ENGINEERING RESEARCH INSTITUTE UNIVERSITY OF MICHIGAN ANN ARBOR FERROMAGNIETIC AND FERROELECTRIC TUNING Technical Report No. 2 Electronic Defense Group Department of Electrical Engineering By: L. W. Orr Approved by: B.W (I H. W. Welch, Jr. Project 2262 TASK ORDER NO. EDG-4 CONTRACT NO. DA-36-039 sc-63203 SIGNAL CORPS, DEPARTMENT OF THE ARPY DEPARTMENT OF ARMY PROJECT NO. 3-99-04-042 SIGNAL CORPS PROJECT NO. 194B PLACED BY: SIGNAL CORPS ENGINEERING LABORATORY, FORT MONMOUTH, NEW JERSEY July, 1954

TABLE OF CONTENTS Page LIST OF FIGURES iii ABSTRACT iv FOREWORD 1 1. GENERAL REMARKS 2 1.1 Electronic Tuning 2 1.2 Magnetic Tuning 2 1.3 Electric Tuning 2 2. MATERIAL PROPERTIES 3 2.1 Properties of Ferrites 3 2.2 Properties of Titanate Ceramics 5 2.3 Comparison of Properties 7 3. COMPARISON OF MAGNETIC AND ELECTRIC TUNING 12 3.1 Manufacturing Problems 12 3.2 Costs 13 3.3 Temperature Stability 14 3.4 Time Relaxation Problem 16 3.5 Frequency Range and Tuning Ratio 19 3.6 Scanning Rates 22 3.7 Very Rapid Scan Systems 25 4. CONCLUSIONS 26 REFERENCES 27 DISTRIBUTION LIST 28 ii

LIST OF ILLUSTRATIONS Fig. No. Title Page 1 Definition of Magnetic Parameters 4 2 iMu Surface for Ferramic H. 6 3 Epsilon Surface for Aerovox Hi-Q 40 at 26~C 8 4 Epsilon - Temperature Surface for Aerovox Hi-Q 41 9 5 Comparison of Properties 10 6 Increductor Tuning Temperature Curves 15 7 Aerovox Hi-Q 41 Tuning Temperature Curves 17 8 Tuning Units 23 iii

ABSTRACT Ferromagnetic and ferroelectric tuning are compared on the basis of presently known materials and techniques. Conclusions are based on present knowledge in two rapidly growing fields. In general, magnetic tuning appears more satisfactory on most counts below 50 me, but at higher frequencies electric tuning has certain advantages. Such aspects as frequency stability, manufacturing problems and cost are considered in detail. iv

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN FERROMAGNETIC AND FERROELECTRIC TUNING Foreword This report attempts to make a critical comparison between methods of ferromagnetic and ferroelectric tuning. The comparison is based on present knowledge in two rapidly developing fields, and because it is impossible to predict properties of materials now in development, the conclusions drawn here must be regarded as subject to modification as new materials are developed. At the present time, magnetic tuning appears generally more promising than ferroelectric tuning, except for certain features which are discussed in detail, but it must be recalled that magnetic tuning has been in development for five years or more and is consequently more advanced. The ideal tuning element may be thought of as having such low loss that this may be neglected. Its tuning capabilities should extend to 1000 megacycles or better with little change in Q. It should have a zero temperature coeffecient of frequency from -50~ C to 1000 C. It should be small, compact and capable of large scale manufacture with only a small fraction of rejects. It should scan a frequency ratio of 10:1 or better with relatively small control power, and this requency ratio should not be reduced at the higher frequencies. The control power for rapid scan (up to 50 megacycles per microsecond) should not become excessive. As usual the ideal does not exist, and it is highly unlikely that it willbe developed within the next few years. Therefore the present choice for a particular tuning method and tuning material must be a compromise to obtain a

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN close approach to the ideal in those properties most important for a specific application. For the purpose of this report "ferromagnetic tuning" will be abbreviated to "magnetic tuning," and "ferroelectric tuning" to "electric tuning." 1. GENERAL REMARKS 1.1 Electronic Tuning To add versatility and avoid mechanical difficulties, electronic control of the tuning of resonant circuits is desired for search receivers and other allied applications. Magnetic tuning uses the principle that the inductance in the resonant circuit can be varied by means of a magnetic bias field usually furnished by a variable current in a solenoid. Electric tuning uses the principle that the capacitance of the resonant circuit can be varied by an electric bias field usually furnished by a variable voltage. Both of these principles adapt themselves readily to electronic control. 1.2 Magnetic Tuning In magnetic tuning, a special magnetic core material is used which has a wide variation of permeability with applied field. It is the selection of a suitable core material which determines the success of the device. A class of magnetic ferrospinels or ferrites have enjoyed prominence in this field because of their high resistivity and freedom from eddy current losses at high frequencies. 1.3 Electric Tuning In electric tuning, advantage is taken of the ferroelectric property

