ENGINEERING RESEARCH INSTITUTE UNIVERSITY OF MICHIGAN ANN ARBOR MAGNETIC MODULATION DESIGN EMPLOYING MU SURFACES FOR FERRITES Technical Report No. 37 Electronic Defense Group Department of Electrical Engineering By: L. W. Qrr Approved by: a /; J. A. Boyd Project 2262 TASK ORDER NO. EDG-4 CONTRACT NO. DA-36-039 sc-63203 SIGNAL CORPS, DEPARTMENT OF THE ARMY DEPARTMENT OF ARMY PROJECT NO. 3-99-04-042 SIGNAL CORPS PROJECT NO. 194B PLACED BY: SIGNAL CORPS ENGINEERING LABORATORY, FORT MONMOUTH, NEW JERSEY September, 1954

TABLE OF CONTENTS Page LIST OF ILLUSTRATIONS iii ABSTRACT iv 1. INTRODUCTION 1 2. INCREMENTAL PERMEABILITY 2 3. BUTTERFLY LOOPS 6 4. LA -H LOOPS 6 5. 1WJ SURFACES 9 6. MAGNETIC MODULATOR DESIGN 6.1 Sinmple Balanced Modulator 14 6.2 Calculation of Transimpedance 18 6.3 Other Design Features 20 6.4 Application of Mu Surfaces 21 6.5 Practical Design of a High-Zt Modulator 22 7. CONCLUSIONS 23 APPENDIX A 24 BIBLIOGRAPHY DISTRIBUTION LIST 26 ii

LIST OF ILLUSTRATIONS Fig. No. Title Page 1 Definitions of Magnetic Parameters 4 2 Multiple Hysteresis Loops at Constant AB for Ferramic G 5 3 Butterfly Loop at Constant AB for Ferramic G 7 4 y -H Oscillograms for Ferramic G b 5 Variation of Incremental Permeability with Bias Field in Ferramic I 10 6 Mu Surface for Ferramic G 11 7 Mu Surface for Ferramic H 12 8 Mu Surface for Ferramic I 13 9 Simple Balanced Magnetic Modulator 15 10 Operating Characteristic of Modulator 17 111

ABSTRACT A practical design equation is derived for the transimpedance of a balanced magnetic modulator for small modulating current signals. Design is simplified by presenting incremental permeability as a three dimensional mu surface. Mu surfaces are shown for Ferramics G, H and I, and their application to magnetic modulator design is discussed. iv

ENGINEERING RESEARCH INSTITUTE * UNIVERSITY OF MICHIGAN MAGNETIC MODULATOR DESIGN EMPLOYING MU SURFACES FOR FERRITES 1. INTRODUCTION Magnetic modulation of an rf carrier is very useful in a variety of applications. For example, it may be used in low level dc amplification, as in the measurement of very low current signals from low resistance thermocouples. In this instance it serves the same purpose as a high frequency chopper, but Iwuith several advantages over other chopper techniques. Balanced magnetic modulation may be used to eliminate the audio amplifier in the modulator of rf transmission systems, such as in telemetering and in single sideband systems where the carrier is absent. A balanced magnetic modulator delivers an output voltage at the carrier frequency, the phase being determined by the polarity of the modulating signal, and the amplitude being proportional to the amplitude of the modulating signal. Since the input signal may be a varying dc current, the carrier may be at all times present. When an ac current signal is applied, the carrier is absent, and nly sidebands are present. By proper core selection and circuit design, a magetic modulator may be constructed with extremely small drift from the balance oint. The presence of a dc unbalancing current in the input signal introduces carrier voltage in the output for no ac signal. By adjusting the amount of inbalance, a normal modulation of any amount (up to 100l if desired) may be

