|
FOIA DOCUMENTS on the COBRA MIST (AN/FPS-95) PART 2 OF 4 PARTS
|
|
{SECRET NOFORN} UNCLASSIFIED TABLE 4. Receiver/Signal processor parameters. (Table unclassified.) ---------------------------------------------------------------------------
--------------------------------------------------------------------------- verter. These data were then available for analysis off line by the extensive programs that were specially developed as part of the DVST activity or for replay through the on-line system. Some of the main receiver and signal processor parameters are listed in Table 4. DISPLAYS (U) The signal processor outputs contained data on target range, azimuth, velocity, acceleration, and signal amplitude. These parameters, together with a time-history dimension, could be shown on a number of cathode ray tube (CRT) displays. Intensity modulation was not employed on these displays, with the result that only two of the foregoing six parameters were displayable in the chosen x-y format at any one time. Some of the remaining parameters could be thresholded by manual selection to restrict the number of displayed data. From among all the possible combinations of the six parameters taken two at a time, the AN/FPS-95 had the capability of displaying 14 such pairs. (U) On those displays where the signal amplitude was not one of the exhibited parameters, an amplitude threshold had to be chosen. Thus, only those signals that exceeded this threshold could be "detected" and displayed, as in a classical radar signal detection process. Cursors were provided to allow readout of parameter values from the displays, and cameras were available for a permanent display record. In addition to the presentations on the cathode ray tubes, certain data could be recorded on magnetic tape or automatically typed. Figure 6 shows a view of the radar control console with its associated displays. Figure 6. Radar control console (Courtesy RCA Corp.) (Figure unclassified) ANCILLARY EQUIPMENT (U) To support the AN/FPS-95 operation in the selection of radar operating frequencies, the site contained a vertical ionospheric sounder and a panoramic radio receiver. RADAR CAPABILITIES AND LIMITATIONS EXPECTED CAPABILITIES (U) {(S)} The AN/FPS-95 was expected to detect and track (a) aircraft in flight over the westerly part of the Soviet Union and the Warsaw Pact countries and (b) missile launches from the Northern Fleet Missile Test Center at Plesetsk. Aircraft detection and tracking at ranges of 500 to 2,000 nmi, corresponding to one-hop ionospheric propagation, were considered feasible. Missile launches from Plesetsk were also within one-hop range from the radar. A searchlight mode was provided for high-priority targets whose approximate locations were known a priori. These targets could be single aircraft, compact formations of aircraft, or missile launches. In this mode, the radar continuously illuminated a small geographical area to obtain the maximum data rate on the selected targets. As an alternative, a scanning mode was provided, which allowed the radar to search in azimuth and range over any chosen sector of the radar coverage {SECRET NOFORN} UNCLASSIFIED JDR 297 ============================================================= {SECRET NORFORN} UNCLASSIFIED at a reduced data rate for any particular area. Scanning in azimuth was implemented by switching the antenna among the 13 discrete beam positions. Range-scanning was implemented by two methods: (a) switching between the lower and upper elevation beams and (b) varying the transmitted frequency as required by ionospheric propagation conditions to reach the desired range. Frequency selection was facilitated by an oblique sounding mode of the radar and by a separate vertical sounder. The scanning mode was intended for detection of targets whose locations were unknown a priori and for time-sharing the radar among different missions. (U) {(S)} Both accuracy and resolution of the AN/FPS-95 were expected to be considerably lower than for a typical Picowave search radar. The azimuth beamwidth of 7 deg determined the angular resolution. Monopulse beam-splitting provided a nominal 1-deg angular accuracy, but this was further limited by ionospheric tilts, which could amount to several degrees, particularly in the northern beams that approached the auroral region. The unmodulated radar pulses provided only coarse range resolution. With the longest pulse, 3 milliseconds in duration, the range resolution equaled 240 nmi. The shortest pulse, 250 microseconds., provided a nominal 20-nmi range resolution, but this pulse could be used only under a limited number of conditions. Pulse-splitting in range could be performed manually on the displays to obtain a slant range accuracy of perhaps one-third of the range resolution, but the accuracy with which slant range could be converted to ground range was limited by uncertainties in virtual height of the ionosphere. (U) Against this coarse spatial accuracy and resolution was set the fine Doppler resolution of the radar. Coherent integration times of 10 sec were frequently allowed by the ionospheric propagation medium, providing a range-rate resolution of 1.5 knots at the midband frequency of 20 MHz. The fine Doppler resolution was intended both for separation of multiple targets and for discriminating moving target returns from stationary ground backscatter. (U) The expected capability of the AN/FPS-95 were based primarily upon experience with the Madre OTH-B radar constructed in Maryland by the Naval Research Laboratory. Aircraft detection and tracking over the North Atlantic was reported by Madre experimenters at one-hop ranges on several occasions. (2) Missiles launched from Cape Canaveral were also said to have been detected. (3) Since the AN/FPS-95, like Madre, was a monostatic pulse Doppler HF radar with high transmitter power, coarse spatial resolution, and fine Doppler resolution, its detection and tracking performance was expected to be roughly similar. The siting of the AN/FPS-95 provided two operational differences between it and Madre, however: (a) Interference in the HF band was worse in Europe, particularly at night, and (b) ground backscatter was usually received by the AN/FPS-95, rather than sea backscatter. (U) {(S)} Performance of the AN/FPS-95 was projected (9) using the ITS-78 ionospheric propagation prediction computer program, (10) as modified by MITRE personnel for radar use. This program was designed to support long-range HF communications, and its utility for predicting OTH-B radar performance, a more demanding application, was not established. Nevertheless, it was the best tool