A brief history of the NTDS.
To say the Navy was an early adopter of radar would be an understatement. As of 1940, few if any US Navy ships had radar.
That’s a Benson class destroyer. That was the main model in production at the outbreak of the war. And yes, that’s a crows nest up top on the mast.
By the end of the war, this Fletcher class destroyer had an air search radar, a surface search radar, gunfire control radar on her main director, radar intercept receivers, and a jamming system, as well as at least one navigational homing beacon.
Virtually every combatant ship by 1943 had multiple radar systems, and most auxiliaries did as well. Airplanes had surface search and air intercept sets, submarines had search radars, and even periscope mounted ranging sets. Land based units of course had radars as well. Radar was everywhere the Navy went. Not an inch of the surface they sailed wasn’t assailed by trillions of electrons flung out in search of the foe. And today’s reader may be surprised to learn that those sets were very effective. Their range and sensitivity over open ocean would compare quite well even with todays radar systems. Oh, there have been improvements, to be sure. Greater power, greater reliability, and such. But the range of radar hasn’t really improved much since World War II. Physics sets those rules. Radar, for the most part, is a line of sight device. And since the earth is curved, a radar on at surface level has in inherent limit to its range. Toward the end of World War II, the Navy’s prime target was enemy aircraft, particularly the Kamikaze. The range at which a destroyer’s radar detected a Kamikaze was largely a function of the altitude a Kamikaze was flying at. Few Kamikaze raids at Okinawa went undetected. The relatively short detection range of a surface mounted radar against low flying raids led the Navy to station numbers of destroyers as early warning pickets further along the most likely avenues of enemy approach to give more warning time. The Navy was also developing airborne early warning radar systems to better detect low flying aircraft at longer ranges.
As valuable as radar was, it wasn’t without its limitations. Radar would tell a ship where an airplane was, in terms of range and bearing. But that information was of little use without knowing where the target was heading, and how fast. By manually plotting range, bearing and time information and using a little trigonometry and a maneuvering board, the radar operator could determine the targets course, speed, and best course for friendly fighters to intercept. Against one or two piston powered airplanes, that was sufficient. But plotting took significant time, and more than one or two targets overwhelmed the manpower available .
One way to avoiding this task saturation was to assign each escort a limited sector to plot, so more fighter control teams were available to plot more interceptions. A master plot of the total air picture was maintained on the force flagship on a Plexiglas board. The pickets would call in information via voice radio. Don’t forget, every bit as important as tracking and plotting enemy air was tracking and plotting friendly aircraft. Only by continuously tracking every airborne contact could a task force be reasonably certain that every enemy raid would be intercepted or taken under anti-aircraft fire by escorts. Further, by plotting the location of friendly combat air patrols, fighter direction officers would know the best CAP to task to intercept a given raid.
But the Navy realized that with increased speeds of the jet age, the teams manually plotting air tracks would quickly become overwhelmed. The engagement chain, from detection, tracking, handing off to fire control radars, and developing a firing solution took time, time that the higher speeds of jets just didn’t allow.
Almost immediately after the end of WWII, a handful of radar specialists in the Navy began to search for a solution to the problem. They quickly realized any mechanical/analog computer system similar to those used in gunfire director/computer systems would be overwhelmingly complex. But a few had heard of the first forays into the electronic (digital) computer systems entering service such as the ENIAC. The plotting and tracking functions of any system of automation would largely be relatively simple mathematic calculations. The challenge wasn’t the complexity of the math, but the volume of it. And the ENIAC and its brethren were designed for the sole purpose of performing large numbers of mathematical computations. The problem was, the ENIAC was as big as a house. The Navy’s codebreakers were also deeply interested in digital computers. Codebreaking is again an arena where the mathematical computations themselves aren’t terribly complex, but the sheer volume of calculations needed overwhelm both humans, and the primitive electromechanical devices used in World War II.
Working hand in hand with universities and private industry as well as pioneers in digital computing such as Seymour Cray, the Navy, over the course of several years (with Moore’s Law in effect) managed to procure digital computers that were little larger than a refrigerator. Massive effort also went into developing modulator/demodulator (modem) systems that would convert the analog input from radars and other sensors into the digital information a computer could process. Another massive hurdle was to develop displays for the operators. It wasn’t enough to simply show the raw radar picture. The system also had to generate symbology showing which targets were being tracked, and display the course/speed information the computer had divined on them (known as a “vector”). It also had to generate and display the interception vector computed for a friendly aircraft to intercept any hostile tracks. It quickly became apparent that even with computer support, one operator could only manage one or two interceptions concurrently. Accordingly, the system would have to provide multiple workstations for several operators, and accept input and generate output for each one separately. Finally, since the CAP was unlikely to defeat all possible raids, the system had to be able to cue and point fire control radars, gun and missile systems for close in engagements. Since these systems were analog, the computer had to go through another modem to convert its information back to an analog format that they could accept. Basically, the digital information would be converted to a DC electrical current that would vary, and drive a servomotor that energized the traverse and elevation drive motors of a fire control radar or gun mount. Once the track had been handed off, the fire control system would engage in its traditional (analog) manner, though the system still had to maintain its track information on the target until destruction.
