THE UNIVERSITY OF MICHIGAN Reports Control Symbol 1084-2-Q OSD-1366 OSD -1366 Technical Report ECOM-0547-2 March 1968 Azimuth and Elevation Direction Finder Techniques Second Quarterly Report 1 October - 31 December 1967 Report No. 2 Contract DAAB07-67-C0547 DA Project 5A6 79191 D902-05-11 Prepared by J. E. Ferris and W. E. Zimmerman The University of Michigan Radiation Laboratory Department of Electrical Engineering Ann Arbor, Michigan For United States Army Electronics Command, Fort Monmouth, N.J. DISTRIBUTION STATEMENT Each transmittal of this document outside the Department of Defense must have prior approval of CG, U. S. Army Electronics Command, Fort Monmouth, New Jersey, 07703, ATTN: AMSEL-WL-S.

THE UNIVERSITY OF MICHIGAN 1084-2-Q ABSTRACT This report discusses the construction and operation of the aximuth-elevation direction finder. It considers the proposed antennas and the antenna feed networks and the techniques being developed to achieve the design specifications. While the theory is well understood there are experimental design problems which are discussed along with their effect on the basic feed network. Several methods of signal detection are described with relative merit for each system presented. The report illustrates the function of the computer in the signal detection system and inthe computation of aximuth and elevation angles of the monitored signal. Several small computers and analog-to-digital converters were considered during this period and the final choice is described. A brief survey of the effects of ground reflection on the operation of the system is given. However, due to the complexity of the problem no definite conclusions have been reached. ii

THE UNIVERSITY OF MICHIGAN 1084-2-Q FOREWORD This report was prepared by The University of Michigan Radiation Laboratory of the Department of Electrical Engineering under United States Army Electronics Command Contract No. DAAB07-67-C0547. This contract was initiated under United States Army Project No. 5A6 79191 D902-05-11 "Azimuth and Elevation Direction Finder Techniques". The work is administered under the direction of the Electronics Warfare Division, Advanced Techniques Branch at Fort Monmouth, New Jersey. Mr. S. Stiber is the Project Manager and Mr. E. Ivone is the Contract Monitor. The authors wish to express their thanks to Dr. B. L. J. Rao for his theoretical contributions; to Messrs. A.J. Loudon, D.R. Marble, E. C. Bublitz and C. D. Spragg for their efforts in the experimental work performed during this period, and to Mr. P.H. Wilcox for preparing the computer program required for the three-dimensional analysis. The material reported herein represents the results of the continuing investigation into the study of techniques designing a broadband circularly polarized azimuth and elevation direction finder antenna. 111

THE UNIVERSITY OF MICHIGAN 1084-2-Q TABLE OF CONTENTS ABSTRACT ii FOREWORD iii LIST OF ILLUSTRATIONS v I INTRODUCTION 1 II ANTENNA SYSTEM AND PHASING NETWORK 6 III SURFACE REQUIREMENTS 27 IV ELECTROMECHANICAL SWITCH 28 V COMPUTER AND DETECTION SYSTEM 21 VI GROUND REFLECTIONS 37 VII CONCLUSIONS, 39 REFERENCES 41 DD FORM 1473 42 iv

THE UNIVERSITY OF MICHIGAN 1084-2-Q LIST OF ILLUSTRATIONS 1-1 Spherical Coordinate System Used for Azimuth - Elevation Direction Finder 3 2-1 Antenna Feed Network 8 2-2 Single Section Quarterwave Directional Coupler 10 2-3 Three Section Quarterwave Coupler 13 2-4 Theoretical Coupling Curve for Seven Section Coupler 19 2-5 Graph of Measured Coupling Curve and Coupling Curve Predicted From Experimentally Determined's 21 2-6 Irradiated Dielectric Stripline after Etching Process 23 4-1 Engineering Model of Electromechanical Switch 29 4-2 Top Secticn Indicates Geometry of Rotation Capacitive Probe and Rotary Joint 30 v

THE UNIVERSITY OF MICHIGAN 1084-2-Q INTRODUCTION This report describes the construction and operation of the components in the azimuth and elevation direction finder presently being designed and fabricated by the Radiation Laboratory. The azimuth-elevation direction finder consists of an antenna system, electromechanical switch, receiver (GFE), Automatic Signal Recognition unit (GFE), an A to D converter, computer, and a visual display. The antenna system consists of 16 to 17 antennas mounted on an arbitrary surface. Each antenna will have associated with it a particular 0 and 0 coordinate of the spherical coordinate system. Direction finding is accomplished by assigning to each antenna a vector whose amplitude is that of the received signal with the direction given by the outward pointing normal for the antenna receiving the signal. The signal information from the 17 antennas is fed into the computer where it is vectorially added, and the sum vector has 0 and 0 coordinates in the direction of the incoming signal. The antennas are multiplexed by a rotating electromechanical switch. Tenta tively, the switch will have three rotational rates, 10, 100, and 1000 rpm. The operator will then be able to select the optimum switching rate, which will depend upon the types of signals being interrogated by the azimuth and elevation direction finder. At present, signals which are considered to be of the greatest concern are those from search radars. Since the antennas are time multiplexed, one would like to have a very high switch rotational rate to assure interception of all radar pulses, especially those of a long range ground based rotating antenna. However, one must be careful not to switch at a rate faster than the computer cycling rate or the response time of the receiver. The cycling rate of the computer determines the switching rate since the inverse of the switch rotational rate times the number of antennas gives the time allowed for the computer to accept data from each antenna. Consideration must also be given to direction finding signals with CW, AM, and FM 1~ ~ ~ ~ ~~~n _______________M_____________

THE UNIVERSITY OF MICHIGAN 1084-2-Q modulation. However, it is generally agreed that these type signals will not present as severe a problem to this system as those of the radar pulses and do not directly affect the switch rotational rate. The switch will consist of 16 or 17 input ports with one output port. In addi- tion to the signal input, an identification or reference signal must be available in the switch to identify to the computer which antenna is being interrogated. This identificeation is conceived to be in the form of a light beam to be interrupted as the switch output port slews past each of the input ports. It is anticipated that the switch will interrogate each of the antennas such that information will be collected and stored in the computer in a sequential format. For example, the first antenna would be associ iated that 0 = 00,, = 0~ in the spherical coordinate system of Fig. 1-1; the second antenna would be at 0 = 450~, 00; the third antenna at position 0 = 450, = 450, I etc., to the last antenna. However, at the present time the effect of picking antennas that are widely separated in space orientation is being considered, This sequence of interrogation may have advantages for radar pulse detection in that a greater portion of the sky would be scanned in a shorter period than the cycle time of the switch, thereby increasing detection probability for fast moving beams. The accuracy per single scan probably would not be within contract requiremnents of 20 for azimuth and 50 for elevation, but consolidating information for several scans should give improved accuracy. At present the mounting surface has not been chosen but as is shown in Section III, on mounting surfaces and SectionVI on ground reflections there may be improved accuracy by using a surface that is planar rather than hemni spherical as originally proposed. The electromrechanical switch has a dual function to perxforrmn 1) switch the mricrowave energy, and 2) provide a reference signal for the computer. The referenLce signal identifying the coordinates of the interrogaed te nna is to be traenswc feried from the switch directly to the computer, After passing through the switch, the microwave signals are amplifed and reduced to an IF frequeincy by a receiver. I~~~_ __ __ U c P-'111 ~gr~if