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN - obtained for a given input signal level. The input circuit of the modulator may be wound for a wide range of impedance levels to match the modulating source. When the input impedance is higt a large inductance in the input circuit is generally implied; therefore, the upper limit of modulating frequencies may be restricted by circuit considerations. The design equations derived for the simple magnetic modulator apply equally well to any magnetic material. However, the specific data given are for ferrites, or ferromagnetic spinels which are particularly well suited to high frequency operation., Carrier frequencies up to 50 mc or more may be used with ferrite cores, -and satisfactory operation of a magnetic modulator at 10 me carrier frequency has been obtained with little difficulty. Modulating frequencies are generally limited to an order of magnitude lower than the carrier frequency. However, there may be additional limiting of tt bandwidth of the modulating signal because of the inductance of the signal winding The input to a magnetic modulator is generally a current source, while the output is a modulated carrier voltage. In the balanced condition the ratio oI output voltage (peak) to signal input current is called the transimpedance. An expression for the transimpedance of a simple magnetic modulator is derived. It will be shown that mu surfaces are extremely useful in magnetic odulator design. In order to introduce the subject of mu surfaces, some basic roperties of magnetic materials will first be considered. 2. INCREMENTAL PERMEABILITY The basic physical property employed in magnetic modulation is the variation of incremental permeability At with bias field Ho. Incremental

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN permeability is defined as the ratio of A B to AH, iilen a toroidal specimen is cycled around a minor hysteresis loop such as 12341 in Fig. 1. A H is the range Df excursion in applied magnetic field, while AB is the total excursion in flux density caused by the combined ac and dc fields. In Fig. 2 the dashed curve shows The position of the saturation B-H loop, or major hysteresis loop,for the material, md for any combination of ac and dc fields the specimen operates within this major loop. HaL, the slope of the chord of a minor loop is generally smaller;han the corresponding slope of the major loop at any particular value of Bo, but )therwise there is no particular relation between these slopes. The manner in which the incremental permeability varies as a specimen f ferrite is cycled around a major B-H loop is demonstrated by the oscillogram, ig. 2A. This shows a series of small excursions (minor loops) superposed on a jor B-H loop for a toroid sample of Ferramic G. This was obtained by a method which the driving magnetic field applied to the sample contains two frequency mponents. The two frequencies used here were 60 cycles and 3000 cycles, so that'e specimen performed 50 minor excursions for every excursion around the major op. To make the oscillogram clearer, half of the pattern (the negativeoing portion of the major loop) was blanked off by a synchronized blanking pulse,. he blanked protion would otherwise fall within the upper half of the major B-H oop (dashed line). Since the audio amplifier used for driving the core was essentially a onstant voltage generator, the size of AB for each small excursion was approxiately constant over the range of H, while the size of AH varied because of the riation of tz ~ Ferramic G is a high frequency ferrite, Body No. 254, General Ceramics and Steatite Corporation, Keasbey, New Jersey. L ~~~~~~~~~~~3

-l\ > ~..~~~l~~ \\ |I o I\:> > o i\ rl~. Ia II, —' I'C z0 Hn w I N b

:I''':" ":'"' / Bo BI - 19..52 11320. 500 | asB= 270 GAUSS / I I"..21 | 5401 880 |.035 -810 950 /=-. ---- J -___..20 - 1750 200. ~, I/ I: l I A. ANALYOSCILLOGRAM MULTIPLE HYSTERESIS LOOPS AT CONSTANT AB FOR FERRAMIC GUSS C~B~270 GAUSS 21' F 540 880 ~5. t.'.,";:".:. "!?:'...'';"''"':""'"'

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN - Fig. 2B shows the values Ho, Bo and Lzi for several of the minor loops in the oscillogram. Since these values were obtained by scaling, the data are somewhat inaccurate. The data are plotted in Fig. 3, and the curve thus obtained is a tz -H curve, frequently termed a Butterfly Loop. 3. BUTTERFLY LOOPS To obtain a continuous plot of the variation of tLA with H, which gives greater accuracy than the previous method, the envelope of a radio frequency or high audio frequency wave, derived from the incremental flux excursions, is displayed as the abscissa with an ordinate proportional to H. A typical Butterfly 0 Loop obtained in this manner2 (in this case at constant A H) is shown in Fig. 4A. If the variations in Ho are confined to positive values only, a "half butterfly loop", or one-way /A.-H loop, is generated, as shown in Fig. hB. 4. /LA -H LOOPS As. will be shown later, the mode of operation in magnetic modulation is such that the total magnetic field (the sum of the modulating signal, dc bias, and rf driving fields) does not change sign. For this reason, we will consider only the one-way uAd -H loop. For semiquantitative work, the data may be obtained as in Fig. 4B. For more accurate work, the variations in IzA must be found from 1 The term Butterfly Loop is more usually applied to a plot of the variation of th/ when.iH is held constant. L. W. Orr, "Permeability Measurements in Magnetic Ferrites" EDG Technical Report No. 9, University of- Michigan, September, 1952.

