the target when the captain judged the probability of making hits to be optimal.
In World War II a torpedo’s gyro angle was set mechanically while it was in the tube. A shaft, known as the spindle, slipped into a socket near the housing of the torpedo’s course gyroscope. When the fire control system rotated the shaft, the gyroscope rotated. After being fired, the torpedo traveled on a straight course for a known distance called the reach. A delay in the release of the torpedo’s gyro steering mechanism by a threaded shaft determined the magnitude of the reach. Once engaged, the steering mechanism brought the torpedo to a new course based on the angular offset of the gyroscope.
The Mark III computer consisted of two sections, the position keeper and the angle solver. The position keeper tracked the target and predicted its current position. To do this, the position keeper automatically received input of the ship’s own course from the gyro compass, and own ship’s speed from the pit log. The position keeper had hand cranks on its face that set the target length, estimated speed, and angle on the bow. It also contained a sound bearing converter that calculated the target’s location based on sonar measurements.
The position keeper solved the equations of motion integrated over time. The result was a continuous prediction of where the target was at any instant. Successive measurements of the targets’ position were compared to the position keeper predictions and corrections for error were introduced with the hand cranks. The predicted target position became more accurate as more measurements made the corrections smaller. It was typical to get an accurate track on the target after about three or four observations under good conditions.
The angle solver automatically took the target’s predicted position from the position keeper, combined it with the tactical properties of the torpedo, and solved for the torpedo gyro angle. Values calculated from this solution were returned to the position keeper in two feedback loops. The gyro angle automatically went to each of the torpedo rooms and set into the torpedoes continuously. The TDC controlled both torpedo rooms and all 10 torpedo tubes at once.
The U.S. Navy thus had a system that would point the torpedoes at a target as the fire control problem developed. The TDC Mark III was the only torpedo targeting system of the time that both solved for the gyro angle and tracked the target in real time. The comparable systems used by both Germany and Japan could compute and set the gyro angle for a fixed time in the future, but did not track the target. Thus the idea of the position keeper, and its iterative reduction of target position error was unique to the U.S. Navy, and represented a distinct advantage.
TDC Development History
The U.S. Navy contracted with the ARMA Corporation for the first TDC. The first Mark I was installed and tested in USS SEAL in 1938. The Mark I was a large device, and could not fit in the small space available in the fleet submarine’s conning tower. Instead it doubled as the navigator’s chart table in the control room, and had to be cleared off when running an attack problem because the dials showing the calculations were under the glass table top.
To install a Mark I in the submarine’s control room required it to come in pieces, and be reassembled in place. To make up for the computer being in the control room, an electrically controlled remote plotter in the conning tower kept the captain up to date on the attack. The captain and the executive officer running the computer would yell at each other through the open conning tower hatch. The Mark I worked, but was too big. Plotting the development of the attack in both the control room and the conning tower split up the attack party and limited their effectiveness. It became apparent that a truly integrated system had to fit in the conning tower. ARMA only produced 28 Mark I machines. Before the end of production the design of a smaller machine started.
During the same period, the Ford Instrument Company developed an alternative model, the TDC Mark II. Its use overlapped that of the Mark I developed by ARMA. Designed by the head engineer of Ford, William Newell, the Mark II machine featured a very innovative mechanical solution for the targeting problem. This permitted the device to be small enough to fit in the conning tower where the action was. Ford was too busy with surface fleet computer contracts to even consider bidding on a contract for the Mark II model. It appears that only 12 Mark II TDC computers were built.
Before Mark I production was over, and not knowing of the Mark II project, ARMA accepted a contract for the development of the TDC Mark III. This device was very successful and turned out to be the major submarine computer in World Warn. As the U.S. entered the war most of the earlier models of TDC were replaced with the TDC Mark mas machines were available and submarines came in for refit. A testimony to the significance of the design was that during the entire war period only five alterations were made to the original TDC Mark m design.