THE UNIVERSITY OF MICHIGAN 1084-2-Q z y /X~E0 FIG. 1-1: Spherical Coordinate System Used For Azimuth-Elevation Direction Finder. #i~~~~~

THE UNIVERSITY OF MCHIGAN 1084-2-Q A video amplifier may be required between the receiver and the analog4to-digital converter to amplify the IF signal to the voltage level required by the analog- ton digital converterter. In the event that more than one frequency is received within the bandwidth of the receiver, an automatic signal recognition unit (ASRU) would be required. It is understood that such units are available and have been employed by the military for this purpose. The data from the ASRU is then transferred to the analog-to-digital converter where it is digitalized in the proper format for the coma puter. Due to the requirements for detecting radar pulses which may be quite narrow in time, a standard analog-to-digital converter will probably not be used. Furs ther discussion oif the detection method for detecting narrowv pulses during the entire switch aperture will be discussed in a later section (SectionV on the computer and the analog-to-digital converter). In the computer the data is stored sequentially similar to that discussed above, For example, the first storage compartment would consist of data for the pole antenna, or the first antenna. The second storage cornpartment would have data from the second antenna or as in the example, 0 - 45~, 0 - 0~, Each compartnrent would then have the data stored for a definite space orientation of that antenna throughout all 17 antennas. When all 17 antennas have been interrogated the computer will then vectorially add all the 17 vectors in storage and take the arc tangents to obtain the 0 and, direction of the arriving signal. I Accuracy of this system depends upon a well controlled antenna pattern with the limitations being discussed by Ferris, et al (1967). The present status of this antenna developrrent and the feed system for the antenna is discussed in Section II on the antenna and feed system design.

THE UNIVERSITY OF MICHIGAN 1084-2-Q II ANTENNA SYSTEM AND PHASING NETWORK During the first part of the experimental program a cavity-backed spiral was tested over the 5:1 band from 600 to 3000 MHz. The cavity-backed spiral was used to achieve the circular polarization required by the contract, but early in the program it became evident that a cavity-backed spiral could not give the antenna pattern control necessary over the required 5:1 frequency band. A cavity-backed spiral with the X/4 deep cavity has the proper boundary condition for reflected wave reinforcement only for a narrow range of frequencies.Operating such antennas over more than one octave results in pattern deterioration with the main beam of the antenna -pointing off-axis. It is desired that the antenna for this direction finding system should have a cosine pattern and that the maximum of the beam should be normal to the axis of the antenna. Because of the poor pattern control over the 5:1 band, the cavity-backed spiral was discarded and an investigation of the log conical was started. As was reported in the previous quarterly (Ferris, et al, 1967), a bifilar log spiral will set up moding at the higher frequencies. This moding results from higher order current bands that formed when the spirals electrical circumference is equal to (n + m) X where n is equal to 1 for the summation mode and 2 for the delta mode and m is equal to the number of filaments. A bifilar antenna operated in the sum mode (i. e., pattern maximum on-axis) has higher order modes at the third and fifth harmonics of the fundamental. Of course at all frequencies above the third harmonic, there would be some region of the antenna which will support the third harmonic mode. For a multifilar element the radiating current bands become more randomly distributed as the number of filaments increases. The quadrafilar spiral, whose four terminals are fed 00, 900, 1800, 2700, willhave aradiating band when the electrical circumference is 1 wavelength in diameter and the next higher order radiating band occurs when the spiral is 5 wavelengths in diameter. Although the quadrifilar spiral will be operating at the high frequency limit of the 5

. THE UNIVERSITY OF MICHIGAN 1084-2-Q 5:1 frequency band near the region of the 5 wavelength radiating band, ii.t.is felt that for a feasibility study this antenna is a compromise between desired. operating characteristics and cost for phasing networks. Better pattern control can be achieved by a further increase of the number of radiating windings, but a more complex phasing network and increased research and development time is required with the add ditional number of radiating arms. The quadrifilar feed network requires a power split and a phase shift which is more easily accomplished over these broadbands by stripline type transmission lines. A stripline phasing network that would give 00, 90, 180 and 2700 phase distribution across the 5:1 band, would employ three 3 db couplers and one 900 phase shifter as shown in Fig. 2-1. A 3 db coupler in stripline has the property that the output arm has a 900 phase shift from the non-coupled port, and this phase shift is relatively constant across the bandwidth of operation of the 3 db coupler as long as the 3 db coupling is maiaintained. By going through a 3 db coupler the outputs display (l equal power and have a 90~ phase shift with respect to each other. A total phase shift of 1800 is achieved by placing a 900 phase shifter in the output ports of the 3 db coupler. The addition of the 900 phase shifter causes the 3 db hybrid to display the characteristics of a,magic tee. If 3 db hybrid couplers are placed in each of the two outputs from the first 3 db coupler and 900 phase shifter, the resulting four outpluts ro-In the directional couplers are now divided equally in power and the phase of thesE four equal amplitudes is 00, 90, 180 and 270~ The 3 db coupler and 900 phase shifter are constructed from the data that has been documented by Shelton and Mosko (1966). By using a technique originally cong ceived by Shiffman (1958), a broadband phase shifter or hybrid coupler can be constructed to have an arbitrarily large frequency ba.nd. As the bandlwidth increases, there is a simrilar increase in the numnber of quarterwave seceions that are required in the directional coulpler and in the phase shifter. These sections are a quarter! w-avelength at the mid frequency of the operating band. It is well known that a 3 db 6..

THE UNIVERSITY OF MICHIGAN 1084-2-Q T3 d b CoBupler 0Q. 5ej(3 0. 5Vej03 3db Coupler O3db Coupler 2.0e V3 Terminated.......... Termminated Ve 0.(707Vej(-2r 0X.5Ve-J(~3 I) 0. 5VeJ(P3 ) 0. 707VeJO2 Coupled Section Reference LM 900 Phase Shifter 7 FIGo 2-1o Antenna Feed Network

THE UNIVERSITY OF MICHIGAN, 1084-2-Q coupler can be made with a single quarterwave length section. However, the bandwidth at most for such a single quarterwave section is approximately one octave. Figure 2-2 is a drawing of a simple quarterwave length coupler. If energy is fed into Port 1, it will flow out of Port 2; Port 4 is the coupled port and Port 3 is an isolation port. When the section is one quarterwave length long, the induced voltages will add at Port 4 to give the coupled port, but the equations show that at Port 3 the voltages cancel giving an isolation port. Coupling in the directional coupler is accomplished by perturbing the even allae odd mode of fields. Equation (2. 1) shows that the coupling coefficient ris proportioi al to the impedance discontinuities of either the even or the odd mode. oej +1 oej F Zooj - +1 00] oej+l oej oej+1 etl j Zeoej+i oej + 1 o ej $.eo+ 1+ oej + tI; mak9es no difference whether we talk of the even mode or odd mnode, since the couplers are designed to have a constant characteristic impe.dance, Z, a.ccordil'to equation (2, 2). From equation (2. 1) we see that a wave entering, from a 50 iS2'lin. into the coupler at Port 1 would be going from matched even and odd modt:iJmpe, c{, anee into the same characteristic impedance with a mode im.pedance nmism atch.:t even z \z z (-2) 6-1 Z oe o ( 00 mode impedance of tlhe utlarterwave line is higher than the feeod line so that fth l or the induced voltage at Port 4 will be a positive r while at IPort 2 the fields are going from a higher even mode impedance to a lower even mode impedance and will be negative as shown in Eq. (2. 3). Each of these induced voltages will cause a wave to travel to the right and to the left at Ports 3 and 4. Equation (2.3) shows the OF nt;ti-ffix$]Ca -E- - 8A