,gS-~-6 aa00 L8-b9-V 9ZZ 1500 AB: 270 GAUSS 1200 DATA FROM FIG.2 900 600 op,~Ho 12'9 6.. ~.9 1.2 FIG. 3 BUTTERFLY LOOP AT CONSTANT AB FOR FERRAMIC G 3 0 0,..,.~~~~~~~

A. BUTTERFIY I ooP B. ONE —WAY,m-tH LOOP FIG. 4 p-H OSOILLOGRAMS FOR FERRAMIC G

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN pointwise data using impedance measurements or a resonance method. This variation of tLA for a typical ferrite is shown by the one-way tza -H curves in Fig. 5. It is seen that each curve is double-valued, the higher value of permeability being obtained for Ho increasing. Curve B indicates less difference between the permeability for rising and falling Ho than does curve A. This serves to illustrate the fact that this discrepancy decreases as A H is increased. However, it is clear that there is also a variation in 11/ at constant Ho as A H is varied. In this case, the value of Ft increases for curve B where AH is larger. In many instances, such as in the design of a magnetic modulator, it is important to examine the manner in which,LA varies when both Ho and A H are variable. A display of such data may be accomplished by plotting a family of curves, such as A and B in Fig. 5, but a much clearer presentation is a threedimension plot, or mu surface. 5. MU SURFACES By plotting,tLA as a function of both H and A H in three dimensions, mu surface1 is obtained. Isometric projections of three mu surfaces are shown Ln Figs. 6, 7 and 5. These surfaces were obtained from Type F109 toroid samples )f Ferramic bodies G-254, H-419 and I-141. The permeability was measured at 25~ C Lsing a frequency of 10 kc. Each surface was obtained by taking a series of oscil-.ograms similar to Fig. 4B. From these the surface contours were plotted in isoetric projection. The curve for A H = 0 (dotted line, Figs. 5, 6 and 7) was obained by extrapolating the curves of constant H back to the plane AH = O. The L. W. Orr, "Permeability Measurements in Magnetic Ferrites" EDG Technical Report o. 9, University of Michigan, September, 1952.._, _,_ _._ _ _._._.,,_.,_. 9

3000 FERRAMIC I A - AH =.12 OERSTED B - AH.50 OERSTED 2000 l 1000 0 0.2 0.4 0.6 0.8 1.0 Ho BIAS FIELD - OERSTEDS FIG. 5 VARIATION OF INCREMENTAL PERMEABILITY WITH BIAS FIELD IN FERRAMIC I. 10

Il 9 OIW.13- SOi 3OV1-3nS nw 9'9149. o~~ OZ6Z = XoW o

, max 3700 zoo0 420 6l0 0ooo 600 FIG. 7

~1 I OIVW8t3A 80AO 3OVflnS ns 8'91J 469;. 0~0~ %%%%.0_ %%0~]...~~~~~~~~~ =t, ) ~S

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN intercept of this A/ H = 0 curve at the origin is the value oL to distinguish it frdm the true initial permeability,,uo. The values of ALo and / max are given in each figure. Since it is seen in Fig. 5 that each /4A -Ho curve is double valued, the mu surface is really a double surface. The upper branch of this surface is for Ho increasing, while the lower branch is for Ho decreasing. To avoid confusion, only the upper mu surface is plotted in Figs. 6, 7 and 8. The spacing between the two branches of the surface becomes quite small for large values of Ho and AH, the trend being indicated in curve B of Fig. 5. In most regions of the surface, the slope d, LA /dHo is very nearly the same for the upper and lower branches. This is an important point, since the derivative is the important factor in modulator design. Mu surfaces for various magnetic ferrites make it possible to solve readily several types of magnetic design problems, and to answer such questions as the following. What is the best material for the job? What is the operating point for the maximum sensitivity? What is the operating point for the minimum distortion at a specified signal level? The application of mu surfaces in magnetic modulator design will be discussed in detail in the next section. 6. MAGNETIC MODULATOR DESIGN 6.1 Simple Balanced Modulator One of the simplest forms of balanced magnetic modulators is shown in Fig. 9. It consists of two similar ferrite toroid cores (a and b, Fig. 9) of |niform rectangular cross section. These are wound with excitation windings l(a) nd l(b) of N1 turns each and connected in series. These windings are excited by