available at the time and was therefore used. Single-dwell probability of detection for aircraft was estimated for a representative assortment of target areas, and it appeared adequate for a searchlight mode with repeated dwells in a given geographical area, but marginal for a scanning mode having a low data rate on any particular target. (U) {(S)} It was also recognized at the time that ground backscatter would be orders of magnitude larger than aircraft returns or missile skin returns. Ground backscatter cross section Xs, is given by Xs = 1/2 Xu R Tb C r sec Ai (1) Take typical values of the parameters for an example. Setting backscattering cross section per unit area Xu equal 0.02 (-17 dB), range R equal 1,350 nmi, azimuth beamwidth Tb equal 7 deg, pulse length r equal 1 msec, and angle of incidence Ai equal 8 deg, gives Xs, equal to 1.06 X 109 m2 (90.2 dBSM). Given a typical aircraft radar cross section Xt of 30 m2 (14.8 dBSM) at HF, the ratio Xt/Xs, equals -75.4 dB. In order to achieve a 10- to 15-dB signal-to-clutter ratio for high single-dwell probability of detection in a scanning 298 JDR UNCLASSIFIED {SECRET NOFORN} ============================================================= {SECRET NOFORN} UNCLASSIFIED mode, the pulse Doppler radar signal processor was required to suppress the ground backscatter by 85 to 90 dB relative to aircraft returns - that is, to provide 85 to 90 dB of subclutter visibility (SCV). Somewhat lower probability of detection, and hence less subclutter visibility, would suffice for the searchlight mode, where the radar continuously illuminated a given target. In an attempt to achieve the required subclutter visibility, great care was taken in the design of the radar transmitter to minimize spectral noise and in the receiver and signal processor to minimize intermodulations and cross modulations and to provide a large linear dynamic range. DESIGN VERIFICATION SYSTEM TESTING (DVST) (U) {(S)} Following construction of the AN/FPS-95 and its acceptance by the government, a one year research and development program was planned (4) to assess its capabilities. The 12 aircraft detection and tracking experiments assigned to MITRE during the DVST will be described briefly as a further indication of the expected capability of the radar. A number of other experiments, including all of the missile detection and tracking experiments, were assigned to Naval Research Laboratory and have been documented by that organization.(11) This paper will therefore discuss only aircraft detection and tracking, with which the authors have firsthand experience. (U) {(S)} Experiment 202, Radar Aurora, was intended to determine experimentally the effects of HF radio aurora on OTH radar design and operation. Experiment 104, Signal Detectability, and Experiment 502, Target Detection and Calibration, were to determine probability of detection, probability of false alarm, and signal-to-noise ratios of detected targets, as well as develop procedures to estimate radar cross section of the detected targets. Three experiments dealt with real-time tracking of aircraft at the radar consoles and were designed to develop and evaluate this capability: Experiment 501, Evaluation of Target Window Printout; Experiment 505, Tracking Through Azimuth Beams; and Experiment 508, Track Capability and Track Sample Rate. One experiment dealt with automatic tracking of aircraft, conducted off line on a digital computer. This was Experiment 405, Track-While-Scan Feasibility. Experiment 506, Range and Azimuth Calibration, was intended to provide an absolute spatial calibration using ground transponders. (U) {(S)} These eight experiments were intended to assess the general capabilities of the radar for aircraft detection. Four other experiments were directed toward specific intelligence objectives. Experiment 306, Vertical Velocity Estimation with Aircraft, was to exploit the fine Doppler resolution of the radar to measure vertical velocity. The Doppler difference between alternate ground reflected propagation modes was to be utilized for this purpose. Experiment 312, Intelligence from Test Range Calibration Flights, surveyed aircraft activity near Plesetsk and other missile test centers. Experiment 314, Reconnaissance Aircraft Surveillance, tracked friendly aircraft over the Baltic Sea area, providing the only source of over water aircraft tracking data. Experiment 315A, Aircraft R&D Test Intelligence, observed aircraft at the Ramenskoye and Vladimirovks Flight Test Centers. (U) {(S)} Of these 12 experiments, three were carried out and documented: Experiments 202, (12) 405, (13) and 506, (14) The rest were not completed for either of two reasons: (a) The experiment as conceived proved too ambitious for the actual capability of the radar or (b) the scientist assigned to the experiment was reassigned to efforts to improve the radar capability. OBSERVED CAPABILITIES AND LIMITATIONS (U) {(S)} Once DVST got under way at Orford Ness, it became apparent to the MITRE team (and others) on site that the actual radar capabilities were a good deal less then the expected capabilities. In the searchlight mode, aircraft detection and tracking were marginal, even when aircraft flight plans were known a priori. When the radar was carefully operated, with due regard for range and Doppler ambiguities and ionospheric propagation conditions, tracking trials on known aircraft in the searchlight mode produced tracks less than half the time. Furthermore, the tracks obtained were discontinuous, the aircraft return usually being above the noise level only near the peaks of the Faraday rotation and multipath fading cycles. Additionally, routine observations of areas of high air traffic density, such as air routes near Moscow, in the searchlight mode often produced few or no target detections at times of day when the propa- {SECRET NOFORN} UNCLASSIFIED JDR 299 ============================================================= {SECRET NORFORN} UNCLASSIFIED gation was good and the aircraft density in the illuminated area was known to be high. (U) {(S)} Capability in a scanning mode was essentially nonexistent, because of both the unexpectedly low signal-to-noise ratio on aircraft returns and the difficulty of making appropriate frequency selections for the large number of scan positions. In practice, the poor performance in the searchlight mode discouraged the experimenters from making much use of the more ambitious scanning mode. (U) {(S)} Some general observations during aircraft detection and tracking are worth recording, in light of the later discussion on the physical phenomena limiting radar performance.