THE UNIVERSITY OF MICHIGAN 1084-2-Q r= oej+l -oej j oej + Z1 Zoe rj~o re-jO -rej0!4 31 output O0 output 2jrsin Z=Z ZZ - - ~, oe oo l l 2o - input lej0 output (1-2 r+r2 )ej0 FIG. 2-2: Single Section Quarterwave Directional Coupler

THE UNIVERSITY OF MICHIGAN 1084-2-Q magnitude and phase of these induced voltages. at Port 4 +7 e j, at Port 3 - e (2. 3) = electrical length of the stripline l? voltage coupling coefficient Equation (20 4) indicates the result of summing these two traveling voltages waves at Port 3. -+ j1 eJ "1eJo -0' ej= (2.4) The summation of the voltages at Port 3 causes cancellation of the voltage, therefore no energy will appear at this port. However, in equation (2. 5) the voltages add at Port 4. It is evident from equation (2. 5) that the voltages will add giving a maximum voltage at Port 4 when the line is X/4 long, i. e., the magnitude of the induced voltag is proportional to sin P where 0 is the electrical length of the line. Note that the voltage output from Port 4 is +2j l eCJO sin 0 and that of Port 2 is (1-2[ + 2 )e-j0 +1 e"Jo 7 20 e- 0 (+j e- ) 1 1 1-j (2.5) = +2j 1 e sin. (2.5 There is a net phase difference of 900 between these two as indicated by the j in the Port 4 term. This X/4 coupler gives the desired 3 db coupling and the 900 phase shift between the output ports, but has a limited range of 3 db coupling due to the sin 0 term which is proportional to frequency. When one wants to build a square wave with a Fourier series, it takes higher order terms to build up the flat response across the top and extend the width of the square wave. Frequency response of a directional coupler is broadbanded in much the same way by adding more terms or more sections to the primary center section.

THE UNIVERSITY OF MICHIGAN 1084-2-Q Figure 2-3 is a schematic of a three section directional coupler. Equation (2. 6) gives the voltage response at Port 3 and Port 4. Here again 2 at Port 4 is positive and; at Port 3 is negative for the same reason as discussed above for the single section coupler; at Port 4 + 2re-j0 at Port 3 2- j 0 (2.6) Equation (2. 7) shows the summation of these voltages waves at Port 3. As in the equations for the single quarterwave coupler, there is a cancellation of these voltage waves giving no power out of Port 3: j32 _ 0. (2.7) + 2 e-j30 -2 e-j30 = 0 (2.7 Equation (2. 8) illustrates the addition of the voltages at Port 4. The wave induced at Port 3 traveled an electrical distance of 30 from the reference at Port 1, and an additional 30 from Port 3 to Port 4 for a total distance of 60. A combination of these two waves, the one induced at Port 4 and one that traveled from Port 3 to Port 4, gives a term that is proportional to sin 30. From this, it is evident that each time two symmetrical sections are added to a directional coupler, the coupling is increased by adding odd harmonics in 0. In this way, the coupling will start with sin 0 for the central term, the next symmetric sections couple proportional to sin 3 0, the next sections couple proportional to sin 5 0, etc. ejO e 0 ejj3 (ej30 -j3) - 2j j30 (2 8) 2 2 2 2 sin Looking at equation (2, 8) it is also apparent that this three section coupler has the desirable quality that the two output ports are separated in phase by 900. By choosing the proper mode impedance mismatches, therefore giving the proper refle tion coefficients at these various junctions, it is possible to design uniform coupling across a broad frequency band in the same manner that a square wave is constructed with a Fourier series. Coupling between sections depends upon reflections established 11

TH1E UNIVERSITY OF MICHIGAN 1084-2-Q Coupled Portre e U efJe 2 eJ output t w 4 3 ou=tput 2j r sin b +2C2 sin 3j (*t 2r2 +4C1 E22 2 2 22 2 -j 30 FIGo 2-3~ Three Section Quarterwave Coupler 12

THE UNLVr R$ TY OF MCH~GAN - 10es4-2Q by impedance mismatches caused by the variation of strip width mand overlap of the coupled section. All calculations are based on the assumption of no higher order reflections, i.e., no mutua coupling between sections. For this reason it is imperative that all sections be desiged with a constant characteristic impedance. As in the case of Fourier series, the first term has the largest coefficient. The succeeding terms build uap the square wave response at the edges and reduce the ripple. For the 7 section 8l1 band 3 db directional coupler, the major contour is established by the center section and the first section outside the center section. The remaining terms are added to deerease the ripple. To cover the 5,l bandwidth we have chosen an 8,1 band directional coupler listed in the design tables by Mosko and Shelton. These design tables were derived by Mosco and Shelton using a computer iteration technique that would choose the values of the coefficients, P etco An error curve was obtained by summing the response from theIs in terms of 0 and comparing with a desired curve. Computer analysis of the error curve changed the Vs and the process was repeated until the response of the coupler was within the specified requirements in the computer program. These variables specified in the program are bandwidth, ripple, numer of sections allowed and toleranceo The ripple is the average deviation from the desired coupling, Of course, some peaks will be higher than this and the tolerance indicates how much higher than than the average these will beo A 3 db coupler is more difficult to build than those of lower coupling, as they require closer coupling and greater construction aceuracyo For this tighter coupling it is generally the accepted practice to design the coupler as two tandem -8, 3 db couplers. Two similar -8, 3 db couplers connected in tandem will display =3 db coupling over the bandwidth of the operation of the -8, 3 db couplers. The tandem desig allows a rarger separation between the ground planes and less coupling in the center sections. Inc6reased distance between the two ground planes minimizes variiation in the coupler due to mechanical deformnation in the two ground planes. ____ i13