2g-IZ- S VY3r 8Q-tS-V 0L6-W (o Io Io +I, SINwt (b) N 2 I b s --- E2 ( e, - 0 L C FIG. 9 SIMPLE BALANCED MAGNETIC MODULATOR. 15

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN rf and do current generators giving a combined current of I1 sin cAt + Io amperes. A do signal current, Is, is fed through isolating inductor L, to the secondary windings 2(a) and 2(b) connected in series opposition, and when properly balanced, the output voltage eo is an rf carrier having its magnitude and phase related respectively to the size and polarity of the signal current Is. The operating characteristic of each core is shown in Fig. 10. When Is = 0 both cores operate at the point P by virtue of the do bias current Io, and there is zero output voltage. When a positive signal current flows, core "a" is made to operate at point A while core "b" is made to operate at point B (see Fig. 10). An expression for the peak amplitude Eo of the fundamental component of the output wave may be obtained by assuming a sinusoidal variation of flux density in the two cores. The resulting expression is N2 Ao(,L b - tL a) N1 I O-108 volts peak. (1) E = (1) 5r Where N1 and N2 are the number of primary and secondary turns on each core, A is the cross section area in square centimeters of each core, co is the angular frequency of exciting current of amplitude It, qLa and tLb are the incremental permeabilities at points A and B in Fig. 10, and r is the mean magnetic radius2 in centimeters of the toroids. This expression is derived in Appendix A. 2 The mean magnetic radius r of a toroid of uniform rectangular cross section and inner and outer radii r1 and r2 is given by r r rl r= 12 loge r2 -loge r1

~G-I;Z-G VJ3r 8'-t9-V OL6-1W AL'a AL P r- a — GA Hs Hs O Hb Ho Ha H FIG. 10 OPERATING CHARACTERISTIC OF MODULATOR. 17

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN 6.2 Calculation of Transijmpedance Over a small range in H, in which the operating characteristic may be considered linear, we may write (Lb -.La) 2Hs i d H0 (2) where the derivative is taken at the point P, and Hs is the small additional field derived from the signal current I as in Fig. 10. Noting also that N2 I H 2 s (3) s 5r we may substitute (2) and (3) in (1) to obtain 2 A i N I N2 I dA -8 E y10 volts (4) 2 In most applications, the input exciting voltage E1 is restricted to practical limits by oscillator design considerations. It is convenient to introduce a design peak value of E1 such that E = E +E E1 Ela 2a = 2c L1 I1 volts peak. 2 Ag. N -8 where A FLP 1 10 hen., for I O. Sr 2 pcN I -8 So E = 1 10 volts peak. (5) 5r By substitution in Eq. 4, we have ~~~~E.1 N ~~dI Kvolts peak. (6) 5rl A N1 2 s d H 18

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN'where 1L and ( are the values of incremental permeability and its d Ho derivative taken at the operating point P. The transimpedance, Zt, of the device is the ratio of change in peak output voltage E to a change in signal current Is. Thus, Eo E1 2 ( dH ) Z- t N (7) Is 5r1 ~Z^p d H 0 P This equation illustrates the following factors in modulator design: 1. The transimpedance increases with E1. 2. The transimpedance decreases as 1,A increases. This is caused by the fact that the inductive reactance of the driving windings increases with t P thus reducing the exciting current I. 3. The transimpedance increases with N22. Since the inductance 2 L2 of the output windings also increases with N2, E will vary directly with L2. 4. The transimpedance varies with the derivative (dtZA /dHo). This is one of the critical. variables and must be selected by a proper choice of magnetic material and operating conditions. For the maximum sensitivity (highest transimpedance), this derivative must be as large as possible. On the other hand, if minimum distortion is required, the operating point should be chosen for little or no variation in (dLza /dHo) over the desired range of operation. The next section slows how mu surfaces may be used to facilitate the selection of a suitable material and operating ___________________________9,_, 9