(U) {(S)} As experience was gained with the AN/FPS-95, the factors limiting radar performance became clear. These were clear channel availability, a lack of radio frequencies below 6 MHZ, the two-level radar displays, and (worst of all) the limited subclutter visibility. A clear channel is a 5-kHz frequency band in which no significant interference exists; that is, the noise power spectral density approximates the atmospheric and galactic noise predictions of CCIR Report 322 (15) or the rural man-made noise model of the ITS-78 report, (10) whichever is larger. In the searchlight mode, one could take a few minutes to look carefully for a clear channel, but this was not possible for the scanning mode. During the day, clear channels were generally available at Orford Ness, particularly, for illuminating the longer one-hop ranges with radio frequencies above 20 MHz. Clear channels were less often available at night, when radio frequencies from 6 to 8 MHz were required. The large number of HF radio stations in Europe all operated in the same frequency region at night, Particularly the 41-meter and 49-meter international shortwave bands. The fact that the radar could not operate below 6 MHZ also limited performance at night, since the maximum usable frequency (MUF) for the shorter one-hop ranges was sometimes below 6 MHz. The shorter ranges could not be illuminated by the radar when this occurred. (U) {(S)} The radar displays were also a significant limitation of real-time aircraft detection and tracking. The Doppler-time and Doppler-range displays had only two intensity levels - on and off. The lack of a gray scale to show a dynamic range of signal and noise amplitudes made target detection difficult. Setting the on/off amplitude threshold appropriately for fading targets in a fluctuating noise level that varied from range bin to range bin was a challenging task. Better target detection and tracking were obtained offline with computer generated displays having an amplitude scale, some of which are illustrated later in the paper. An improved online radar display was developed (16) and brought to the AN/FPS-95 site just before the radar was shut down. While this had only a very brief trial, it did show promise of considerably improved detection and tracking capability compared to the radar displays. (U) {(S)} The most important constraint on radar performance was the limited subclutter visibility. Instead of a projected 85- to 90-dB subclutter visibility, the radar was found to have only 60- to 70-dB subclutter visibility. Target detectability was degraded by a noiselike clutter residue which 300 JDR UNCLASSIFIED {SECRET NOFORN} ============================================================= {SECRET NOFORN} UNCLASSIFIED often could be 20 dB, and in some cases even 30 dB, higher than the level of external noise received by the radar. Figure 7 is a photograph of the AN/FPS-95 Doppler-range display taken early in the DVST. (17) The range scale (horizontal) extends from 0 to 2,000 nmi, the nominal unambiguous range at a PRF of 40 Hz. The Doppler scale (vertical) extends from 3 to 20 Hz, with approaching and receding Doppler shifts folded together. A Doppler shift of 20 Hz corresponds to a radial velocity of 264 knots at the radio frequency of 22.1 MHz employed to obtain these data. Ground backscatter in the 0- to 3-Hz region is suppressed by the digital clutter filter. In some range bins, corresponding to the skip zone for ionospheric propagation, the noise level is below the display threshold in all Doppler bins. In the succeeding range bins, generally corresponding to the ranges of first-hop ground backscatter, the noise level in all Doppler bins is much higher, hindering target detection. (U) {(S)} That the noise seen on the radar was in fact clutter-related was demonstrated clearly, by turning off the radar transmitters. This caused the display of Fig. 7 to go black. When the threshold was readjusted to observe the noise level, it was observed to be constant with range, as one would expect from external noise. After a number of such observations, it became apparent that even if a clear channel could be found, even if ionospheric propagation to the desired geographical area existed at the clear channel frequency, and even if the radar display limitations could be overcome, the excess noise would still provide a severe limitation on radar performance. Therefore, in parallel with DVST, an effort to characterize the excess noise was undertaken on site. CHARACTERISTICS OF THE EXCESS NOISE (U) {(S)} The radar displays presented the excess noise in a dramatic way, but a quantitative characterization of the phenomenon required the use of off-line digital signal processing programs. (18) The output of one of these programs is illustrated in Fig. 8 for data recorded near 7:00 Greenwich mean time on March 4, 1972 (Day 64) in beam 11 with vertical polarization at 22.1 MHz - the same data as previously illustrated in Fig. 7. [ Fig.7 too poor an image to reproduce ] The variation of ground backscatter (clutter) and excess noise amplitude with slant range is plotted in Fig. 8. Note that the clutter curve is moved downward by 50 dB to facilitate comparison with the noise curve. Ground-clutter amplitude was computed by peak selection in a +-1.5-Hz Doppler window. The amplitude of excess noise was computed by averaging the squared modulus of the digital signal processor output over all Doppler bins from 3 to 20 Hz on either side of the carrier, that is, over all those Doppler bins outside the radar clutter filter rejection band. The digital signal processor performed a fast Fourier transform (FFT) over 512 successive radar pulses in each 12.8-sec coherent integration interval. The plotted clutter and noise powers were then further noncoherently averaged over 15 successive coherent integration intervals. (U) {(S)} One sees in Fig. 8 a marked variation of the excess noise amplitude with slant range. Strong excess noise exists at short range, in the skip zone just ahead of the ground clutter, and at the range of the ground clutter. The excess noise near the range of peak ground backscatter varies with range in direct proportion to the backscatter, {SECRET NOFORN} UNCLASSIFIED JDR 301 ============================================================== UNCLASSIFIED {SECRET NORFORN} Figure 8. Clutter and noise amplitude versus slant range. (17) {(Figure Classified Secret.)} (U) Figure 9. Approach/recede difference in noise amplitude. (17) {(Figure Classified Secret.)} (U) Figure 10. Approach/recede difference in noise amplitude. (17) {(Figure Classified Secret.)