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN of certain dielectrics -- the dielectric constant changes as an electric field is applied. There is a variety of ceramics falling in this category, but most of the commercial capacitors available are in the barium-stronium titanate class. The choice of a particular dielectric material for electric tuning is generally a compromise, since very high dielectric constant, low temperature coefficient and low losses are very often mutually exclusive, as will be shown. 2. MATERIAL PROPERTIES 2.1 Properties of Ferrites In comparing magnetic and electric tuning, we must first compare the corresponding properties of existing magnetic and ferroelectric materials. In considering the most important properties pertinent to magnetic tuning, we will confine our attention to the magnetic ferrites (ferrospinels). These materials have such high resistivities (103 to 109 ohm-cm.) that they have negligible eddy-current loss at all rf frequencies. For the maximum tuning range, the largest possible change in permeability is desired. Regardless of the initial permeability, it is theoretically possible to obtain a permeability approaching unity by driving the specimen hard into magnetic saturation. Therefore one should first look for materials having a large initial permeability, Lo. The quantity actually concerned in magnetic tuning is the incremental permeability, ILy. This is the permeability observed when the specimen is subject to combined ac and dc magnetic fields applied in the same direction. This is illustrated in Fig. 1. 4A changes with both bias field Ho and ac field A H. A convenient way of presenting these variations is by means of a three dimension

+ B AB! / AH -H I H, I H3 +H /I TtHo0 C. ~~ -BH HO 1/2 (H3 + Hl) a. mom..- /sowAB FIG. I DEFINITIONS OF MAGNETIC PARAMETERS

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN plot called a mu surface1. The mu surface for a typical ferrite is shown in Fig. 2. The mu surfaces are very useful in comparing the properties of two ferrites in a particular application, such as a swept oscillator where both Ho and AH are expected to vary. The permeability is also affected by a number of other,parameters 1 such as temperature, frequency and magnetic history as previously reported 2.2 Properties of Titanate Ceramics The titanate ceramics may be considered as representative of materials suitable for electric tuning. Because of this and the fact that most of the available data is concerned with them, our attention will be confined to the titanate ceramics in the present report. The important parameter in electric tuning is the dielectric constant, E. This is the parameter corresponding to permeability in magnetic tuning. For the maximum tuning range, materials showing the largest change in E with applied electric field E, are desired. When a very large electric field is applied to a ferroelectric material it is driven into saturation or maximum polarization, but the theoretical limit of E is not unity. This is because the dipole moment at saturation may still be increased by further increasing the electric field. In a practical situation however, the limit of E is generally determined by the electric breakdown field of the specimen. In addition, when both ac and dc electric fields are applied, thermal failure may occur at dc fields much lower than the ordinary dc electric breakdown field2. Corresponding to the mu-surface, a surface representation of E 5_ _ _ _ _ _ _ _ _ _

Z'91J a A-so~ooO''"4~ -%'ozO = 07O XD ---— 0Z. = xow'r/

ENGINEERING RESEARCH INSTITUTE * UNIVERSITY OF MICHIGAN may be used to show the variation with de field Eo and ac field A E. A typical Epsilon surface is shown in Fig. 3. It is noted that the variation in E withAE is not nearly as great as for the corresponding variation of ALA with AH. Although it is possible to design ceramic materials with very small or even zero temperature coefficient of E over a wide temperature range3, such materials are not particularly field-sensitive, and hence not suitable for electric tuning. Titanate ceramics in general exhibit a wide variation of E with temperature and dc field. This variation is conveniently presented by a three dimension31 Epsilon-Temperature surface4. A representative surface of this type is shown in Fig. 4. The point of maximum capacitance at each value of field is indicated )y a small circle on the figure. These points show that as the dc field is in2reased,the point of maximum capacitance (i.e. zero temperature coefficient) Shifts to a higher temperature. This Curie shift6), is an important factor to c*onsider in the design of electric tuning units.?.3 Comparison of Properties A graphic comparison of the properties of ferrites and titanates is given Ln Fig. 5. Fig. 5A indicates the typical properties of a high-mu ferrite. Note hat there is the desired large change in lL with a relatively small change in agnetic field H, but that this is offset usually by a large temperature cofficent of AL0 and some variation of AL with frequency up to the useful limit 0 c where the losses become excessive. A great improvement in temperature cofficient of Ato may be made by special composition low-mu ferrites7,8 as in B. owever it is found that a much smaller tuning range is possible because of the ow sensitivity of AL to bias field H. Although the useful frequency limit fc ay be larger than in A, the losses are usually larger at all frequencies below c which is undesirable. Because of longer and more extensive research in ferrite 7