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN point with regard to (d UZ /dHo). 5. The transimpedance is unaffected by varying the frequency. This follows since c is absent from Eq. 6. It is often believed that a higher transimpedance is obtained by operating at a higher carrier frequency, and Eq. 4 would lead to this conclusion. However, it should be pointed out that as the frequency is raised, the inductive reactance L1 of the driving windings increases. This causes a reduction in the driving current I1 with the result that the product coI1 in Eq. 4 remains constant as co is varied. Therefore, the transimpedance is invariant with frequency. 6.3 Other Design Features Although the transimpedance is generally the most important factor in modulator design, several other factors must be considered. For example, the core material must be chosen to have a relatively low total loss at the carrier frequency, otherwise additional driving power and overheating of the cores may become a problem. In addition, the turns on the signal winding, N2, may not always be made as large as desired because of excessive inductance in the input circuit. The choice of N2 and of the isolating inductance L, thus depend to a large extent upon the bandwidth of the modulating signal to be used. A further restriction on N2 may come from the output circuit, where particular requirements are needed. If the output is to drive a vacuum tube, it must work into the capacitive input of the tube. It may be occasionally desirable to resonate the output inductance with this capacitance to obtain the argest possible output voltage. In general this is avoided however, so that a ritical adjustment is not required, and so 20.

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN that small fluctuations in output inductance, driving frequency, or stray capacitance do not affect the operation of the modulator. Certain assumptions were made in deriving the transimpedance which may not be precisely true. For example, it was assumed that the incremental permeability was a single valued function of Ho although it is shown in Fig. 5 to be double valued. A first consequence of this is that the application and removal of a rather large pulse of signal current to a balanced modulator will slightly upset the balance giving a small amount of carrier for no signal. A second consequence is that there will generally be a small phase shift between the input current waveform and the envelope of the modulator output. However, these effects are generally small enough to be ignored. It was also assumed that the rf excitation current was sinusoidal. However, slight variations of the total impedance of the driving windings, and the nonlinearity of the core material may cause the excitation current to be modulated by the signal current, and to contain harmonics of the carrier frequency. It may be necessary in some applications to remove the harmonic content1 of the modulator output by means of a low-pass filter. 6.h Application of Mu Surfaces As noted above, the derivative (dAt /dHo) is the important factor in modulator design. A proper choice of magnetic material,and operating conditions may be directly obtained from mu surfaces. For maximum small signal transimpedance, the material having the largest derivative is selected. By examining Figs. 6, 7 and 8 it is found that Ferramic I In another type of magnetic modulator called a magnettor, the carrier is filtered out and only the second harmonic is used as the modulator output. This is described in Article 8 of the Bibliography. 21

ENGINEERING RESEARCH INSTITUTE * UNIVERSITY OF MICHIGAN will give the greatest Zt. Although this material has a maximum permeability (a m = 2900) smaller than others, such as Ferramic H, its maximum derivative [(d/ /dH) = 12,700 per oersted) is largest. The point on the mu surface where (dLtA /dHo)max is found also gives the correct operating point for each core of the modulator. In this case the cores should operate at H = 0.05 oe., and AH = 0.37 oe. 6.5 Practical Design of a High-Zt Modulator As an example of the method of using the design equations and mu surfaces, the design values for a typical high sensitivity modulator are given. It is assumed that the F-109 core size is satisfactory, and this determines the values of r and A. Ferramic I is chosen as the best material of the three for which mu surfaces are shown in this report. This determines the values of (d/A /dHo) and /zA, found from the mu surface as described above, and fixes the operating point at H = 0.05 oe, and AH = 0.37 oe., ihich determines 0 the size of Io and I1 once the excitation windings (N1) are designed. In the design given, N1 was chosen as 50 turns, and N2 was chosen as 200 turns. The choice of N1 is not arbitrary, but is chosen to work with a particular driving frequency. In this case, however, we shall choose N1 arbitrarily, and calculate the required driving frequency. It is assumed that rf power is available at 100 volts peak. These design values are tabulated as follows: r = 0.9 cm d = 12.7 x 103 dH A = 0.1 cm N1 = 50 turns = 2 x 103 N = 200 turns H = 0.05 Oe 2 o El - 100 v. peak AH = 0.37 Oe 22

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN To satisfy the operating point conditions, the required values of Io and I1 are found. o 5rHo/N1 = 4.5 ma. I1 5r(1/2AH)/N1 = 16.2 ma. (peak) Since we do not have arbitrary control of the current I1, the frequency must be adjusted so that the proper value of I1 will flow. By rearranging Eq. 5 for frequency determination, we have oX 5 x 10' E r f 1 h4kc. 2n 4- I= h Ap A N1 1 p Finally,the transimpedance is found from Eq. 7, and has the value Zt = 0.112 x 10Y volts per ampere. 7. CONCLUSIONS The design of magnetic modulators is greatly simplified by the use of mu duo surfaces. Although the derivation is of primary importance in modulator design, it was noted that other factors must be considered. Although not described in detail in this report, the mu surface may be applied to the design of many types of magnetic modulators and other magnetic devices (e.g., magnetic tuning of rf circuits) whose operation depends on the variation of incremental permeability.

