Aerofax: Lockheed Martin F/A-22 Raptor, by Jay Miller, Midland Publishing 2005, ISBN 185780158X
YF-22 and YF-23 Advanced Tactical Fighters: Stealth, Speed and Agility for Air Superiority, by Bill Sweetman, Motorbooks International 1991, ISBN 0879385057

ATF Programme History
The specification for the ATF radar evolved gradually, taking advantage of progression in technology, especially in relation to computer processing power and miniaturisation of electronic circuitry. The ATF radar descended from the Ultra-Reliable Radar (URR) project. A core requirement that surfaced as a result was that the radar antenna be a completely integral part of the sensor suite rather than a separately functioning unit; it was to be linked to the aircraft's computer system in such a way that it could be used for tasks other than traditional radar scanning and detection. It was to be used for ECM, ECCM, and have the ability to hand off tasks to other sensors and aerials in a reciprocal fashion.

The ATF would not have a completely separate dedicated radar like the F-15 does. Instead, the antenna was to be regarded as one of a number of sensors in conjunction with the electronic warfare and infrared search and track systems. Signals from these sensors would be processed by a number of JIAWG compliant common signal processors, which would extract target returns and provide that data to the main mission processor, the cockpit display processor, and the rest of the avionics system.

An important step forward was that information from the radar sensor would be fused by the computer system with information coming in from other sensors, to present a single symbol on the pilot's cockpit display in the case of a single target being detected. The pilot would no longer have to expend concentration effort trying to build a tactical picture in his head from the different gauges and displays in the cockpit. This concept of sensor fusion was first fielded successfully on the F-117A Nighthawk, which had 2 laser designation systems that needed to be synchronised.

Another core requirement was that the radar system should have a far higher level of reliability and lower maintenance costs compared to previous generations of radar. The radar concept that emerged did away with mechanical movement. It was based on the concept of Active Array, which employs many thousands of miniature transceivers arranged in a matrix and all electronically linked together. Instead of the scanner dish turning mechanically to provide coverage, the array is mounted flat and is aimed electronically, which means that the scanning ability becomes extremely rapid. The radar beam is steered by adjusting the wave phase of the transceivers from one edge of the array to the other. This method allows the radar to work in several modes or frequencies at once, flipping between each in milliseconds. This facilitates near real time scanning for targets, tracking of detected targets, terrain mapping, and frequency hopping to avoid hostile ECM. Because the radar is an array of miniature transceivers, failure of one or two of these small devices does not seriously hamper performance. The radar has more precise scanning ability due to the greatly diminished distance between transmitter and receiver aerials, and because there are so many little aerials on the one unit. Incoming signals have a much higher chance of being picked up, and the computer processing capacity allows for a much finer resolution of targets, which assists in identification of other aircraft even if their IFF systems are not broadcasting. An active array can concentrate a radar beam more intensly to counteract hostile ECM, by placing so much energy on a target that a real return overwhelms the fake signal. If the array encounters continuous wave noise jamming, a few transceivers can be employed to transmit a half-phase-out signal back, to cancel out the original signal.

In concert with the concept of stealth, the radar would have Low Probability of Intercept (LPI) characteristics. The objective of LPI is to prevent signals from being intercepted by hostile surveilance systems, and ensure that any signals inadvertently picked up are dismissed as background noise. Methods employed to facilitate this include using the absolute minimum amount of power required, highly agile frequency hopping, and forming several different beams simultaneously in one burst of power to perform different tasks in the minimum amount of time required.

Texas Instruments and Westinghouse both bid for the ATF programme radar contract. The arrangement was that they would team together, with either company acting as lead for each airframe team: in the case of the Northrop McDonnell Douglas YF-23 effort, Texas Instruments would lead. So, the contract for the solid state array for the 'ATF-23 Team' was awarded to Texas Instruments in April 1983. The MTBF goal was 400 hours, to demonstate that it could operate in an aircraft from an austere base for 1 month, flying six 2.5 hour sorties per day, with a 90% probability of still being able to perform after the end of that month. The prototype array modules used on the Dem/Val ATF radars cost about $12,000 each, and it was anticipated that they would work for an average of 2,500 hours between failure. There were plans to reduce the cost to $400 each on production modules. The prototype radar was tested in 1988.

Dem/Val Phase
It was eventually announced by the AF SPO that there would be a Demonstration Validation phase to the ATF programme, and as part of that phase the competing radars would be evaluated in a distinct aspect of Dem/Val, called Avionics Ground Prototypes (AGP). AGP was a demonstration of the complete range of avionics, not just the radar. Its purpose was to demonstrate practicality of fully integrated avionics suites, real-time fusion of multi-sensor data, a computer core processing speed 100 times faster than previous generation fighters, self-diagnostics and fault isolation, and system reconfiguration. This would be done intially in ground tests, then in the air. AGP was a test of the entire suite of sensors, transmitters, processors, and cockpit displays for the ATF. An unprecedented requirement was putting together a complete prototype of the avionics system and demonstrating it while it was being used by a pilot.

The avionics suite was to be tested completely separate from the PAVs, in a dedicated flying lab. By the time the ATF teams were ready to install radars in their AGP labs, Texas Instruments and Westinghouse were working on next-generation versions of the radar. Northrop used a modified BAC-111 (N162W) for the AGP flight tests, commencing 17 July 1989. Approx 100 flight hours were achieved in 1990. These tests validated sensor data fusion: the ability to detect a target using multiple sensors and show it as one symbol in the pilot's cockpit reliably and consistently. In addition to the radar, Northrop demonstrated all-aspect threat missile launch detection and tracking capability, and an IRST system.

The picture immediately below is a rarely depicted picture of the prototype version of the APG-77 being installed, for some unknown reason, upside down. Notice that, upside down, it has a shape that very closely conforms to the cross section of the nose of an F-23A. This is about as close as we can get, guys...