29-1-01 MY3r qv-os-v o.LG1w O0 00,00 00/'coo~~~~~~~~~~~~~~~O 0o/ 00.000/ 0~~~~~~~~~~~~~ ~00 00 i0 00?-~~~~~~~~~~~~~~~~~~~~~~~~t Oo I~~~~f.O:50 ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~ FIG. 3 EPSILON SURFACE FOR AEROVOX "HI-Q" BODY 40 AT 26"C. SPECIMEN THICKNESS= 0.02 INCH. 8

29-0~-6 VY~r ~,-tS-0 OL6-4Y ~~~~~~~~~~~~~~;~~~~~~~~~~~~~~~~~~~~~~,~ ~21 X~~~~~~~~~~~~~~~~~~~~.lo. _tp~~~~~~~~~~~~~~~S FIG. 4 EPSILON-TEMPERATURE SURFACE FOR AEROVOX HI-0 41. EPSICO~~~~~~O- ~ TEM RTRESRFe FOR AEROVOX HI —9 41.~ ~ ~~~ —

ILo I fc TEM P. H (OERSTEDS) FREQ. A. HIGH IL FERRITE Lol TEMP. H (OERSTEDS) FREQ. B. LOW IL FERRITE TEMP. E (VOLTS/MIL ) FREQ. C. HIGH E TITANATE TEMP. E ( VOLTS/ MIL) FREQ. D. LOW E TITANATE FIG. 5. COMPARISON OF PROPERTIES. 10

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN tilning units, it has been possible to make a fairly good compromise between temperature sensitivity, and tuning ratio (See Section 3.3 below). The corresponding curves for a titanate ceramic are shown in C and D. The high-E titanate in C has a large field sensitivity but the temperature variation is even worse than in A. However Eo stays reasonably constant with frequency up to the useful limit fc which is generally a few hundred megacycles for both C and D. The temperature sensitivity is improved in D but at a cost of smaller field sensitivity. However in this case the losses are lowered from C to D instead of increased as was the case from A to B. Fig. 5 illustrates the situation in a general way, and indicates the sort of compromises which must be made. However this does not demonstrate certain other considerations which must be dealt with in detail. It also does not bring out the important point that the specialized materials now being developed may greatly alleviate the situation. Because the demands of magnetic tuning have been felt over a period of several years, there has been considerable attention to the development of lerrite materials having large field sensitivities in conjunction with other de3irable properties. On the other hand, the main trend in titanate ceramic develop. bent over the past few years has been to reduce the field sensitivity. This is 4ighly desirable for the producer of linear capacitors, since it permits him to $ssign larger dc voltage ratings without reducing the GMVi capacitance, but not Pt all helpful to the designer of electric tuning units. However it is safe to say that at present there is no ideal material for tither electric or magnetic tuning. There is also no one material which reprebents the best compromise for all frequency ranges. *GMV = Guaranteed minimum value 11

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN 3. COMPARISON OF MAGNETIC AND ELECTRIC TUNING 3.1 Manufacturing Problems A major problem in tuning unit manufacture is that of obtaining a supply of ferrite cores or titanate capacitors with uniform and consistent proper. ties. Because magnetic tuning is farther advanced, ferrite property tolerences for a particular application have become quite rigid, and more is known about the sort of difficulties which arise. In particular, a ferrite core material chosen for a specific application in magnetic tuning tends to vary from batch to batch and from month to month. The ferrite manufacturer in turn has a difficult problem in control of raw materials -- even small differences in the purity of the raw oxides have a profound effect in the variations of product properties. These also vary between individual cores of the same batch unless core production is in unusually good control. A consequence of this second variation is that the tuning unit assembly plant has a difficult core-pair matching problem, and a large wastage or shrinkage due to unusable cores. No such severe problem is encountered in the production of titanate capacitors because of two facts. The first fact is that present-day quality control of titanate ceramic production is quite good. The second is that titanate property tolerances for electric tuning are not as severe as for magnetic tuning. There is nothing corresponding to the core-pair matching problem in the construction of electric tuning units, since an exact capacitor-pair match is not essential to the electric tuning unit. For search receivers, the required tracking of the various tuned circuits creates a tracking problem which imposes further tolerance requirements 12