} (U) 302 JDR UNCLASSIFIED {SECRET NOFORN} ============================================================= {SECRET NOFORN} UNCLASSIFIED Figure 11. Range/Doppler display of excess noise (17) {(Figure classified Secret)} (U) but the excess noise at short range and in the skip zone does not. To distinguish between the excess noise that occur ahead of the ground clutter and the excess noise that occurs at the range of ground clutter, special terminology was used at the site. All sources of excess noise that varied with range were termed "range-related noise" (RRN). The portion of the excess noise that coincided in range with ground clutter was termed "clutter-related noise" (CRN). Although all of the range-related noise was of scientific interest, only the clutter-related noise interfered with detection of aircraft, which was the primary mission of the radar. To better characterize range-related noise, Figs. 9 and 10 were generated from the same data. (17) Here, the average power of range-related noise is computed separately for approaching and receding Doppler bins. In Fig. 9, noise power is averaged over Doppler bins 3 to 10 Hz from the carrier, while in Fig. 10 noise power is averaged over Doppler bins 10 to 20 Hz from the carrier. One sees that the clutter-related noise near the range of peak ground backscatter has a symmetrical spectrum close in (3 to 10 HZ) and a nearly Figure 12. Range/Doppler display of excess noise (17) {(Figure classified Secret)} (U) {SECRET NOFORN} UNCLASSIFIED JDR 303 ============================================================= UNCLASSIFIED {SECRET NORFORN} Figure 13. Doppler/time display of excess noise. (17) {(Figure Classified Secret.)} (U) symmetrical spectrum farther out (10 to 20 HZ). The range-related noise in the skip zone, on the other hand, has considerably more power in receding Doppler bins than in approaching Doppler bins. So does the range-related noise at short range, particularly from the 3- to 10-HZ Doppler bins. By comparing Figs. 9 and 10, one sees that range-related noise does decrease with increasing Doppler shift, although the true spectral extent of the range-related noise is obscured by the spectral replication that occurs at the radar PRF of 40 Hz. (U) {(S)} A more graphic, if less quantitative, presentation of the computer-processed data is given in Figs. 11 and 12. (17) Amplitude versus frequency (Doppler) is plotted over a +- 20-Hz band for successive 80-nmi radar-range bins during one particular 12.8-sec coherent integration interval. Figure 11 shows the approaching Doppler bins, while Fig. 12 shows the receding Doppler bins. This presentation shows the variation of range related noise with range and Doppler quite clearly. To show the temporal fluctuation in noise level, as well as the spectral rolloff, Fig. 13 shows amplitude versus Doppler during successive coherent integration intervals, for radar-range bin 17, which is at 1,280-nmi slant range, near the peak of the ground backscatter. The data in Figs. 11, 12, and 13 are thresholded, which has been found to produce a more easily interpreted display. The Figure 14. Doppler/time display of excess noise. (17) {(Figure Classified Secret.)} (U) 304 JDR UNCLASSIFIED {SECRET NOFORN} ============================================================= {SECRET NOFORN} UNCLASSIFIED threshold for Fig. 13 is 20 dB higher than the thresholds of Figs. 11 and 12 to show the clutter related noise peaks more clearly. (U) {(S)} Figure 14 shows amplitude versus Doppler during successive coherent integration intervals for radar-range bin 23, which is 480 nmi behind the peak of the ground backscatter, but still illuminated by one-hop ionospheric refraction. (17) In particular, this range bin, at a slant- range of 1,760 nmi, represents a ray path elevation of only a few degrees at ground level for one-hop propagation by means of the F2 layer of the ionosphere. The amplitude of range-related noise is much lower in range bin 23 than in range bin 17, which can be seen by noting that Fig. 14 has a threshold 20 dB lower than that of Fig. 13. One also notes in Fig. 14 a number of possible aircraft tracks (large amplitude returns isolated in Doppler and forming a Doppler-time trace) from the geographical area illuminated, which contained a number of Soviet military airfields. All of these apparent target returns in range bin 23 are well below the level of clutter-related noise seen in range bin 17. Thus, if the targets were in range bin 17, 480 nmi closer to the radar, they probably would not have been detected, even allowing for a 5.5-dB greater radar return due to the decreased range. Figure 14 illustrates the contention made earlier that aircraft detectability was not maximized at the range of peak ground backscatter, but rather at somewhat greater ranges, where grazing incidence for ground backscatter was approached. (U) {(S)} The radar data illustrated, taken on a single day early in the DVST period, are reasonably representative of the range-related noise phenomenon. Characteristics of range-related noise observed throughout the period of AN/FPS-95 operation are summarized here:
THE SEARCH FOR SOURCES OF EXCESS NOISE IN THE RADAR (U) {(S)} Once the effects of clutter-related noise on radar performance were understood, the AN/FPS-95 underwent extensive testing to see if the clutter-related noise might be originating in the equipment itself. There were two motives for first testing the radar itself before carrying the investigation to possible external causes of clutter-related noise:
{SECRET NOFORN} UNCLASSIFIED JDR 305 ============================================================= UNCLASSIFIED {SECRET NORFORN}
(U) {(S)} The results of this extensive radar testing are described here. First is a concise account of each of the principal tests conducted. The tests are grouped according to the radar component tested, rather than chronologically. Second, the main equipment-related, hypothesized causes of clutter related noise are listed and compared with the test results by means of a matrix. Third, the overall conclusions of the radar testing are given, namely, that the radar hardware was not the principal cause of clutter-related noise. EQUIPMENT TESTS FOR CLUTTER-RELATED NOISE (U) {(S)} The equipment test descriptions, while not a complete account of all tests conducted on site, represent most of the tests for which documentation is available, and we believe they give a reasonably comprehensive picture of the radar's capabilities. Tests of the radar transmitting chain are described first. The objective of these tests was to measure spectral noise on the transmitted signal, to see if it was comparable in amplitude to the levels of clutter-related noise observed on ground clutter. Test 1: Transmitter Noise Level (19) * (U) Description: Measure spectral noise level of HPA #4 at 10 and 30
Hz from carrier Test 2: Fan Dipole on Sea Wall #1 (20) * (U) Description: Transmit at high power from one or more strings of
the radar antenna. Connect Test 3: Fan Dipole on Sea Wall #2 (21) * (U) Description: Transmit at full power from five strings of beam 13.