Another major hurdle was sharing information. Even if a ship had sufficient computer power to quickly plot aerial targets, and provide vectors to intercept or cueing to fire control systems, unless that information could be shared across an entire task force (or at a minimum, across the with the fleet flagship and the primary anti-air platforms), there was still a very good chance that intruders could slip through. For instance, while each anti-air escort would still be responsible for its own “slice of the pie” sector, attacking aircraft were often inconsiderate enough to cross from one sector to another. Handing off the track from one escort to another in the age of paper plots was time consuming, and without an automated system, the chance of overlooking a handoff was quite high. In fleet air defense exercises in the early 1950s, as many as a quarter of all contacts were not tracked. And of the contacts tracked, as many as a quarter of those contacts were never assigned an interceptor or handed off to onboard weapons for engagement. Giving the enemy a free shot for almost half of its attacking force was simply unacceptable. Again, without an automated system to monitor all the contacts, the task force air defense could quickly be overwhelmed.
But how to share that plot information across multiple ships? Digital computing itself was in just the barest infancy. The idea of networking computers hadn’t even been considered. Data had to be converted from raw radar return video to a digital format suitable for display on the operator’s scope. In addition, to be shared with other ships in the task force, accurate data regarding the source ships position, course and speed had to be injected. First, since contacts were referenced by their position relative to the reporting ship, to figure the contact’s true position, the receiving ship had to know the reporting ship’s position. Secondly, if two ships were both tracking the same target, knowing that positional data allowed the computer system to filter out duplicate tracks of the same target.
Given the wide dispersal of task force ships, long range High Frequency radio (HF) had to be used. That choice also meant a very low bandwidth for data. Smart communications people came up with an ingenious method to transmit information among NTDS equipped ships. A modified Radio Tele-Type would transmit (and receive, of course) information. Using a master/slave timesharing system (a simplex system, as opposed to a multiplex to you commo types) the master would query each NTDS ship in turn to update the shared air warfare picture. Each ship in the network would listen in to each transmission. Having done so, all ships would have the same picture of all tracks in sensor range of the formation, even from distant radar pickets up to 400 miles away.
NTDS would grow in complexity as the capacity of computing power grew. Surface search radar and anti-submarine sonar inputs would be added, expanding NTDS from simply an air warfare tool, to a battle management system. Interfaces between NTDS and airborne early warning aircraft vastly extended the reach of sensors. Tying in the NTDS with similar shorebased systems for the Marines (the MTDS- Marine Tactical Data System) extended coverage over the shore). Eventually the existing analog fire control systems would give way to digital systems, and be tied into the NTDS system.
The NTDS system was expensive, and required considerable volume and personnel in a ship. Clearly it was only suitable for larger ships, generally those larger than a destroyer. But in order to share the common sensor picture, a receive only system was developed to share information with smaller ships such as destroyers and frigates.1
NTDS quickly became the primary information superhighway of the Navy from the early 1960s through the mid-1990s. Any time the Navy needed to keep a clear picture of the air (and surface and subsurface) situation, off Vietnam, Lebanon, or off the shores of Iraq during Desert Storm, NTDS equipped ships were there.
Not until the boom in the use of smaller computers of the 1980s did the NTDS start to face obsolescence. Development of successor systems began in the mid 1980s, and by the mid 1990s, a variety of cheaper, faster, smaller and lighter systems began to replace it. Today several high bandwidth data links working in conjunction with advanced versions of the computer systems used in NTDS perform the battle management function.
In many ways, NTDS can be considered the earliest internet. Further, from a program management point of view, NTDS was also a first. Whereas the Air Force SAGE system used a computer that took up almost a half an acre, the NTDS performed virtually the same functions with between seven and 11 computers each roughly the size of a refrigerator. The NTDS was also the first program for which there WAS a program office. Rather than receiving input from the various bureaus and commands and cobbling together their consensus product, the program office received the input on desired functionality, but then held sole responsibility for defining, designing, procuring and supporting the product, with its own budget.
1. Known at that time as destroyer escorts, or ocean escorts. At the time of NTDS development, a “frigate” was a warship larger than a destroyer, but smaller than a cruiser.
Bibliographical note- I researched any number of sources on NTDS prior to beginning this post, but came across the motherlode (and in effect, sole primary source for this post) at IEEE with this excellent 9 part series of the development of NTDS. If you’re at all interested in the subject, or program management, or simply history, I highly recommend you read the whole thing.