ENGINEERING RESEARCH INSTITUTE * UNIVERSITY OF MICHIGAN on tuning units. For a particular search receiver, the wastage of magnetic tuning units which fall outside tolerance is serious because of the limited market for such specialized units. In the case of electric tuning, however, ferroelectric capacitors which fall outside tolerance for tracking requirements can be diverted to a large number of other uses, and thus do not represent a loss to the manufacturer. Unique to electric tuning is the problem encountered when the dielectric is made very thin. (A thin dielectric is desirable for electric tuning as discussed in Section 3.6 below.) It is more difficult to produce thin sheets of titanate dielectric having apparent homogenious properties. This is because the effect of small, local impurities in thin specimens is very noticeable -- even tending to produce electric breakdown in some cases. In thicker specimens, the effect of small, local impurities is masked by the larger volume of surrounding homogenious material. 3.2 Cost A magnetic tuned resonant circuit consists of a variable inductance element and a fixed capacitance element. The major expense is in the former. First the ferrite cores must be manufactured, sorted and selected. The tuning unit may consist either of two matched cores, or of a single core with special machining. There is therefore the cost of machining, which is large for ferrite materials, or of core matching, which involves semi-skilled labor plus the resuiting shrinkage mentioned above. To this must be added the cost of winding, handling, assembling, and final testing. It is clear that the final cost of the resonant circuit is considerable. An electric tuned resonant circuit consists of a fixed inductance and 13 _

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN two variable capacity elements, plus an inexpensive isolating element such as a resistor or rf choke. Here the cost is approximately equally divided between the inductance and the capacitors. In the manufacture of the capacitors, large sheets of the dielectric may be made at a time and sliced down to size. Electrodes are plated on and electrode wires attached. The capacitors are then coated with a resin covering by dipping. Finally, the finished capacitors are usually sorted on semi-automatic testing bridges, so that the final product for use in tuning units may be held within tolerance. In addition, the rejects may be sold as by-pass capacitors not requiring the tight tolerance. The total cost of a finished electric tuned resonant circuit is therefore considerably less than the corresponding magnetic tuned circuit. Although actual manufacturing costs are not available at present, it is estimated that the cost of an electric tuned resonant circuit is from 3 to 8 per cent of the cost of a corresponding magnetic tuned resonant circuit. 3.3 Temperature Stability Magnetic tuning units are now being produced with very good temperature stability9. Fig. 6, reproduced from the reference article, shows the tuning temperature curves for an Increductor in a resonant circuit operating between 28 and 55 megacycles. When the control current is 10 ma, the temperature coefficient of frequency is approximately zero over a wide temperature range (-40~C to +80~C). Electric tuning, to give the same frequency ratio, must employ ferroelectric capacitors capable of at least 4:1 capacity variation. Presently available titanate capacitors in this class have relatively poor temperature stability. For comparison with Fig. 6, the predicted Tuning Temperature curves for 14

V"-9g W3P OL-VO9-V OZ61 55 IC. 50MA 4 0 MA 30OMA 50 20 MA 45 IOMA | 40 -_ - = CONTROL ci CURRENT LL 4MA 35 25 i- - TEMP. 0C FIG. 6 INCREDUCTOR TUNING TEMPERATURE CUGVES. 15NCREDUCTOR TUNING TEMPERATURE CURVES.