1084-2-Q ourpH.ng t oleari ancees becomre qui te critical for broadband b k igh- vs l age ototul-pli g coeffienrtso Tlhe coupler is designed in terms of reflections frol the even anld odd rode impnedances. Knowing these reflections, one is able to estimnalte t:ie pr oper stri)p w.iib and strip overlap to give the 3 db coupling. Trhis strip width is a.ppro:.i=i mrate for several reasonso 1) the dielectric has variations ti dielectric const;ant that jo t~ecoie appa-ent a m:icrowave frequencies, 2) at rlicrowave frrc wua ies:j; the strip=-.ie teds o radiate so the walls of the stripline are lineid with screwvs foriiingg a -re;ctar.ngu.ar TWEl/i transmission line, 3) not only do thie screws in the si.deawitLs per- rb tarb Ihe waves but they also deform the ground pl.ane locally and possibly. coimrpress i1 the di.elect ric and 4) these theoretical eurves are for exact; tolerances which one is!i not able ito achhieve in practice, TVhe e:feeit of the scre-vvs uipon the inpijedance is o,slwnr by, a- plot of a 50 Q stripiline or a coiupler usirg the lir ie Iorlain lielctometer tX J 2 >lCh a plot re veals a. series of sn al nnl i u diseleoel nteit ilci sg tle uiie. A sirnilar plot of an air line wi ithe TDR systeemi s:ows te air.lfr; has a I-ear.y constant irippeidance level. UJpon investigating'this effeit, itt waras fottnd tihe Ci dscortinuiilities dueo to the screws' cotld be increased or derdease d a. ei:o' diTng to the te onsion on the screws,, AnotLher effect to be st udied is the fl'fetli; na tt. e - inps dace ea'used kby varyinjg the dista nee be-tweeri tie s evs and twLe e e'ht'ente' c'idrdrc. The'esif'n of tie coutpter is based u.pon fnaintair.ling a charna(-:-e'ris'ic 5{ { oip..mp oec at | cj.-,et section -firroghou'; tihe enli:iore coupler, and any undei s:-al..- t' r'tis al chlange the broadband respornse, Anothe: very iry:,"sortart aspect of the corupler do:'sigee is thle hrea.rie-re1nt f-or I.roet~'iClo; T i ~rEee conlectors" Sttip.in:o'nn.ectors s houild:,ha've a vei:Jry t'.'rV?'Si WRl,:If' not t'hey will act,as aon imnpedance mrlismatch producing li.ia',nte: rlect ilon, aly work on the coupler used -modified Type N T]JG/58 —A Dpanel co rtectors. rhie large; co.enate-r pr.in mt[ due('apacitane i ntCo;he junetio sn ith1 t- e sr ii.P inske masain:e, i.ig erYerI o i the problIe-ms in the stripline itself, These reflec;tiions vwere reduaced b-y Io.:tl nging

THE UNIVERSITY OF MICHIGAN 1084-2-Q to Omni-Spectra"': (OSM) stripline connectors. Mosko and Shelton have a very good theoretical representation of the problem of stripline coupler design. However, they do not give a systematic technique for correction of fabrication errors that result as noted above. It is most important to recognize the type of error in the coupler and source of this error. At the present time we know how to correct for these errors and this will be illustrated in Table 2-1. It is noticed there that for the first try we were wanting a 17 of. 559 and that we had an actual of.422. In the second try, the actual F's of the first attempt were determined by taking four points in 0, i. e., frequency, and measuring the coupling at these points. Given the four coupling values and with four unknown coupling coefficients it was possible to determine the individual coupling coefficients that were actually achieved. However, by using only four points any experimental error would also be calculated. To obtain a more accurate representation of the actual coupling, a large number of coupling values are measured as a function of frequency and by this over-determination of the coupling data, the best or most probable coupling coefficients are calculated by the computer. For the second try, a linear error analysis was performed and determined that1 should be. 700 on the design curve to obtain a. 559 coupling. This represented 100 per cent overlap of the center section and consequently the maximum coupling that could be achieved for this ratio of center conductor separation to distance between ground planes, s/b = 1/9. Notice the effect of coupling tolerances on the 4th try, the couplingF, was again given for.700 with the same strip width and overlap but measured I was.519 in this case. When the maximum coupling is required small variations in the strip widths produce large variations in the impedance introducing undesired reflections in the line. A more critical factor is that two center sections must be fully overlapped. Errors of a few mills of the overlapping of the center conductors will produce a large variation in the coupling between the two lines. Shelton (1965) also indicates that a variation'- Trademark brand name...... 15

1084-2-Q of 1 mnil for the center conductor separation will produce 0, 14 db error in coupling. Figure 2-4 shows the individual sin O, sin 30, sin 50, and sin 70 components, their relative amplitudes and the total coupling desired across the frequency band. TABLE 2_ 1 First TL Design 23.559 21=.168 2.=058 2 r 1. i Actual 21 = 422 2.069 -.028 2 4 013 Second Tr Design 2 =.700 2 = 261 2 = 085 24.007 Actual 2K.501 2.249 2 036 029 j 4 ~I ~Fourth Try Design 2F.700 22.298 2 = 073 2 001 2 4 Actual 2 519 2 2 161 2 041. 24-.001 I The overall coupling of a 8.o 3 db represents a voltage coupling of 0. 385o By this,simrple graphical technique, we do predict, with reasonable accuracy, the ripple:iquoted by Shelton (1965), i.eo, approximately 0.2db. Figure 2.-5 is a measured;coupling curve for one -8, 3 db section of the tandem coupler and the second curve is the one produced from the summation of the calculatedF's that were achieved. In Fig. 2-5 we see the values that were obtained by calculating Vs from the experimental curves and then reconstructing this curve to show a slight error due to the experimnental error of the four values measuredo In general the overall curve is fairLy close and is good enough to give an indication of the coupling error due to _________j_ __ 16

.5 2 r sin- -'" J.4 /0000 ~Resultant Coupling Curve - 140e~~~~~~~~~~~~~~~~~~~~~~~~~P o 0 2~~ 0 2r sin5i 3' O X -fit-~~- &Cs -p - _ \___ I) 0 10 20 of30 _ 50 60 7 (Electrical Length Individual SectionsXI _ 00000,~ ~ ~ ~ ~~r sin 37 FIG. 2-4: Theoretical Coupling Curve for Seven Section Coupler. 2r =.478, 2r 129, 2r=.049, 2. = 013 I'23 40

1084-2-Q the improper impedance levels in the coupler. The major difficulty at this time is controlling in the coupling level at the extremes of the frequency band. Referring to Fig. 2-4, it becomes apparent that to increase the coupling at the low frequencies the Fl and 12 must be increasedo r2 cannot be increased alone since it bee comes negative at the center frequency, 1800 MHz, so that the net effect is to cause the coupling to become low at this frequencyo Changing the r1 and the rK will also change the ripple slightly so that the third and fourth coupling coefficient must also be changed to ensure the proper ripple limit across the band, Our past experience has shown that it is difficult to increase the K1 and r2 to improve low frequency coupling without overcoupling at the lower frequencies. This is due to the very tight tolerances that must be obtained in terms of strip width and the overlap for the center sectionso Normally the coupler center conductors are etched from double clad dielectric sheets by a process that requires an accurate negative. The original drawing is drawn oversize, generally 5 to 10o1 and this drawing is photo reduced to minimize errors in the original drawing. To reduce costs of the fabrication of trial m odels, the first models have been made by cutting the conducting center sections outt of i mil brass shimstocko The difficulty in cutting this brass shimstock within 1 mnil tolerance can be appreciated when you consider that the thin rmaterial will de= form to have a somewhat irregular edge. Another difficulty after the shirnstock is cut is the alignment of the two conductors to produce the proper overlap. Early attempts at etching the center conductors have not significantly im- 1 proved the tolerances. The material presently being used is Rexolene S which is an irradiated dielectric with a dielectric constant of 2, 32 The irradiating process lowers the loss tangent at the higher frequencies but this causes the material to buckle (Fig. 2-6) when etched making it quite difficult to hold. close tolerances. The,bucklled material has air gaps in the coupler even with screws around the edge. These a ir gaps have an adverse effect since the change in dielectilc constant affects the odd modes maore than the even nmodeso Rexolite with a dielectric constant of 2, 62 has._ ___. _ _ D o.__o ____,_....._.___=_ 18