Connect fan dipole to (U) {(S)} A number of receiving chains tests are now described, each directed toward a possible cause of clutter-related noise. The objective of these tests was to determine the linear dynamic range of the radar receiver end other components in the receiving chain, to see if the observed clutter related noise could be originating in the receiving chain. The general approach was to inject test signals of high spectral purity at various points in the receive chain and to measure the resulting spectral noise at various output points. Test 4: Receiver Linear Dynamic Range (19) * (U) Description: Input-output testing of receiver was performed to
measure
selectivity, noise Test 5: Receiver Intermodulation (22) * (U) Description: Measure in-band intermodulation (IM) levels at
several
frequencies. Test 6: Receiver and Duplexer Cross Modulation (CM) (23) 306 JDR UNCLASSIFIED {SECRET NOFORN} ============================================================= {SECRET NOFORN} UNCLASSIFIED * (U) Description: Inject two test signals 10 HZ apart at 1 MHZ from the
desired signal. Test 7: Radio-frequency Hardware Measurements (19) * (U) Description: Measure spurious-free dynamic range of
transmit/receive diodes and magnetic Test 8: Electromagnetic Interference Measurements (24) * (U) Description: Measure the power level of interfering HF signals at
the receiver input, mostly Test 9: Simulated Clutter into Receiver (25) * (U) Description: Inject a simulated clutter signal into the receiver
at a range in the skip zone, Test 10: Receiver Attenuation (25) * (U) Description: Attenuate received ground clutter from beam 7 at the
receiver input in 6-dB (U) {(S-NF)} Next, a number of tests of the radar antenna on reception are described. Spectrally clean test signals were radiated toward the radar antenna from various points in the local area, and the received signals were examined for spectral noise of a level comparable to the observed clutter- related noise. One might note that extensive rework of the antenna was undertaken by the contractor (RCA) from Aug. 4 to Sept. 17, 1972. Antenna tests before the repairs were made showed a higher level of spectral noise than subsequent tests, which tended to exonerate the reworked antenna as the principal cause of clutter related noise. Test 11: Loop Antenna at the focal point (25) * (U) Description: Radiate a simulated clutter signal from a loop
antenna located at the geometrical Test 12: Monopole Antenna on Sea Wall (25) * (U) Description: Radiate a test signal from the vertically polarized
monopole. Receive on beam Test 13: Vertical Dipole on Sea Wall (21) * (U) Description: Radiate a test signal from a vertically polarized
dipole. Receive on beam 13 {SECRET NOFORN} UNCLASSIFIED JDR 307 ============================================================= UNCLASSIFIED {SECRET NORFORN} * (U) Results: Spectral noise on test signal was down 91 to 95 dB from
carrier. Test 14: Helicopter-borne Test Antenna (20,21) * (U) Description: Suspend a 1-watt CW transmitter and a vertical dipole antenna
300 ft. below Test 15: SAC Test 1 - Monopole on Sea Wall (26) * (U) Description: Radiate a test signal from a Fluke frequency
synthesizer through a vertically Test 16: SAC Test 1 - Airborne One-Way Measurements (26) * (U) Description: Radiate a CW test signal from an antenna trailed
behind an aircraft orbiting (U) The next group of tests involved simultaneous testing of the transmitting and receiving chains of the radar, utilizing either a specially constructed signal repeater in the local area or actual ground clutter from OTH ranges. Although the radar antenna both transmitted and received during these tests, in some cases a supplemental nine element Yagi-Uda test antenna (27) was also used so that a comparison with the radar antenna could be made. Test 17: Repeater on the Sea Wall (28) * (U) Description: A signal repeater with a 1/8 wavelength monopole
receiving antenna, delay line, Test 18: SAC Test 1 - Aircraft Repeater (26) * (U) Description: The above repeater was placed in an aircraft, which orbited 4 miles from the Test 19: Transmitter Power Reduction (25) * (U) Description: Reduce transmitter power in 3-dB steps to 12 dB,
during normal radar Test 20: SAC Test 2 - Subclutter Visibility with an Auxiliary Antenna. (29) * (U) {(S)} Description: Alternately transmit and receive on string 16
of the radar antenna and the 308 JDR UNCLASSIFIED {SECRET NOFORN} ============================================================= {SECRET NOFORN} UNCLASSIFIED EQUIPMENT COMPONENTS AS POSSIBLE SOURCES OF (U) {(S)} All components of the radar, including its local environment, were considered as possible sources of clutter-related noise. For each component, one or more physical mechanisms capable of generating clutter-related noise were hypothesized. These mechanisms (see Table 5) were then considered in structuring the equipment tests for clutter-related noise. Table 6 summarizes the results of the 20 equipment tests as described earlier with respect to sources of clutter-related noise in each radar component. A minus sign (-) means that a given radar component was found not to be a significant source of clutter-related noise; a plus sign (+) means that a component was found to be significant. Many squares in the table are left blank, indicating no conclusive relationship between a given equipment test and a given radar component. (U) {(S)} Spectral noise on the radar transmitter output could cause clutter-related noise to appear on ground clutter. The ratio of clutter to clutter related noise expected would be approximately equal to the ratio of carrier to spectral noise on the transmitter output, if such spectral noise were the principal cause of clutter-related noise. The transmitter noise level measurement (Test 1) showed a very low level of spectral noise - much too low to account for the observed clutter-related noise. A test using a fan dipole on the sea wall (Test 3) also showed transmitted spectral noise to be much lower than the generally observed clutter-related noise. The two overall system tests using a repeater (Tests 17 and 18) also tended to clear the transmitter as a cause of clutter-related noise. Finally, the observed linearity of clutter-related noise with transmitter power (Test 19) was an indication that nonlinear effects in the transmitter were not a significant source of clutter-related noise. (U) {(S)} Receiver testing was more extensive than transmitter testing, in part because numerous tests of the radar antenna also implicitly tested the radar receiver. Tests of receiver linearity, dynamic range, intermodulation distortion, and cross modulation distortion (Tests 4, 5, and 6) showed that the spectral noise imposed upon received ground clutter by these receiver phenomena should be much lower than the levels of clutter-related TABLE 5. Physical Mechanisms for clutter-related noise (Table unclassified.) --------------------------------------------------------------------------