ENGINEERING RESEARCH INSTITUTE * UNIVERSITY OF MICHIGAN an oscillator using Aerovox Hi-Q 41 20 mil Capacitors as the variable elements are shown in Fig. 7. These curves are plots of the ratio ( emax/E )1/2 which is proportional to oscillator frequency, using the same data as for Fig. 4 E occurs at 410 C and zero electric field. max It should be mentioned that Aerovox Hi-Q 41 does not necessarily represent the "best" material available, but as the curves show, a frequency ratio of 2.4 is possible with the temperature stabilized at 400 C, and a control voltage varying from zero to 2000 volts. It is clear from the curves that there is no value of control voltage that gives a zero temperature coefficient over a wide temperature range as was observed in Fig. 6. The design of an oscillator using electric tuning which must operate over a wide range of ambient temperatures must therefore include temperature control of the capacitors. This is not a serious problem however, because of the very small volume occupied by the capacitors. It is not difficult to build a capacitor tuning unit and a low-watt thermostat-controlled heater in a miniature oven having a volume of 0.1 cubic inch. 3.4 Time Relaxation Phenomena Time relaxation effects are exhibited by both ferrites and titanate ceramics. In the case of ferrites a time decrease of permeability is observed after a ferrite core is subject to a degaussing treatment. When used as a tuning element in a low power oscillator, this effect can cause up to 3.8 per cent increase in frequency in Ferramic G over a period of several hours1. This effect was reported by Snoek in 1947, and is variable from one material to another. In general, those materials showing a large change in hA with applied H, have also a relatively large relaxation effect. __ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 16 _..___ _ _ _ _ _ _ _ __ _ _ _ _ _ _ _ _

3.0 SPECIMEN THICKNESS = 0.020 IN. CONTROL VOLTAGE = Vp cl: u6 2.5 0 p - 2000 v ~c 2.0 0 0., 0 20 40 60 80 100 0 20 40 60 80 I00 TEMP. IN 0C FIG. 7 AEROVOX HI -Q 41 TEMPERATURE CURVES. 17

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN - A similar relaxation effect is noted in titanate ceramics, although this doe not correspond exactly to the effect just described. If an electric field is applied to a specimen of titanate ceramic for a period of several minutes and then suddenly removed, the dielectric constant increases first abruptly and then slowly with time. When such a ceramic is used as a tuning element in a low poweroscillator, the frequency drops abruptly at the time the field is removed. This is followed by a relatively slow decrease in frequencyll. The frequency drift substantially ceases after a period of 5 to 30 minutes, depending on the material In this respect the effect is shorter-lived than the corresponding magnetic relax ation. In general, the longer period is associated with materials which show a large change in E with applied electric field, such as Centralab K 70006. In a swept receiver applications, the consequences of these effects are as follows: A. Reduction of Tuning Range. When swept at 60 cycles or faster, the tuning range will be smaller by several per cent than it is with manual, arbitrarily slow tuning. B. Frequency drift. If the amplitude of frequency sweep is reduced, or stopped altogether, there will be a slow frequency drift. The drift will probabl; be small in both magnetic and electric tuning, but is shorter-lived in the case of electric tuning. C. Sweep waveform. If a sawtooth sweep is employed, it is more satisfactory for electric tuning to use a jump-rise, slow decay wave than a Jump-drop, slow-rise waveform. The reason for this is that the ferroelectric ceramic responds more quickly to a jump-rise step-function of voltage than to a jump-drop step-functionll. A consequence of this is that the ferroelectric-tuned unit is best used to sweep downward in frequency, starting at the highest frequency and 18

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN sweeping to the lowest. In magnetic tuning, the frequency may be swept in either direction but there is some slight preference for sweeping upwards in frequency because of circuit considerations. This mode of sweeping requires a slow rise and abrupt drop in control current. Because the control tube may be rapidly cut off, permitting a high inductive kick from the control winding, the flyback time may be made somewhat shorter than in the other mode. 3.5 Frequency Range and Tuning Ratio The frequency ratio is the ratio of upper to lower frequency of a swept oscillator or amplifier. It is desired that this be as large as possible so a relatively small number of receiver units may cover the entire communications rf spectrum. In both magnetic and electric tuning it is increasingly difficult to obtain large frequency ratios as the starting frequency becomes higher. Tables 1 and 2 indicate representative frequency ranges and ratios for increductor-tuned and ferroelectric-tuned oscillators. In the frequency ranges below 10 me, magnetic tuning gives much larger tuning ratios than are possible with electric tuning. Magnetic tuning offers ratios of 10:1 or better up to 30 me. For starting frequencies in excess of 30 me, magnetic tuning gives ratios of 2:1 or less, and electric tuning can compete in this frequency range. At a starting frequency of 50 me electric tuning gives a tuning ratio of 2:1 which is superior to that available with magnetic tuning. The theoretical limit for magnetic tuning is set by the ferromagnetic resonance phenomenon. This phenomenon occurs in ferrites at some frequency between 500 and 3500 mc. Because of the fact that materials having adequately low losses at very high frequencies have a rather small field sensitivity, the 19