Desitred Voltage CauPtng Level Predicted Coupling Curved UnE r-ze IL2Cy G o, tmientajly Determin 4's M ered CTag'C s e ~~~~~~~~~ 2~~~~~~~~~~~~~~ 0~ ~ ~ ~ ~ ~ ~~~~~~~~~~~~5 O 10 ~~20 __ u6 (Electriczl Len th Ofte ILodividUr% SI Otian FIG. 25or and CoPtg Curve F o ~~~~~~~~2r =0:441, 2r = 0050. 27= D0.0 2F =0 027 ]:Departmentally Determined rus. 1 2 1 403

THE UNIVERSITY OF MICHIGAN 1084-2-Q been ordered in an effort to overcome this problem. The phase shifter is more complex such that its operation is difficult to describe theoretically. The phase shifter is constructed by shorting the two center sections some distance from the input. This distance is an integral number of quarter wavelengths at the center frequency. By shorting the coupled sections, the reflectio coefficient is unity and the phase of these coefficients determines the phase shift characteristics of the component. The basic coupled section configuration for the phase shifter is similar to that of the- directional coupler. The mode impedance mismatches still maintain a constant characteristic impedance. The coupled section is terminated by a short between the two center conductors. This termination "appears" as a short circuit to the odd mode, whose major energy components are confined between the two center conductors. The termination "appears" to be an open circuit to the even mode impedance whose energy is primarily confined between the two center conductors and the groun plane. This oversimplified model can be used to predict the general operation of the phase shifter. Consider the energy in the odd mode. The normalized input impedance for a shorted transmission line is given by Z. = Z tanh 1y where -y = a + j3, (2. 9) Z = odd mode line impedance If the line is assumed lossless this becomes Zin = Z tanh (jft{) = j Z tan 0 since 0 = electrical length of line =Ol and tanh j 0 = j tan. (2.10) From the transmission line theory, the input reflection coefficient is given by Z. -1 in Z. + 1 (2.1)....._~~~~~ | 120

THE UNIVERSITY OF MICHIGAN 1084-2-Q b3 C) 21

THE UNIVERSITY OF MICHIGAN 1084-2-Q Substituting the value for Zin from equtaion (2. 9) the input reflection coefficient becomes - j Z tan - + 1 = amplitude e j. (2.12) The phase of the reflection coefficient (0) is -1 0 = -2 tan (Z tan p) (2.13) 0/2 = tan (Z tan 0) tan 0/2 = Z tan 0 0 = absolute phase through coupled section. Now if a new angle is defined pu = 0/2 - 0 and the tangent is taken of both sides the right hand side of the equation can be expanded in a double angle formulation. After some manipulation the equation reduces to Qi+ )sin 2 tan u _. (2.14) 1+ I+Z1 + cos 2 It is seen that the quantity 1+ Z is the magnitude of the reflection coefficient Fat the impedance step. =sin 2 tan +['cosn 2 (2.15) If the tangent is taken of both sides of the equation one obtains /uta 1+sicons 20 (2. 16) If the reflections and phase dispersion are small and second order effects are neglected the equation (2. 16) can be simplified. If the r' is assumed to be small the ratio is small and the tan of a small number is approximately the small number. 22 ___

THE UNIVERSITY OF MICHIGAN 1084-2-Q Further if is small, the cos 2 P term of the denominator can be neglected. Then p, can be expressed as p Fsin 2 (2.17) With additional sections the net phase shift p becomes u r sin sin42 + n2+ sin+ ( sin2.18) This net phase shift with several terms approaches a straight line slope with frequency. If a reference line with the same frequency slope is used as the other port any phase difference between the two lines is maintained across the linear phase shi region of the coupled region. 23

1084-2-Q III SURFACE REQUIREMENTS In the first quarterly (Ferris, et al 1967) it was reported that a hemispherical surface would be required to mount the antennas. Since then efforts have been expended to reduce the undesirable effects of (the natural earth) ground reflections. As a part of this investigation, consideration was given to the antenna system geometry. Although the ground reflections will be discussed in Section VI, one aspect will be presented here. Since we are performing a simple vectorial addition of the relative amplitudes of antennas and no phase information is involved, it is not mandatory for the antennas to be mounted on a hemispherical surface. The surface can be of any arbitrary shape as long as the 0 and P pointing direction is known for the normal of each antenna in the radiating system. An alternate configuration would consist of a planar surface with 8 antennas in the outer ring pointed near the horizon. On an inner ring there could be 8 antennas pointed at 450 in elevation, and a single antenna in the center pointed at the zenith. There may be some antenna blockage in the configuration. 24

THE UNIVERSITY OF MICHIGAN 1084-2-Q IV ELECTROMECHANICAL SWITCH The electromechanical switch has been designed with capacitive coupling to reduce the noise. generated by a rotating DC contact. The early model for this switc was a simple laboratory model that had linear motion only. At the present time an engineering model of the final configuration has been received from the shop and is being tested. This model (Fig. 4-1) is 10 inches in diameter with an inner rotating section (rotor) 9 inches in diameter. The 17 inputs are to be located onthe stationary section (stator) of the switch on a 3-1/4 inch radius separated by a cord distance of 1-1/4 inches. The rotating port is located on a 3-1/4 inch radius of the rotor, with the energy being extracted from a rotary joint located at the center of the rotor. A stripline is utilized'to transfer energy from the rotating port to the rotary joint at th center section shown in Fig. 4-1. In this figure, the top section is the rotor. The stripline has been removed and TNC connectors installed to aid in making experimental measurements of the individual VSWR values for the rotary joint and switch contacts. At present the engineering model has four input ports located a cord distance of 1-1/4 inches apart as they will be in the final model. Measurements made for these four ports have shown that the decoupling between adjacent ports have at least 20db isolation. The VSWR of individual parts of the switch have been measured without the stripline from 600 to 3000 MHz and exhibit VSWR's that range from 1.3:1 to 2.0:1 with respect to a 50 i2line. Data has been collected for various separations between the input and output ports and the best VSWR seems to occur when the spacing is between 0.003 and 0.004 inches which is felt to be an adequate separation for the 1000 rpm rotational rate. The capacitive coupled rotationg switch is difficult to design since the VSWR is related to: 1) size and geometry of the cavity behind the capacitive probe (Fig. 4-2), 2) the size and geometry of the capacitive probe, and 3) the spacing between coupling probes. 25

THE UNIVERSITY OF MICHIGAN 1084-2-Q.4 0.-4 0 26

THE UNIVERSITY OF MICHIGAN 1084-2-Q.4~... 2~~~~~~~~~~~~~~~78 0 0 ~0:::j:,i~~~~~~~~~~~~~~~~~~~~.-.p.4~~ 2 7''~::::::::::