-------------------------------------------------------------------------- noise actually observed, thus showing that the receiver was not the major cause of clutter-related noise. The electromagnetic interference measurements (Test 8) showed that out-of-band electromagnetic interference occasionally exceeded receiver ratings, which could allow cross modulation in the receiver to cause significant clutter-related noise. However, such large out-of-band electromagnetic interference was rare, whereas clutter related noise was observed all the time when ground clutter was strong. Testing of the receiver with simulated clutter (Test 9) showed spectral noise on the simulated clutter to be smaller than observed clutter-related noise on actual clutter received at the same time. Actual clutter and clutter-related noise were also shown to be linear with received signal attenuation (Test 10), thus indicating that receiver overload was not a cause of clutter-related noise. (U) {(S)} Antenna reception Tests 13 through 16 included the receiver in the test chain. Since low levels of spectral noise were observed, these tests also exonerated the receiver as the principal cause of clutter-related noise. For example, the test employing a vertical dipole on the sea wall as a signal source (Test 13) showed spectral noise down 91 to 95 dB from the carrier, which is far lower than the clutter-related noise-to-clutter ratios commonly observed. Transmit and receive system tests employing a signal repeater (Tests 17 and 18) also tended to exonerate the receiver, although the spectral noise in these cases was not quite so low, because of the limitations of the repeater. {SECRET NOFORN} UNCLASSIFIED JDR 309 ============================================================= TABLE 6. Equipment components versus clutter-related
noise tests. 310 JDR UNCLASSIFIED {SECRET NOFORN} ============================================================= {SECRET NOFORN} UNCLASSIFIED (U) {(S)} The receiving chain and signal repeater tests also indicated that the radar signal processor and displays were not the cause of clutter-related noise. That is, spectrally clean test signals injected ahead of the signal processor were observed on the displays to be not corrupted with spectral noise to anything near the level of clutter-related noise observed on actual ground clutter. There was a further indication that all the radar signal processor circuits after the analog-to-digital converter had adequate linear dynamic range to properly spectrum analyze ground clutter. An off-line digital signal processor was developed by MITRE (18) to supplement the on-line hybrid digital/analog radar signal processor. Careful comparison of clutter related noise at the output of the off-line digital processor with clutter-related noise observed on the radar displays showed very close agreement in amplitude, spectrum, and time variation. While it is possible that both processors might have had an undetected flaw, it is extremely unlikely that both would have had exactly the same flaw. (U) {(S)} A lingering doubt does exist about one component of the signal processor - the analog-to digital converter. A colleague (30) has put forth the hypothesis that analog-to-digital converter transient response errors in following time-varying clutter might account for the spectrally spread clutter-related noise. Since all test signals, both CW and pulsed, had constant amplitude from one radar pulse repetition interval to the next, the transient response of the analog-to-digital converter may not have been adequately tested, according to this hypothesis. At this late date, there appears no way to resolve this question. (U) {(S)} Four tests of the radar antenna, ground screen, and RF hardware were conducted in the spring and summer of 1972, before RCA reworked these components. The transmitter power reduction test 19 tended to rule out nonlinear effects in the transmitting antenna, such as arcing and corona, as the principal cause of clutter-related noise, but it was too limited in scope to be wholly conclusive. Cross-modulation distortion in the duplexers was measured in conjunction with similar receiver measurements (Test 6) and found to be negligible. Two early tests of the antenna on reception gave positive results, however. Spectrally clean test signals that were radiated from a loop antenna at the geometric focal point of the radar antenna (Test 11) showed spectral noise at the signal processor output comparable in amplitude to the clutter-related noise simultaneously observed on ground clutter. When a vertical monopole on the sea wall was used to radiate a spectrally clean test signal (Test 12) to the radar antenna, spectral noise was also observed at the signal processor output. With the radar antenna vertically polarized, spectral noise on the test signal was lower than the clutter-related noise usually observed on ground clutter. However, with the radar antenna horizontally polarized (cross polarized to the test signal), spectral noise on the test signal at some frequencies was higher than the clutter-related noise usually observed on ground clutter. The results of these two tests were taken as an indication that at least some of the clutter-related noise was originating in the radar antenna on reception - possibly in the ground screen, because it could produce cross-polarized spectral noise. (U) As a result of these early antenna tests, a team of engineers from RCA Moorestown, the AN/FPS-95 contractor, came to the site in the fall of 1972. They inspected the antenna, ground screen, and RF hardware, had extensive repairs and rework done, and then participated in further tests of the reworked antenna. Rework of the antenna was conducted between Aug. 4 and Sept. 17, 1972. Expansion sections in the RF hardware and the ground screen clips were both found to generate spectral noise during two-tone intermodulation tests; corroded joints were also found by visual inspection. The expansion sections were replaced, steel towers in the antenna field were rewelded to reduce nonlinear RF effects at joints, grounding connections were improved, and the ground screen clips were welded, Coaxial lines to the baluns were modified and grounded, as were certain conduits and fan plates. All loose metal debris in the antenna field was removed. (U) After the rework, extensive testing of the antenna on both transmission and reception was performed. Transmitting tests were directed toward both linear sources of spectral noise, such as wind vibration, and nonlinear sources, such as arcing and corona. Receiving tests took into account wind vibration also, as well as nonlinear effects, such as rectifying action at joints in the antenna and ground screen. {SECRET NOFORN} UNCLASSIFIED JDR 311 ============================================================= UNCLASSIFIED {SECRET NORFORN} (U) {(S)} Transmission from the radar antenna to a fan dipole on a sea wall (Tests 2 and 3) showed low spectral noise when the radar antenna and test antenna had the same polarization, but higher spectral noise when the two antennas were cross polarized. This indicated that some spectral noise generation mechanisms still existed in the antenna on transmission, especially at the lower operating frequencies, but that these were probably not the principal cause of clutter-related noise. In this regard, it should be noted that the spectral noise uncovered in cross-polarization testing was concentrated in the region of 10 to 30 HZ from the carrier, whereas clutter-related noise associated with ground clutter had a spectrum that fell off slowly, but monotonically, with Doppler shift. While cross-polarization testing was a good way to uncover spectral noise generation mechanisms in the antenna, it differed too much from normal radar operation to be readily interpreted in terms of the clutter-related noise that could be expected on actual ground clutter. The two repeater tests of the whole radar system (tests 17 and 18) showed no noise generation in the antenna, within the limits of the instrumentation. Finally, a comparison of clutter to clutter-related noise ratios between one string of the radar antenna and an auxiliary test antenna having a roughly similar radiation pattern (Test 20) showed no significant differences. This was a further indication that the radar antenna was not the