Frank Dickey
August 31, 2004
Displaced Phased Center Antenna (DPCA) Radar
1953
The following was excerpted from the paper "Development of Airborne Moving Target Radar for Long Range Surveillance," published in "IEEE Transactions on Aerospace and Electronic Systems," Vol. 27, No. 6 November 1991. Images were removed to conserve web space.
INVENTION OF DISPLACED PHASE CENTER ANTENNA
One of the authors, Frank Dickey, was the principal investigator for GE on the Signal Corps contract mentioned above. One of the topics studied was the use of a monopulse antenna in an MTI radar to remove the degradation of the MTI caused by rapid scanning of the antenna. He noted the similarity of this problem to that of the butterfly effect in AMTI, this problem being caused by rotation and the butterfly effect by translation of the antenna. As a result he proposed a method of using a monopulse antenna for motion compensation in airborne MTI and reported this to the Signal Corps.3 The term “displaced phase center antenna” was coined later.
Proposal to Wright Air Development Center
With the approval of the Signal Corps, the idea for AMTI motion compensation was taken by GE to Wright Field. In discussions between GE and Wright Field in early 1953, there was agreement that; (1) platform motion was a severe problem and was made worse by the increasing speeds of tactical jet aircraft and (2) the solution proposed by GE was attractive but a complete analysis was needed.
Within a few months a performance analysis and a preliminary design for the motion compensated AMTI were made at GE. The goal set by the Air Force was elimination of the butterfly effect in the APS-27 AMTI radar at an aircraft speed of 500 knots. The performance analysis showed that this could be achieved. Fig. 3 shows a graph of some of the results. Here computed values of clutter cancellation ratio as limited by the butterfly effect are plotted against the pulse-to-pulse movement of the aircraft normal to the beam, divided by the antenna width. An arrow on the horizontal axis shows the value of this ratio for the critical case of an aircraft speed of 500 knots and assuming APS-27 parameters which for the antenna width are 60 inches and for the repetition rate, 2000. At this point the graphs show an improvement from 12 dB without compensation to 38 dB with compensation. Some unclassified portions of the analysis were published4 at the time. Subsequently, a formal proposal was made to Wright Field, and in 1954, GE was authorized to start work.
APS-71 Research Model
The contract called for GE to incorporate DPCA into an APS-27 X-band radar furnished by Wright Field. The radar included non-coherent MTI which, as explained earlier, meant that ground clutter had to be present to act as a doppler reference before moving targets could be detected. The non-coherent feature was not considered objectionable since it was expected that the radar usually would be used over terrain where ground clutter was present out to the maximum range of interest. The radar as modified for DPCA was to be called the AN/APS-71. George Kirkpatrick’s radar section was well equipped to do this job since they had just completed another program for Wright Field in which an APS-23 was modified for monopulse. Donald H. Kuhn was made Project Engineer and he guided the program until its completion.
A block diagram of the unique portions of the radar after modification for DPCA is shown in Fig. 4. The APS-23 antenna was a folded pillbox about 60 inches wide which gave an azimuth beamwidth of about 1.5 degrees at 9375 MHz. The single feed on this antenna was replaced by a dual feed with a waveguide hybrid combiner to provide sum and difference outputs. These changes required also the use of dual azimuth and elevation rotating joints, dual mixers and dual preamplifiers.
The sum channel preamplifier delivers a signal that is very nearly identical to that which would exist in the APS-27. In the APS-27, this signal would be applied to a lin-log receiver and delay line canceler to provide MTI video, or simply to a rectifier to provide normal video. In the APS-71 radar, controlled amounts of difference signal are properly phased and added to the sum signal to provide platform motion compensation.
At the IF level, the difference signal also is made available with an added 90 degree phase shift. This condition is proper for scan compensation. That is, it causes the pointing angle rather than the phase center to shift in proportion to the amount of difference signal added to the sum signal.
Forward phase center displacements are made for signals that are scheduled to be used with a time delay and backward displacements are made for those to be used without delay. The amount of displacement equals the distance traveled by the aircraft (horizontally and normal to the beam) during the interpulse time. Thus, effectively, the two signals that are combined have been received from the same point relative to the earth. Actually, phase center displacements are made on reception but not on transmission. However, this fact can be taken into account easily by making the “effective” radar location in space the midpoint between transmission and reception. This simply means that the total required phase center displacement is not equal to but is twice the aircraft travel distance.
To compensate in a similar manner for the antenna rotation rate, small angular displacements are made. These are equal and opposite to the angular movement of the beam due to scanning so that, effectively, the beam pointing angle is the same for each of two signals that are combined.