THE UNIVERSITY OF MICHIGAN 1084-2-Q V COMPUTER AND DETECTION SYSTEM Pulsed signals present the most difficult problem in the detection of the various types of communication and radar signals. Analog-to-digital converters typically employ a sampling technique, with a standard sampling rate of 20 KHz/sec. With a 20 KHz sampling rate, there is a possibility that multiple radar pulses could be received in the switch aperture period and not be detected during the sampling intervals. More difficult to detect is a relatively long pulse from a rotating, long range, ground based radar system. If a small portion of this pulse appears in the switch aperture, it would appear to be a very narrow pulse appearing only once in the entire aperture interval. There are alternatives to reduce the possibility of undetected signals. One method of detecting such small pulse width signals in the aperture is to increase the sampling rate up to 500 KHz. Five hundred KHz would appear to be the upper limit for the sampling rate due to the cycling time of most small, general purpose computers. Any sampling rate higher than 500 KHz would not allow the computer to accept the data from the analog-to-digital converter. This does not. assure 100 per cent detection of the signal in the aperture, as there are regions during the conversion of the analog signal where there is no surveillance of the switch aperture. In this method, the sampler converts the voltage level to a digitalized form at all sample points and reads these into the computer, and the computer has to make a decision at the end of the interval as to the largest voltage. Having the computer search through the voltages at the end of the period is a waste of the computer time, as the computer is not only tied up during the entire switch aperture accepting the data from the- sampler, but must spend time after the switch aperture interval sorting information to find a single peak voltage. This method is an example of injudicious application of the computer since it becomes a data storage device with insufficient time allowed for real time computation. 28

THE UNIVERSITY OF MICHIGAN 1084-2-Q Another analog-to-digital conversion method is the simultaneous method described in the Digital Logic Handbook (1967) in which the voltage is measured by a comparison to a reference. This method involves the construction of several leveldetection circuits. Each circuit would be triggered if the amplitude of the voltage peaks in the input signal exceeded that particular level. At the end of the interval the voltage level of the desired peak is indicated by the highest triggered detector circuit and would be determined by the computer through an examination of a single output word. The output word would contain the unique bit position for each level. Thus, the highest bit position containing a bit would indicate the highest amplitude encountered in the interval. With this bit information the computer would correlate the actual voltage reading. It should be noted that this technique does not require an A to D converter or sample and hold amplifier. It employs a continuous surveillance technique in that the peak voltage appearing during the switch aperture would trigger the comparison circuitry to the closest circuit level of the peak voltage. The major problem of this technique is the cost associated with attaining accuracy throughout the whole range of voltages that may be expected from the incoming signals. The number of comparison circuits required is apparent when low level signals are considered. With ten detectors designed to cover a 5 volt range, the best accuracy that can be obtained in detecting the peak voltage is + 0. 5 volt. While this may be acceptable for an interval in which the peak is near the 5 volt level, it is inadequate when the peak voltage is near or below 1 volt. For the5 volt signal, the 0. 5 volt resolution represents a 10 per cent accuracy, but the accuracy of resolution is only about 50 percent for the 1 volt signal. To improve the low signal resolution perhaps a logarithmic increase in the number of level detectors is required in the low voltage region. The best method would appear to be a peak reading voltmeter, and this is the method that has been chosen. The peak reading voltmeter will accept all signals from 29

THE UNIVERSITY OF MICHIGAN 1084-2-Q 0 to 5 volts during the switch aperture and record the peak voltage during this time, provided the voltage spikes are at least 1 microsecond in width. The peak reading voltmeter has a field effect transistor (FET) with the control voltage being the input voltage from the receiver. The input voltage allows a certain amount of current through the FET. This current charges a capacitor in the circuit to the value of the peak voltage impressed upon the FET. The FET with its high source iwmpedance will effectively appear to be an infinite impedance during the short switch intervals,thereby assuring no significant discharge of the capacitor. The detection system with the peak reading voltmeter is controlled by three computer generated pulse outputs and an interval timer. The first pulse from the computer is used to re-zero the circuit and must. occur at least 50 microseconds before the beginning of the switch aperture. The second pulse output occurs at the beginning of the switch aperture and is used to direct the acceptance of the signal from the multiplexer into the input of the FET circuit. The peak reading voltmeter is switched "on" only after the VSWR of the electro mechanical switch is low to prevent transients from producing false readings in the switch aperture. The interval timer in the computer program produces a program interrupt every 100 microseconds and by counting the number of these interrupts the switch aperture interval can be measured to the nearest 100 microseconds. At the end of the switch aperture the computer issues a command with a third pulse to open the FET gate and initiate an analog-to-digital conversion cycle. The A to D converter takes a 200 nanosecond sample of the output of the FET gate and stores this reading in a sample and hold amplifier. The contents of this amplifier are then converted to the 12 bit binary value. The end of the conversion is indicated by a program interrupt request at which time the digital value is available for input to the computer. There are many advantages to using this type of signal detection. One of these is that during the entire aperture interval, the computer is free to be converting the data from the previous antenna into the rectangular coordinate vectors and summing them with the accumulated sum from all previous antennas in the cycle. 30

THE UNIVERSITY OF MICHIGAN 1084-2-Q There is a finite lag time in the analog-to-digital conversion after signals have been received and recorded by the peak reading voltmeter. This time occurs during the time interval that lapses between two adjacent antenna switch positions. Two important features associated with this system are: 1) the computer is free to perform calculations, and 2) continuous surveillance of the switch aperture ensures the interception of all signals, within the 1lp sec response time of the peak reading voltmeter. In choosing the A/D converter, one must take into account the probability of circulating currents. Circulating currents may provide either noise or erronous readings to the computer. A/D converters with transformer inputs convert the data without interference from such circulating currents. For this reason we have chose a Texas Instrument A/D converter with transformer inputs. This is an economically priced A/D converter with provisions for eliminating circulating currents. The main function to be performed by the computer is the calculation and display of the azimuth and elevation angular location of the received signal measured from the 17 readings produced by the peak detection system. Although it is expected that most trigonometric functions required can be stored in table form, there are sufficient other calculations to warrant the addition of the multiply and divide hardware to the basic configuration. The addition of the multiply and divide hardware, of course, will increase the speed of the multiplication and division, and also reduce the size of the memory required. Therefore, the 4096 words of memory on the small computer are felt to be adequate. The hardware multiply and divide has a cycle time of approximately 2 microseconds per operation as compared with up to 300 microseconds for complicated multiplication or division in the computer without this option. Several computer types were considered, as to their ability to perform the task required for this project. Three computers felt to be, competitive were the PDP-8, the Varian 620I, and the Hewlett-Packard 2115A. Of the three, the PDP-8 has been exposed to the most years of service in the field. However, it has several.... 31

THE UNIVERSITY OF MICHIGAN 1084-2-Q limitations which would make it less desirable for-the DF problem. These limitations are: 1) its 12 bit word size limits single precision accuracy to 3 digits, 2) its weak command structure of the only 5 standard memory reference instructions, and 3) its absence of hardware index registers substantially increases the time and cost of programming as well as increasing the size of programs. If double precision calculations are required, to achieve the desired accuracy, the size of the program becomes even greater, possibly to the extent that the 4096 word memory would be inadequate for the DF problem. The Hewlett-Packard 2115A is an integrated circuit computer having the following features: it has a greater variety of basic commands, up to 6410 slots in addition to being an extremely versatile computer. The Varian 620I computer has comparable calculation capacity of the Hewlett-Packard 2115A at a substantial reduction in cost. The 16 bit word size of the 620I allows for single precision accuracy of 5 digits and contains 32 standard memory reference instructions, thereby reducing the size of the programs. It has hardware index registers that allow direct addressing of all memory as well as facilitating table look up operations. The cost of the Varian 620I ($18, 000. 00) with hardware multiply divide is comparable to the PDP-8 without.this feature. Employing the peak reading voltmeter, which will permit adequate computation time, it is felt the calculations required for the direction finder will be no problem for the Varian 620I with the hardware multiply divide option. At present the output display is envisioned to consist of an Industrial Electronic Engineers (IEE) unit which employs a 2 digit readout for elevation and a 3 digit readout for azimuth. The control for each unit is provided by two unique output transmission instructions. When executed, the instructions for the two digit display initiates the transfer of the low order 8 bit binary coded decimal (BCD) coded through software, of the accumulator or specified word of core memory to the IEE display interface. The instruction for the 3 digit display performs the same function --.. 32