principal cause of clutter-related noise. (U) {(S)} Several reception tests were also performed on the reworked antenna. The RF hardware measurements (Test 7) showed no significant source of clutter-related noise, unless the level of electro-magnetic interference at the antenna terminals approached 0 dBM, which rarely occurred. Spectrally clean test signals were radiated toward the receiving antenna (Tests 13 through 16) from the sea wall, from a helicopter, and from an aircraft. The sea wall provided a stable mounting point for a test antenna, allowing very low levels of spectral noise (more than 90 dB down from the carrier) to be observed at the signal processor output. However, the sea wall was not in the far field of the antenna and was lower in elevation than either of the radar elevation beams. The helicopter and aircraft tests were intended to get further from the antenna and in the main lobe of one of the radar elevation beams. These tests also showed the radar antennas on reception to be relatively free of spectral noise generation, but not to the degree of the sea wall tests, because the airborne platforms were not sufficiently stable. It might be noted that some of these tests were done on windy days, so the hypothesis that wind vibration of radar antenna elements caused clutter-related noise was thoroughly tested. Wind vibration was shown to produce some spectral noise, but was a minor contributor to the observed levels of clutter related noise. (U) {(S)} The local environment of the radar system was included in some of the equipment tests for clutter-related noise. Receiving chain tests with the helicopter-borne test antenna (Test 14) and the airborne one-way measurements (Test 16) included some of the sea within the first Fresnel zone of the antenna ground plane. Most of the first Fresnel zone was within the aircraft test range of 4 miles; only part was within the helicopter test range of 1/2 mile. Both of these tests showed received spectral noise almost 90 dB down from the carrier at the Doppler frequencies of interest. Apparently, significant modulation of the signal by sea waves was confined to smaller Doppler shifts. The low level of spectral noise observed during these tests also negated the hypothesis that clutter-related noise might be caused by cross-modulation effects in ancillary equipment near the radar site, such as transmission lines, power stations, and so on. (U) {(S)} The aircraft repeater measurements (Test 18) were conducted with the repeater at a range of 4 miles from the radar. These measurements uncovered no spectral noise that could be attributed to the local environment. Because of instrumentation limitations, this test was not quite as conclusive as the one-way receiving chain tests. Both the aircraft repeater test and the auxiliary antenna measurements (Test 20) made use of a Yagi-Uda antenna located about 1/2 mile south of the radar antenna. During these tests, no significant differences in spectral noise on test signals or in clutter related noise on ground clutter were found between the two antennas. This was a further indication that local environmental effects, which were some what spatially decorrelated between the two antennas, were not the principal cause of clutter- 312 JDR UNCLASSIFIED {SECRET NOFORN} ============================================================= {SECRET NOFORN} UNCLASSIFIED Figure 15. Ray path geometry. (Figure unclassified.) related noise. One further piece of evidence - the absence of nonlinear effects during the transmitter power reduction Test 19 - tended to negate the hypothesis of cross modulation in ancillary equipment. CONCLUSIONS OF EQUIPMENT TESTING (U) {(S)} Numerous tests of the AN/FPS-95 transmitter showed it to have exceptional spectral purity and to be a negligible contributor to the overall level of clutter-related noise. The radar receiver, always a prime suspect as the cause of clutter-related noise, was very thoroughly tested for spectral noise generation. It, too, was exonerated, except when very large out-of-band inteferers were present at the receiver input. Since such interferers were rarely present, whereas clutter-related noise was always present when OTH propagation was good, it was concluded that the radar receiver was not the principal cause of clutter-related noise. The radar signal processor was shown through numerous tests to not be a significant source of clutter-related noise. A minority opinion (30) would have it that these tests did not adequately measure the analog-to-digital converter transient response to time-varying clutter. (U) {(S)} Some spectral noise generation mechanisms were found in the AN/FPS-95 antenna, ground screen, and RF hardware. After extensive rework of these components by RCA, such noise generation mechanisms were considerably reduced, but still present. Extensive system testing on both transmission and reception showed that the antenna, ground screen, and RF hardware were not the principal cause of clutter-related noise. These components had particularly good spectral purity above 20 MHZ, whereas clutter-related noise on actual ground clutter was just as prevalent as it was to lower radio frequencies. Some tests of the radiating system also included its local environment, particularly the sea. The local environment seemed no significant source of such noise. (U) {(S)} Having rather thoroughly exonerated the radar equipment as the limiting source of clutter-related noise, attention turned to factors external to the radar. Both the ionospheric propagation medium and radar reflectors in the target space were considered as sources of clutter-related noise, as discussed in the next section. THE SEARCH FOR SOURCES OF EXCESS NOISE (U) After an introductory discussion of propagation geometry, this section gives brief descriptions of all the relevant experiments and tests, followed by discussions of postulated causes of noise due to reflection effects and propagation phenomena. PROPAGATION GEOMETRY Figure 15 shows an idealized diagram of the propagation ray paths typical of radar operation using the ionosphere F-layer as the reflecting layer (the normal mode of operation). The rays emanating from the radar located at R are shown as being restricted to a range of elevation angles bounded by the lower ray path R-E3 and the upper ray path R-E1. In fact, of course, the actual elevation gain pattern did not have such sharp boundaries. It featured a direction of maximum gain that could be switched between an upper elevation angle of approximately 15 deg and a lower position of typically 5 to 7.5 deg by selecting, respectively, horizontal and vertical polarizations. The measured antenna patterns indicated considerable variations in elevation beam shape as a function of beam number and radar frequency. {SECRET NOFORN} UNCLASSIFIED JDR 313 ============================================================= UNCLASSIFIED {SECRET NORFORN} Figure 16. Radar-range relationships. {(Figure classified Secret)} (U) (U) The radiation pattern gain did approach zero at a zero elevation angle, due to horizon shielding, but there was considerable antenna gain at angles higher than the nominal upper limit of the beam represented by ray path R-E1. The effect of the high-angle radiation was not usually important, however, since these rays would normally pass through the F-layer end thus would not contribute to the normal OTH radar process. In Fig. 15, the highest ray that is reflected from, rather than penetrating, the F-layer is shown by the ray path R-F1-G1. This ray is shown as passing through the E-layer at points E2 and E4. Similarly, the lowest ray follows the path R-F2-G2 and