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN TABLE 1 MAGNETIC TUNED OSCILLATORS12 Maximum Frequency Range Frequency Ratio Oscillator Type 15 kc -- 2 me 130 Wien Bridge 0.2 mc -- 2 mec 10 Tuned L C Type 3 me -- 30 me 10 20 me -- 60 mc 3 75 mc -- 95 me 1.27 100 mc 1.1 to 1.3 200 mc -- 210 mec 1.05 Table 2 ELECTRIC TUNED HF OSCILLATORS5'6 *Observed Frequency Range Frequency Oscillator Applied control field imc Ratio Type volts/mil 50 -- 100 2.0 Hartley 0 -- 100 95 -- 115 1.21 Push-pull O -- 60 110 -- 140 1.27 Colpitts 0 -- 60 135 -- 160 1.18 Push-pull 10 -- 60 365 -- 375 1.03 Ultra-audion 0 -- 60 *These figures do not necessarily represent the maximum range, but represent behavior of oscillators which have been constructed in the laboratory.

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN tuning ratio is considerably reduced as the top frequency is raised, as indicated in Table 1. Increductor-tuned oscillator units9 have been operated at frequencies in excess of 300 minc. The top frequency for dielectric tuning is limited by piezoelectric resonance phenomena and increased losses at high frequencies, but the ultimate theoretical limit is determined by a ferroelectric resonance which appears somewhat above 1000 minc. A major problem in the design of very-high-frequency power oscillators using electric tuning is that of heat removal from the dielectric. Development work is now in progress on high frequency electric tuned oscillators at the University of Michigan, and results on bread board models are shown in Table 2. Satisfactory operation up to 375 me has been achieved in one experimental model using the Ultra-audion circuit5'6. Because of the increase in losses with frequency, the Swept Oscillator has a higher frequency limit than the Swept rf Amplifier. This is due to the fact that a low impedance tube may be used to furnish tank losses in the Oscillator, whereas such losses would be excessive for satisfactory rf amplifier operation. This consideration applies to both magnetic and electric tuning. The present practical limit of frequency for magnetic tuning is of the order of 100 me for amplifiers, and 300 mec for oscillators. An interesting development in wide band swept oscillators has recently been released by the Kollsman Instrument Company, Elmhurst, New York. Their Type 2144 Wide range Sweep Generator will presently be available in three ranges: (a) 225 to 420 me, (b) 470 to 890 me, and (c) 850 to 1275 me. The output is 10 mw from a 50 ohm source impedance, and the full range is swept at a 60 cycle rate with ~ 1 db variation in output. The actual circuit has not been released. 21 _

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN - 3.6 Scanning Rates Magnetic tuning is accomplished by varying the dc magnetic control field in a magnetic tuning unit, Fig. 8A. This is done by varying the current in a control winding, and electronic means must be provided to furnish the required variable current i. To obtain sufficient maximum magnetic field at a reasonable control current, a large number of turns is required on the control winding. This generally results in a control inductance of several henries. If L is the inductance of the control winding, the electronic means must furnish a voltage L di/dt to produce the desired rate of change of current. Electric tuning is accomplished by varying the dc electric field in the titanate tuning capacitors, Fig. 8B. Electronic means must be furnished to do this by changing the voltage e. To obtain sufficient maximum electric field at a reasonable voltage the dielectric in the tuning capacitors is made thin. If C is the effective capacity in the control circuit (generally four times the rf resonating capacity), the electronic means must furnish a charging current of C de/dt to produce the desired rate of change of electric field. With relatively low scanning rates (i.e. 60 cycles) there is no serious problem in furnishing the electronic means for either magnetic or electric tuning. When the scanning rate is increased, the control circuit design problems become more difficult. For high scanning rates a short "flyback" or return time, say on the order of one microsecond, is desired. This presents a rather severe problem in magnetic tuning as illustrated by the following example: A rapid scan magnetic tuning unit tunes from 50 to 100 mc signal frequency and has the following properties: Control Inductance - 2.0 henries Control Current = 50 ma (max.) 22,,,

CONTROL WINDING T ~t L~ RF A. MAGNETIC TUNING UNIT dt RF;0~ =L2~~~~. dCIRCUIT CONTROL e CIRCUIT C2'i CAPACITOR B. ELECTRIC TUNING UNIT FIG. 8 TUNING UNITS 23