THE UNIVERSITY OF MICHIGAN 1084-2-Q when executed except that the low order 12 bits (BCD coded through software) of the accumulator or specified word of the core memory and transferred to the digital display. The advantage of this type readout over Nixie tubes is that for each number there is a single light source that illuminates the proper number behind the receivi window. The lens control of the light gives a single readout at the front face. If the light burns out it is replaced by simply inserting a new bulb rather than by replacing the entire assembly as the case of a Nixie tube. Cost of the two types of display, the IEE and the Nixie tube are comparable therefore, the IEE display was chosen for its dependability and ease of maintenance.... 33

THE UNIVERSITY OF MICHIGAN 1084-2-Q VI GROUND REFLECTIONS The most difficult problem associated with the azimuth-elevation direction finder is the ground reflections from the broad beamwidth antennas. Antenna patterns have been calculated assuming the DF antenna is in the presence of the earth. The earth is modeled as a uniform, infinitely thick, perfect conductor. Predicted power patterns for antennas above a perfectly conducting earth have been computed by Jones (1948). An antenna with horizontal polarization is oriented with the maximum toward the horizon. These patterns are felt to be a worst case in that they do not take into consideration non-uniformities such as hills, trees, etc. to disperse the reflections so that reflections add to cause deep nulls in the antenna patterns. When the height of the irregularities is greater than the space wavelength (Rayleigh Criterion), the scattered field may come from an extensive region of the surface of the earth, Beckmann (1963). This causes a random scattering with accompanying loss of phase coherence in the scattered wave. As the antenna is raised above the ground, the number of nulls increases in the pattern due to the separation of the antenna and its image. However, the Fresnel region of reflection is farther removed from the antenna, becomes larger, and is more influenced by variations inr the terrain, further reducing the effect of the scattered wave upon the free space antenna pattern. There has been a considerable investigation of the affect of the terrain, but it does not appear practical to attempt to calculate the effect of terrain upon the antenna pattern. We have discussed the problem of scattering by a rough earth with several outstanding authorities in the field on this subject. All were in agreement that there would be considerable effect from the irregular terrain in the 600-3000 MHz region, but felt the extent of the effect could not be predicted with a high degree of accuracy employing theoretical methods now available. Based on the work of Beckmann and Spizzichino (1963), it becomes apparent that the patterns cannot be predicted, i. e., the nulls cannot be predicted within 10 db of the actual value. 34

THE UNIVERSITY OF MICHIGAN 1084-2-Q It has been recommended that a practical solution is to build the direction finder and measure the results experimentally. This becomes a difficult task within the budget allowed for the DF contract. To really findthe effect of ground reflections on the DF system, an rf source is required at a sufficient distance to establish ground reflections and this should have the ability to change its elevation angle. The only apparent solution would be to employ an aircraft with a signal source and have the azimuth-elevation direction finder plot the azimuth and elevation direction for this signal as a function of its known position. Its known position could be revealed by a radar system such as the NIKE site which is presently available for use at The University of Michigan. 35

THE UNIVERSITY OF MICHIGAN 1084-2-Q VII CONCLUSIONS During this reporting period emphasis was directed toward developing the antenna feed network required for the quadrifilar log conical antenna. This time was spent gaining a better understanding of the theory and practical characteristics of stripline couplers and phase shifters. We have developed the correction techniques necessary to perform a design iteration to achieve the desired operating characteristics. At present we are waiting on the shipment of stripline material to be employed in the construction of the final model of the stripline feed network. A problem remaining to be solved is a technique to exert a uniform pressure between the ground plane surfaces of the strip transmission line to prevent transmission line discontinuities. Discontinuities affect the phase shifter to a greater extent than the electrical characteristics of the coupler. An engineering version of the electromechanical switch has been electrically tested. These tests have shown that adjacent ports are adequately decoupled. However, additional effort is required to lower the VSWR in the 600 MHz region. As for the mechanical design the spindle and clutch for the switch are in the final design stage. The total computer package: 1) the computer, 2) A/D converter, 3) peak reading voltmeter, 4) digital display, and 5) all necessary interfacing has been ordered and is to be delivered within 120 days. In the selection of the computer package, a particular effort was made to 1) obtain an optimum A/D converter, 2) to minimize noise due to circulating currents between the A/D converter and computer, and 3) minimize overall cost while maintaining adequate performance level. An investigation has been initiated to determine what effect the natural earth will have on the operating characteristics of the DF antenna system. It is apparent the presence of the earth will affect the broad beam antenna patterns but the extent of the influence is not known. The greatest affect on the azimuth-elevation direction finder is reasoned to be on the accuracy of elevation predictions. It is to be noted 36

THE UNIVERSITY OF MICHIGAN 1084-2-Q that an exact solution to the reflection characteristics of the natural earth (including its natural roughness, trees, and variable dielectric constant) is impractical.Therefore, much of this study will require making intelligent assumptions as regards to analytical approximations of the true earth characteristics. 37

THE UNIVERSITY OF MICHIGAN 1084-2-Q REFERENCESBeckmann, P. and A. Spizzichino, (1963) The Scattering of Electromagnetic Waves From Rough Surfaces Volume 4, A Pergamon Press Book, The MacMillan Company, New York. No Author, Digital Logic Handbook, (1967), C-105. Ferris, J.E., B. L.J. Rao and W.E. Zimmerman, (1967) "Azimuth and Elevation Direction Finder Techniques", Quarterly Report No. 1, ECOM-00547-1, The University of Michigan Radiation Laboratory Report 1084-1-Q. Jones, E.A., (1948) "Model Techniques for Determination of the Characteristics o Low Frequency Antennas", Ohio State University Final Report, Project 247, Contract W 36-039-sc-32049. Shelton, J. P. (1965) "Tandem Couplers and Phase Shifters for Multi-Octave Bandwidths", Microwaves, pp. 14-19. Shelton, J. P. and J.A.'Mosko, (No Date) "Design Tables for Wideband EqualRipple TEM Directional Couplers and Fixed Phase Shifters", ADI Document 9017. Shiffman, B. M. (1958) "A New Class of Broadband Microwave 90 Degree Phase Shifters", IRE Trans. on Microwave Theory and Techniques, MTT-6, No. 2, pp. 323-327. 38