Modulators which act as analog multipliers, are used to control the scan compensation and the velocity compensation. The control voltage for the velocity modulator varies with ground speed and with the sine of the instantaneous beam azimuth relative to the ground track. The sine is obtained from a potentiometer servoed to a selsyn transmitter on the antenna pedestal. The ground speed and the drift angle were set in by the radar operator or, if he lacked that information, he could adjust his dials to minimize ground clutter. The control voltage for scan compensation is derived from a tachometer on the azimuth drive motor and thus is proportional to the scan rate for either continuous or sector scan.
Both of the control voltages alternate, plus or minus, as governed by a square wave generator. Proper compensation is obtained only on every other cycle, so the square wave also is used to blank the display on every other cycle. This scheme of reversing the correction on each cycle was adopted merely to simplify the delay line canceler. Later designs provided proper correction during each output radar cycle.
The APS-27 radar was set up on the roof of a building at Electronics Park in Syracuse for initial testing. A simulator for testing the displaced phase center principle was installed on the roof of another building about 500 feet away. It consisted of three small horns aimed at the radar and spaced apart such that, when the middle horn was at the center of the radar beam, the other two would be near the beam edges. The middle horn was fed by a signal generator while the outer horns received the same signal after it was phase shifted. Two rotary phase shifters driven by a variable speed motor were used. The signal for one outer horn was shifted forward and the other was shifted backward. The continuous phase changes produced a rough simulation of the differential doppler shift between ground clutter in different portions of the beam.
The action of the simulator also can be viewed as the generation of a multi-lobed pattern by the three horns, with the pattern sweeping past the radar antenna at twice the speed of the aircraft. This viewpoint also is useful as an alternative way of describing the theory of DPCA. The clutter fluctuation observed in airborne radar and usually ascribed to differential doppler shift can be viewed equally as well as being the diffraction pattern of the ground at any given range. The pattern moves backward as the aircraft moves forward so that it moves across the receiving antenna at exactly twice the ground speed of the aircraft.
The simulator was conceived and designed by Donald Kuhn. It was valuable since it permitted adjustments to be made at the roof site, as well as demonstrations and tests of performance as a function of simulated aircraft speed. The radar and the roof top tests were described by Kuhn5 at the Michigan Radar Symposium of February, 1955.
Flight Tests
When plans for flight testing were made it became apparent that a high speed aircraft would not be available. The aircraft that was available was a B-50, similar to a B-29, which cruised at about 225 knots. To simulate the butterfly effect at 500 knots, the PRF of 2000 was reduced to 900 by using a 1111 microsecond delay line. Two indicators were used and circuits were arranged for displaying motion compensated MTI on one indicator and for displaying normal MTI on the other. Provision was made for taking simultaneous photographs of each indicator to compare the compensated to the uncompensated case. The display rate was reduced to one third the PRF or 300. This was necessary in order to provide simultaneous comparison. Provision also was made for gating the compensated and the normal MTI video at any desired range. The resulting samples, after stretching, could be recorded on a dual channel magnetic tape recorder.
During a two month period in 1956 the radar was installed in the B-50 aircraft at Griffiss AFB in Rome, NY and more than twenty hours of flight tests were logged, flying first out of Rome and then out of Dayton. The results were consistent with theoretical predictions. It was found that, under the test conditions, the butterfly effect was reduced by the use of DPCA to the point that it was not troublesome, although it was not completely eliminated.
Fig. 5 is an example of photographs taken during the tests. Here the aircraft is north of Detroit, at about 8000 feet, and is heading south. On the lower right is a portion of a road map centered on the area displayed in the scope photographs. The scope photographs have all been printed with the aircraft heading at the top and the road map has been rotated to correspond. In each picture, the city of Detroit is ahead and slightly to the right of center. There also is a large lake to the left. The radar displays a 30 mile range. A few returns, especially over the lake, were caused by interference from another radar.
The lower left picture shows normal video or non-MTI. The city of Detroit produces solid returns for at least 10 miles ahead and to the right. The upper left picture shows AMTI without motion compensation. Here the solid return in the forward region is replaced to a large extent by a pattern of returns which is consistent with the assumption that most of the returns in this region are from moving vehicles on the major highways. The upper right picture shows AMTI with motion compensation by DPCA. Here the moving target pattern extends to the right as well as ahead. Notice also that ground clutter in the lower left quadrant, which is from across the lake, is greatly diminished in this picture. Both these features are the result of improved cancellation of fixed targets in regions near 90 degrees to the ground track.
The scan compensation feature that was built into the APS-71 radar had been tested previously on the ground with good results. However, the lack of cancellation due to scanning did not show up unless the antenna was rotated much faster than usual, so this feature was not normally used.
The flight test results were presented at the Michigan radar symposium6 of February, 1957.