intersects the E-layer at E3 and E5. The ionospheric layers do not behave exactly as simple mirrors, as indicated in Fig. 15. The effective reflection heights, such as those of points F1 and F2, are a function of the radar frequency, the incidence angles of the radiation, and the ion density profile. Usually, at a fixed frequency the rays having the larger elevation angles at the radar will penetrate higher. Thus, point F1 would be slightly higher than F2. This phenomenon gives rise to a focusing effect, with the result that the effective gain of the antenna, when viewed looking back up the ray path from the region following the ionospheric reflection, is modified from that of the prereflection region. Usually, the effect is to increase the gain along rays in the region close to the uppermost ray (F1-G1). Figure 16 shows, along the abscissa, the relative time delays from the radar for reflections assumed to occur at the labeled points corresponding with the notation in Fig. 15. (U) The radar-range dependence of the observed amplitude of OTH ground clutter, as sketched in Fig. 16, may be accounted for as follows. On the radar side of point G1, no ground reflections are possible, because the ionosphere does not reflect the rays leaving the radar at elevation angles higher than that of R-F1. Moving from point G1 away from the radar, the ground reflections, as received back at the radar, build up in a particularly abrupt manner, since this is the region of enhanced gain due to the ionospheric focusing action. Beyond this point, the combined effects of increasing range, reducing antenna gain, and diminishing earth-grazing angle produce a rapid reduction in the received backscatter amplitude. For the depicted one-hop propagation mode, the ground backscattcr should effectively disappear at ranges greater than that appropriate to the point G2. DESCRIPTION OF EXPERIMENTS AND OBSERVATIONS Synoptic Measurements (U) {(S)} During the approximately 18-month life of the AN/FPS-95, many observations were made and recorded in the form of notes, photographs, computer printouts, and magnetic tapes. Much of this information was relevant to the investigation of clutter-related noise. Unfortunately, these data were taken using a variety of radar operating parameters and analyzed by a number of different methods. The consequence was that the usefulness of the data for investigating the relationship between clutter-related noise and any single radar or operational parameter was impaired. (U) {(S)} To supplement the above data base, during the clutter-related noise investigation of the Scientific Assessment Committee, a concentrated synoptic data-gathering activity was conducted during February 1973 for a period of 19 days.(31) A daily schedule of 12 data-gathering runs was made on 314 JDR UNCLASSIFIED {SECRET NOFORN} ============================================================= {SECRET NOFORN} UNCLASSIFIED each of beams 1, 7, and 12. During and beyond the duration of this test, local weather, solar flux, and the geomagnetic index were recorded to permit the investigation of possible correlation with clutter-related noise. The results of synoptic test data analysis clearly confirmed the persistent existence of the short-range, precursor, and clutter-related noises. They did not, however, reveal any clear correlation between the clutter related noise and local weather, solar, or geomagnetic parameters. (U) {(S)} An interesting effect noted in data recorded between September 1972 and May 1973 is that the ratio of ground clutter to clutter-related noise appeared to vary distinctly as a function of beam azimuth. The relative amount of noise was lowest in beam 1, rose gradually through beam 9, then dropped again until the most southerly beam 13 was reached. (32) The maximum variation (beam 1 to beam 9) was approximately 10 dB. Land/Sea Test (33) (U) {(S)} The object of this test was to investigate the hypothesis that the clutter-related noise was generated, through the modulation and backscattering of radar energy, by objects situated on or near the earth's surface, at ranges normally illuminated by the one-hop OTH radar propagation modes. Because of the importance of this experiment and its results, it is described in greater detail in the appendix at the end of this paper (U) {(S)} The test was arranged to measure the clutter-related noise powers from range-azimuth resolution cells within an area of AN/FPS-95 coverage encompassing both land and sea areas. The greatest variations in clutter-related noise levels were found to occur between adjacent land and sea areas. These results were not inconsistent with the assumption that no clutter-related noise was generated within the resolution cells located over the sea. Short-Range Noise Test (34) (U) The primary purpose of this test was to identify the sources of the component of range related noise observed to occur at short radar ranges (less than approximately 600 nmi). The particular postulated mechanisms investigated were transmit/receive switch transients, transmitter-induced corona, antenna vibration, and meteor effects. (U) Examination of off-line processed data showed effects were confined to extremely short radar ranges and that they could be ignored at the ranges of the observed short-range noise. Earlier in the AN/FPS-95 testing program, the presence of more serious switching transients had been observed using the on-line signal processor. These were subsequently reduced by an equipment modification. In this connection, it should be noted that the vast majority of data used in the investigations of range-related noise were analyzed by off-line techniques. (U) Although antenna arcing had previously been observed at lower radar operating frequencies, measurements at 23 MHZ, the frequency used for most short-range noise tests, failed to reveal any evidence of the phenomenon. (U) A measurement made at a frequency of 23 MHZ in beam 13 using vertical polarization contained a surface wave clutter signal at a range of 40 nmi. The amplitude of this signal was sufficiently higher than the noise background of the spectrally analyzed data to permit the conclusion that any spectral spreading of the signal (by antenna vibration) would be down by at least 66 dB. This conclusion does not, of course, necessarily exonerate the antenna at other frequencies, beam positions, and polarizations. (U) {(S)} The main effort in this test was devoted to an examination of the meteor theory of short-range noise generation. The noise was recorded in beam 1 and beam 13 for each of the two available antenna polarizations (vertical and horizontal). Changing the polarization had the effect of raising the beam from a lower position to a higher position. The radar ranges of the recorded short-range noise were seen to shift in toward the radar when the beam was raised, in accordance with the hypothesis of backscattering occurring within the E-layer. The recorded data were used to calculate the antenna vertical beamshape for subsequent comparison with independent measured patterns. A good correspondence was thus obtained. The above measurements were performed both above and below the maximum usable frequency, at a frequency of 23 MHZ, by choosing the appropriate diurnal time. Some of the measurements, when operation was below the maximum usable frequency, were made at a low pulse repetition rate {SECRET NOFORN} UNCLASSIFIED JDR 315 =============================================================
|
| to Top of this Page | On to Part 3 | Back to Part 1 | To CUFON Main Page |
| C
U F O NSM The Computer UFO Network http://www.cufon.org/ SYSOP - Jim Klotz UFO Reporting and Information Service
|