DISTRIBUTION LIST FOR REPORTS UNDER DAAB07-67-C0547 U of M Project 01084 Destination Number of Copies Technical Library, Rm. 3E-1039, Pentagon Dir., Defense, Research and Engineering Washington, DC 20301 1 Defense Intelligence Agency ATTN: DIARD Washington, DC 20301 1 Director, National Security Agency ATTN: C31 Ft. George G. Meade, MD 20755 20 Naval Ships Systems Command ATTN: Technical Library 20526 Main Navy Bldg., Rm. 1528 Washington, DC 20325 1 Dir., U. S. Naval Research Laboratory ATTN: Code 2027 Washington, DC 20390 1 Rome Air Development Center ATTN: EMTLD, Documents Ibrary Griffiss AFB, New York 13440 1 Electronic Systems Division, ESTI L. G. Hanscom Field Bedford, Mass. 07130 2 Hq, AFSC ATTN: SCTSE Bolling, AFB, DC 20332 1 CG, U. S. Army Materiel Command ATTN: R and D Directorate Washington, DC 20315 1 Redstone Scientific Information Center Attn: Chief, Document Section U. S. Army Missile Command Redstone Arsenal, Ala. 35809 1 CO, 52nd USASASOC Ft. Huachuca, Ariz. 85613 1 CO, Aberdeen Proving Ground Technical Library, Bldg. 313 Aberdeen Proving Groun, MD 21005 1

U of M Project 01084 Distribution List (continued) CG, U. S. Army Combat Developments Command CDCMR-E Ft. Belvoir, VA 22060 1 CO, U. S. Army Combat Developments Command Communications-Electronics Agency Ft. Monmouth, NJ 07703 1 CO, U. S. Army Security Agency Combat Developments Activity Arlington Hall Station Arlington, VA 22212 1 U. S. Army Security Agency OAC of IS, DEV (IARD-EW) Arlington Hall Station Arlington, VA 22212 3 U. S. Army Security Agency Processing Center IAVAPC- R and D Vint Hill Farms Station Warrenton, VA 22186 1 CO, U. S. Army Nuclear Defense Laboratory Attn: Library Edgewood Arsenal, MD 21010 1 Harry Diamond Laboratories Attn: Library Connecticut Ave and Van Ness St Washington, DC 20438 1 CG, U. S. Army Electronic Proving Ground Attn: Technical Information Center Ft. Huachuca, Ariz. 85613 1 Assistant Secretary of the Army R and D Department of The Army Attn: Deputy Assistant for Army R and D Washington, DC 20315 1 CO, U. S. Army Limited War Laboratory Aberdeen Proving Ground, MD 21005 1 CO, U. S. Foreign Science and Technology Center Attn: AMXST-RD-R, Munitions Bldg. Washington, DC 20315 1 Office, AC of S for Intelligence Department of the Army Attn: ACSI-DSRS Washington, DC 20310 1

U of M Project 01084 Distribution List (continued) CG, U. S. Army Electronics Command Attn: AMSEL-MR 225 South 18th Street Philadelphia, PA 19103 1 Director, Electronic Defense Laboratories Sylvania Electric Products Inc. ATTN: Documents Acquisition Librarian P.O. Box 205 Mountain View, California 94040 1 Chief, Intelligence Materiel Develbpmen t Office Electronic Warfare Lab., USAECOM Ft. Holabird, MD 21219 1 Chief, Missile Electronic Warfare Tech. Area EW Lab., USAECOM White Sands Missile Range, NM 88002 1 Chief, Willow Run Office CSTA Lab, USAECOM P. O. Box 618 Ann Arbor, MI 48107 1 HQ, U. S. Army Combat Developments Command Attn: CDCLN-EL Ft. Belvoir, VA. 22060 1 USAECOM Liaison Officer Aeronautical Systems, ASDL-9 Wright-Patterson AFB, Ohio 45433 1 USAECOM Liaison Office U. S. Army Electronic Proving Ground Ft. Huachuca, Ariz. 85613 1 CG, U. S. Army Electronics Command Ft. Monmouth, NJ 07703 ATTN: AMSEL-EW 1 AMSEL-IO-T 1 AMSEL-RD-MAT 1 AMSEL L-RD-LNA 1 AMSEL-RD-LNJ 1 AMSEL-XL-D 1 AMSEL-NL-D 1 AMSEL-HL-CT-D 2 AMSEL-WL-S 4

U of M Project 01084 Distribution List (continued) NASA Scientific and Technical Info. Facility Attnm Acquisistions Branch S-AK/DL P. O. Box 33 College Park, MD 20740 2 Battelle-Defender Info. Center Battelle Memorial Institute 505 King Avenue Columbus, 0. 43201 Remote Area Conflict Info. Center Battelle Memorial Institute 505 King Avenue Columbus, 0. 43201 1 Naval Communications Command ATTN: N2 5827 Columbia Pike Bailey's Crossroads, Virginia 22041 1 TOTAL 72

Security Classification DOCUMENT CONTROL DATA- R&D (Security classificatIon of title, body of abstract and indexing annotation must be entered when the overall.eport is classified) 1. ORIGINATING ACTIVITY (Corporate author) 2a. REPORT SECURITY CLASSIFICATION The University of Michigan Radiation Laboratory, Dept. UNCLASSIFIED of Electrical Engineering 201 Catherine Street 2b Ann Arbor, Miciigan 48068 3. REPORT TITLE AZIMUTH AND ELEVATION DIRECTION FINDER TECHNIQUES 4. DESCRIPTIVE NOTES (Type of report and inclusive date.) Second Quarterly Report 1 October - 31 December 1967 5. AUTHOR(eS) (Lest name, first name, initial) Joseph E. Ferris and Wiley E. Zimmerman 6. REPO RfT'AT'..7.. TOTAL NO. OQF AGE5 7b. NO. OF REF, March 1968 3 8 7 8a. CONTRACT OR GRANT NO..4. ORIGINATO'R;" FEPORT NUMBE'R(S) DAAB07-67-C0547 b. PROJECT NOa 1084-2-Q 5A6 79191 D902-05-11 c.' b. goTH' R R IPORT Ni(".) (,'n~yoth tnumbere. Tt mayR be asigned this report d. ECOM-0547-2 1 O. AVA IL ABILITY/.!JMITA?'ON NOTICES Each transmittal of this document outside the department of Defense must have prior approval of CG, U. S. Arnrrly Electronics Command, Fort Monmouth, New Jersey 07703 ATTN: AMSEL-WL-S t 1. SUPP.- EMENTARY NOTES.... g. POSOR Mi lITARy ACTIVITY U. S. Army Electronics Command AMSE L-WL-S Ft. Monrnmuth, New Tersey 07703 13. ABSTRACT This report discusses the construction and operation of the azimuth-elevation direction finder. It considers the proposed antennas and the antenna feed networks required. Emphasis is placed on the theoretical operation of the feed networks and the techniques being developed to achieve the design specifications. While the theory is well understood there are experimental design problems which are discussed along with their effect on the basic feed network. Several methods of signal detection are described with relative merit for each system presented. The report illustrates the function of the computer in the signal detection system and in the computation of azimuth and elevation angles of the monitored signal. Several small computers and analog-to-digital converters were considered during this period and the final choice is described. A brief survey of the effects of ground reflection on the operation of the system is given. However, due to the complexity of the problem no definite conclusions have been reached. D DJAN64 1473 -Security Classification

S.ecurity Classification KEY WORDS RO L E W T ROLE WT ROLE WT Azimuth Elevation Direction Finder Broadband Directional Coupler Broadband Phase Shifter a~~~~~~~~~~~~~~~~~~~~~~~~] w

UNIVERSITY OF MICHIGAN 3 9015 02826 7162