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN APPENDIX A Derivation of expression for peak amplitude Eo of fundamental component of output voltage. (Eq. 1). Assumptions 1. The r-f flux varies sinusoidally in both cores. 2. The operating characteristic (Fig. 10) is single-valued. 3. The isolating inductance L produces no loading upon the output. According to assumption 1, the flux density B = B1 sin wt so that dB= co B cos ot dt 1 In each core, the voltage e2 in the secondary is -8 e = N 10 8 volts 2 2 dt = N2 A o B cos cot 108 volts 2 1 = 2 A ca H cos ct 10 volts N1 I1 ButH1 5r Oe. Thus, the peak value of voltage across winding 2a is N2 AX L a N1 I1 101 3~~2a =.......volts 5r A similar expression may be written for E2b. These two voltages are 1800 out of phase so that the peak output voltage Eo is their difference, thus N2 & (tHa ctL b) N1 I1 -8 E =.......... 10'volts.

Bibliography 1. E. Peterson, J. M. Manley, and L. R. Wrathall, "Magnetic Generation of a Group of Harmonics." B.S.T.J., 16, 437, October 1937. 2. A. S. Fitzgerald, "Magnetic Amplifier Circuits." Jour. Franklin Inst., 2h4, p. 249, October 1947. 3. A. U. Lamm, "Some Fundamentals of a Theory of the Transductor or Magnetic Amplifier."' Trans. AIEE, 66, p. 1078, 1947. 4. L. W. Buechler, "Magnetic Amplifiers for Shipboard Applications." Elec. Engr., 68, p. 33, January 1949. 5. R. E. Morgan, "The Amplistat —A Magnetic Amplifier." Elec. Engr., 68, p. 663, August 19h9. 6. Gunnar Wennerberg, "A Simple Magnetic Modulator for Conversion of Millivolt D-C Signals." Elec. Engr., 70, p. 144, February 1941. 7. J. M. Manley, "Some General Properties of Magnetic Amplifiers." Proc. IRE, 39, p. 242, March 1951. 8. E. P. Felch, V. E. Legg and F. G. Merrill, "Magnetic Modulators." Electronics, 25, p. 113, February 1952. 25

DISTRIBUTION LIST 1 copy Director, Electronic Research Laboratory Stanford University Stanford, California Attn: Dean Fred Terman 1 copy Chief, Electronic Warfare Department Army Electronic Proving Ground Fort Huachuca, Arizona 1 copy Chief, Engineering and Technical Division Office of the Chief Signal Officer Department of the Army Washington 25, D. C. Attn: SIGJM 1 copy Chief, Plans and Operations Division Office of the Chief Signal Officer Washington 25, D. C. Attn: SIGOP-5 1 copy Countermeasures Laboratory Gilfillan Brothers, Inc. 1815 Venice Blvd. Los Angeles 6, California 1 copy Commanding Officer White Sands Signal Corps Agency White Sands Proving Ground Las Cruces, New Mexico Attn:- SIGWS-CM 1 copy Commanding Officer Signal Corps Electronics Research Unit 9560th TSU Mountain View, California 75 copies Transportation Officer, SCEL Evans Signal Laboratory Building No. 42, Belmar, New Jersey FOR - SCEL Accountable Officer Inspect at Destination File No. 22824-PH-54-91(1701) 26

1 copy H. W. Welch, Jr. Engineering Research Institute University of Michigan Ann Arbor, Michigan 1 copy Document Room Willow Run Research Center University of Michigan Willow Run, Michigan 11 copies Electronic Defense Group Project File University of Michigan Ann Arbor, Michigan 1 copy Engineering Research Institute Project File University of Michigan Ann Arbor, Michigan 27

3t90ll511." O3095.43. "5"0095 0433,

]LIATA'a,-',?:C 4.'Y tg 2 - e ho uli% be Fig t F %e. o, Se@tioIn 7 tinemS 2..........':auld be i "a~~ 2~ GsC.' " -f3 d