From personal interviews and memoirs of submarine captains, one is left with an impression of respect and appreciation for the TDC Marks I and III. Even early in the war when the torpedoes failed to explode, they were usually on target. A Japanese captain after the war recalled that in the beginning U.S. submarines made their ships look like porcupines with impaled torpedoes, and that they knew right away when the exploder started working. [Editor’s Note: Early in WWII the Mk 6 magnetic exploder, in use with the standard Mk 14 torpedo, failed to perform as designed and it was not deactivated for almost two years (eight months later in the Southwest Pacific theater).
About a year after the end of World War ll the TDC Mark IV was introduced as a field installed upgrade kit for the existing Mark ID systems. The modification added a third piece called the Receiver Section, inserted between position keeper and angle solver. This new attachment worked as a master switch between all of the submarine’s sensors. It also simultaneously indicated all of the sensor readings, available at any instant from radar, sonar and optical, permitting a cross reference check.
The Mark IV upgrade also expanded the range of torpedo tactical settings available by changing some gearing. This directly accommodated the new, slower electric torpedoes. Prior to the Mark IV upgrade the TDe Mark III had to be set up to indicate twice the speed and half the range of the true solution for these slower shots. Most of the fleet submarines still in use after the end of the war were upgraded to the Mark IV TDC. Because this was the pool from which most of the fleet submarine museums came, there are now only two unmodified TDC Mark III left installed in submarines.
USS PAMPANITO went into moth balls only two months after the end of World Warn. It remained in this state for 15 years, well after the me Mark IV upgrade program was over. As a result it never received the upgraded Mark IV me. The only other museum ship with an unmodified original Mark m me installed is USS BOWFIN on display at Pearl Harbor.
This restoration effort would have been impossible without the TDC Mark III manual available in the PAMPANITO’s library. There are only seven known copies of this ordnance pamphlet (OP 1 056). The manual for the me Mark IV (OP 1442) is even scarcer, with only two known original copies in existence. In addition, the access to other PAMPANITO volunteers like fleet submarine veteran Joe Senft, familiar with fleet submarine wiring, was invaluable.
The TDC Mark m handbook gives a detailed account of its theory, and examples of how its parts work. There is a detailed discussion of how to dismantle and reassemble a me. Along with the detailed diagrams and pictures, are the directions for checking, servicing and operating the TDC.
The first order of business was to restore Ae shore power to the me heater circuit. All TDCs have an electric heater to maintain an even temperature of 74 degrees inside the position keeper case. This prevents the buildup of moisture and maintains the mechanical tolerances required for accurate operation.
Hundreds of gears, shafts, bearings, and closely machined surfaces must match each other perfectly for the TDC to work. Every moving shaft and gear runs on finely made miniature ball bearings. The surfaces of the integrator wheels look like mirrors because of their finish. Indeed, first hand accounts of the building of these fine machines verify that most of the sub-assembly fitting was done by skilled machinist’s hands. The required fit and touch of each sub-assembly must be as soft as a baby’s behind.
After manually checking the machine’s operation, the next problem was lubricating a machine that had not seen an oil can in 30 years! We were able to obtain a copy ofOP 3000-U.S. Navy Lubrication from the library of USS COBIA in Manitowoc, Wisconsin. This document bas a table that converts the 1944 Navy lubrication numbers used in the TDC manual into the names of lubricants available today. A large number of Gier tubes feed oil by capillary action into key places inside the very close recesses of the TDC. Lubrication was introduced over a period of several months to assure that the oil had time to penetrate, by capillary action, the fairly long distances into the machinery.
The single largest challenge to the restoration of the TDC was providing electrical power. Connecting AC power to the beating circuit is simple compared to starting the machine up. The TDC uses two power sources. One source is DC 115 volt at 10 amps required to run the time motor in the position keeper section. The angle solver section must also have single phase 115 volt AC 60 cycle power for the follow-up heads that make up the feedback loops.
