Ship Gun Fire-control System - US Navy Systems - MK 37 Gun Fire Control System (GFCS) - Ford Mark 1A Fire Control Computer

Ford Mark 1A Fire Control Computer

The Mark 1A Fire Control Computer was an electro-mechanical analog ballistic computer. Originally designated the Mark 1, design modifications were extensive enough to change it to "Mk. 1A". The Mark 1A appeared post World War II and may have incorporated technology developed for the Bell Labs Mark 8, Fire Control Computer. Sailors would stand around a box 62 inches long, 38 inches wide, and 45 inches high. Even though built with extensive use of an aluminum alloy framework (including thick internal mechanism support plates) and computing mechanisms mostly made of aluminum alloy, it weighed as much as a car, about 3125 lb, with the Star Shell Computer Mark 1 adding another 215 lb. It used 115 volts AC, 60 Hz, single phase, and typically a few amperes or even less. Under worst-case fault conditions, its synchros apparently could draw as much as 140 amperes, or 15,000 watts (about the same as 3 houses while using ovens). Almost all of the computer's inputs and outputs were by synchro torque transmitters and receivers.

Its function was to automatically aim the guns so that a fired projectile would collide with the target. This is the same function as the main battery’s Mk 8 Rangekeeper above except that some of the targets the Mark 1A had to deal with also moved in elevation — and much faster. For a surface target, the Secondary Battery’s Fire Control problem is the same as the Main Battery’s with the same type inputs and outputs. The major difference between the two computers is their ballistics calculations. The amount of gun elevation needed to project a 5-in shell nine nautical miles (17 km) is very different from the elevation needed to project a 16-in shell the same distance.

In operation, this computer received target range, bearing, and elevation from the gun director. As long as the director was on target, clutches in the computer were closed, and movement of the gun director (along with changes in range) made the computer converge its internal values of target motion to values matching those of the target. While converging, the computer fed aided-tracking ("generated") range, bearing, and elevation to the gun director. If the target remained on a straight-line course at a constant speed (and in the case of aircraft, constant rate of change of altitude ("rate of climb"), the predictions became accurate and, with further computation, gave correct values for the gun lead angles and fuze setting.

Concisely, the target's movement was a vector, and if that didn't change, the generated range, bearing, and elevation were accurate for up to 30 seconds. Once the target's motion vector became stable, the computer operators told the gun director officer ("Solution Plot!"), who usually gave the command to commence firing. Unfortunately, this process of inferring the target motion vector required a few seconds, typically, which might take too long.

The process of determining the target's motion vector was done primarily with an accurate constant-speed motor, disk-ball-roller integrators, nonlinear cams, mechanical resolvers, and differentials. Four special coordinate converters, each with a mechanism in part like that of a traditional computer mouse, converted the received corrections into target motion vector values. The Mk. 1 computer attempted to do the coordinate conversion (in part) with a rectangular-to polar converter, but that didn't work as well as desired (sometimes trying to make target speed negative!). Part of the design changes that defined the Mk. 1A were a re-thinking of how to best use these special coordinate converters; the coordinate converter ("vector solver") was eliminated.

The Stable Element, which in contemporary terminology would be called a vertical gyro, stabilized the sights in the director, and provided data to compute stabilizing corrections to the gun orders. Gun lead angles meant that gun-stabilizing commands differed from those needed to keep the director's sights stable. Ideal computation of gun stabilizing angles required an impractical number of terms in the mathematical expression, so the computation was approximate.

To compute lead angles and time fuze setting, the target motion vector's components as well as its range and altitude, wind direction and speed, and own ship's motion combined to predict the target's location when the shell reached it. This computation was done primarily with mechanical resolvers ("component solvers"), multipliers, and differentials, but also with one of four three-dimensional cams.

Based on the predictions, the other three of the three-dimensional cams provided data on ballistics of the gun and ammunition that the computer was designed for; it could not be used for a different size or type of gun except by rebuilding that could take weeks.

Servos in the computer boosted torque accurately to minimize loading on the outputs of computing mechanisms, thereby reducing errors, and also positioned the large synchros that transmitted gun orders (bearing and elevation, sight lead angles, and time fuze setting).These were electromechanical "bang-bang", yet had excellent performance.

The anti-aircraft fire control problem was more complicated because it had the additional requirement of tracking the target in elevation and making target predictions in three dimensions. The outputs of the Mk 1A were the same (gun bearing and elevation), except fuze time was added. The fuze time was needed because the ideal of directly hitting the fast moving aircraft with the projectile was impractical. With fuze time set into the shell, it was hoped that it would explode near enough to the target to destroy it with the shock wave and shrapnel. Towards the end of World War II, the invention of the VT proximity fuze eliminated the need to use the fuze time calculation and its possible error. This greatly increased the odds of destroying an air target. Digital fire control computers were not introduced into service until the mid-1970s.

Central aiming from a gun director has a minor complication in that the guns are often far enough away from the director to require parallax correction so they aim correctly. In the Mk. 37 GFCS, the Mk1 / 1A sent parallax data to all gun mounts; each mount had its own scale factor (and "polarity") set inside the train (bearing) power drive (servo) receiver-regulator (controller).

Twice in its history, internal scale factors were changed, presumably by changing gear ratios. Target speed had a hard upper limit, set by a mechanical stop. It was originally 300 knots, and subsequently doubled in each rebuild.

These computers were built by Ford Instrument Company, Long Island City, Queens, New York. The company was named after Hannibal C. Ford, a genius designer, and principal in the company. Special machine tools machined face cam grooves and accurately duplicated 3-D ballistic cams.

Generally speaking, these computers were very well designed and built, very rugged, and almost trouble-free, frequent tests included entering values via the handcranks and reading results on the dials, with the time motor stopped. These were static tests. Dynamic tests were done similarly, but used gentle manual acceleration of the "time line" (integrators) to prevent possible slippage errors when the time motor was switched on; the time motor was switched off before the run was complete, and the computer was allowed to coast down. Easy manual cranking of the time line brought the dynamic test to its desired end point, when dials were read.

As was typical of such computers, flipping a lever on the handcrank's support casting enabled automatic reception of data and disengaged the handcrank gear. Flipped the other way, the gear engaged, and power was cut to the receiver's servo motor.

The mechanisms (including servos) in this computer are described superbly, with many excellent illustrations, in the Navy publication OP 1140.

There are photographs of the computer's interior in the National Archives; some are on Web pages, and some of those have been rotated a quarter turn.