Restoring power required that someone understand the wiring of PAMPANITO’s IC switchboard. Over the years much of PAMPANITO’s wiring has been modified. There are few wiring diagrams, and no way to know what the original intent of the builder was. Much of the restoration time was spent wedged behind PAMPANITO’S IC switchboard tracing wires and checking continuity. Fortunately, PAMPANITO’s cabling systems have well-preserved circuit number tags which speeded up the task. Slowly, an IC switchboard wiring diagram was developed.
Power for the TDC time motor on PAMP ANITO could come from three separate sources, and one of those sources was an AC to DC selenium rectifier stack. Although age had long ago caused the selenium crystals to break down, it was possible for Joe Senft to replace them with a solid state device that easily fit into empty space in the power supply cabinet. After considerable testing of the remaining wires, and some repair to the original circuits, we were able to provide both AC and DC power to the TDC for the first time in 40 years.
Operating the TDC
After carefully testing the mechanical travel of the TDC, and years of input crank fiddling by the well-meaning curious, the machine was well out of alignment. The TDC is a classic example of two electromechanical feedback circuits connected to each other. As the position keeper computes the current position of own ship and the target, the results are forwarded to the angle solver as rotating shafts. The angle solver in tum computes the gyro angle and a projected pseudo run for the torpedo to hit the target. The results of the calculated torpedo’s run are fed back to the position keeper as a new input. In this way the TDC iterates the solution of two differential equations with two unknowns.
Once DC power was applied to the time circuit the time motor started to compute the progress of an imaginary target represented by the current settings of the hand cranks. Adding AC power caused the machine to start computing the total solution. Because most of the mechanism was out of alignment many of the dials started to rapidly tum in every direction at once. In a few seconds the dials started to slow down, and in a few seconds more they started to seek equilibrium.
Once the machine settled into a steady state the generating light came on and the machine began to track a solution. This was quite remarkable after so many years of inactivity! In order to test the accuracy of the TDC, we upset the most extreme test problem available in the manual. This is where the target and submarine are approaching each other at high speed. We shut down the machine and set the initial measurements into the hand cranks.
Upon starting up the computer with these extreme initial conditions loaded the TDC did remarkably well. Most of the variables change at a high rate of speed as the target and subma-rine pass each other. It is fascinating to watch the machine compute continuous solutions to simultaneous differential equations that have rapidly changing variables. The TDC kept up with the problem’s rates, and produced a result that was acceptably close to the required answer. It is most amazing when one realizes that this machine is mostly wheels, gears, and shafts, and pre-dates the invention of the digital computer.
What is Next?
The project on USS PAMPANITO is far from over. We plan to complete the restoration of the balance of the fire control system. This includes rebuilding the gyro angle indicating and setting regulators (GJSR-also known as Mickey Mouse because of how it looks) in each of the torpedo rooms that act as output devices for the TDC. These devices receive the electrical gyro angle order generated by the TDC. The machine converts the order into rotation of a jack shaft. This shaft is geared to axles that run up the inside of the torpedo tubes and turn the torpedo gyro angle setting spindles. The GISR does this with a 1 HP motor that uses 40 amps of 110 volts DC.
In order to operate the GISR we will have to build new power supplies for PAMPANITO that replace the missing battery. In addition, tracing the wires for the much longer runs between the conning tower and the torpedo rooms will present a challenge. There are junction boxes in each compartment for both the DC power and the computer generated signal. All of these connections must be identified and tested before connecting power.
We are also developing a museum display of the World War II fleet submarine fire control system. There are 10 interested museum locations around the country that have vintage torpedoes on display with no explanation of how they were targeted. We hope to cooperatively develop a display explaining this remarkable system to the general public. Only then will we have accomplished the mission of illustrating this historic machine and its effect on history.
Finally, we are attempting to develop a book on this subject. Computational mechanical analog computers had a very short history. They were only prevalent for the 50 years between the turn of the century and the invention of the digital computer at the end of World War II. These devices played a significant role in most of the historical events of the period. The fact that they were built changed the rules. By understanding these devices we can start to see how the ability to compute with a machine fueled the desire for even more machines with greater abilities.