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN For a flyback time of 1.0 microsecond, the flyback voltage is found to be Ldi = 2.0 x.05 x 106 = 100,000 volts. This serves to illustrate the seriousness of the problem not only in control circuit design, but also in voltage insulation in the design of the control winding itself. In order to improve the situation, we first note that the magnetic field is proportional to the number of ampere turns NI. The inductance of the control winding is proportional to N2, so the flyback voltage may be reduced at the expense of a larger control current i. Let us re-examine the calculation just made if N is reduced by a factor of 10. We then have a unit which has the following control constants: Control Inductance = 0.02 henries Control Current = 500 ma (max.) (For same NI max) For a flyback time of 1.0 microsecond the flyback voltage is found to be L di - 0.02 x.5 x 106 = 10,000 volts. dt Thus an improvement is acheived in flyback voltage, but at a price of increased control current maximum, which imposes a difficult control circuit designproblen There is another serious effect in flyback with magnetic tuning. The control inductance resonates with its self capacity C in Fig. 8A, and generates a serious flyback transient if not properly damped. Since it is highly desirabl to use pentodes for the control circuit because of their constant current property, an additional damping circuit is generally required for rapid flyback to suppress the flyback transient. A corresponding example of rapid scan electric tuning is a unit 24 ~,,

ENGINEERING RESEARCH INSTITUTE * UNIVERSITY OF MICHIGAN tuning from 50 to 100 me signal frequency with the following properties: Capacitance in control circuit = 200 uuf. Control Voltage Swing = 500 volts (max.) For a 1.0 microsecond flyback time, the flybackcurrent is found to be C = 200 x 102 x 500 x 106 0.1 ampere. dt This flyback current may be easily obtained from a conventional vacuum tube, so the design problem is reasonably simple. Even faster flyback times are possible by using thyratrons. There is also no problem of flyback transients with electric tuning. It is true that the stray circuit inductance might tend to resonate with the control circuit capacity, however this is easily damped. A series resistor is generally present for isolation puropses, so that no additional damping is required. There are thus a number of definite advantages in electric tuning for rapid scan applications. It may also be noted that as the signal frequency is increased, the control capacity C generally decreases. In this case it is seen that faster flyback times are possible at higher signal frequencies without increasing the flyback current pulse. This is not the case in magnetic tuning where the control inductance does not decrease as the signal frequency is increased. 3.7 Very Rapid Scan Systems Scanning search receivers at very rapid rates, such as 50 megacycles per microsecond, has been shown to have certain unique propertiesl3 which may be desirable for particular services. At present it is not known how valuable such systems will be, but their investigation is important.

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN - Scanning at 50 megacycles per microsecond may be performed if it is possible to scan continuously the units given as examples in Section 3.6 at their flyback rates. In this case a triangular wave of 500 kc and the proper amplitude might be used. From the preceeding discussion it is clear that electric tuning might be used at these rates, but that magnetic tuning is unsuitable. 4. CONCLUSIONS In performance, magnetic tuning is generally more advanced and therefore more satisfactory than electric tuning in a large number of categories. The category in which electric tuning is definitely superior is that of rapid scan. Here magnetic tuning is restricted to scanning rates which are low compared to those possible with electric tuning. In cost, electric tuning is extremely attractive, and as a rough estimate is between 3 and 8 per cent of the corresponding cost of magnetic tuning. The cost of electric tuning may be decreased even further if production methods and printed circuits are employed. The chief difficulty with electric tuning at present is that of obtain. ing a large tuning ratio together with low temperature sensitivity. But it is possible that a combination of new materials and techniques may solve this difficulty as electric tuning becomes further advanced. These conclusions are based on presently known materials and techniques in two rapidly growing fields, and therefore represent the situation only as it appears at the present time. _ _ _ _ _ _ _ _ _ __ ~26

References 1. L. W. Orr,. "Permeability Measurements in Magnetic Ferrites" Technical Report No 9, Electronic Defense Group, University of Michigan, Sept. 1952 2. Private communication with Mr. N. A. Terhune, Squier Signal Laboratory 3. Glenco Corporation, "Development of Close Temperature Coefficient Tolerance Capacitors for Temperature Compensation" First Quarterly Report Contract DA-36-039 sc-42502, March 13, 1953 4. L. W. Orr, H. Alperin, H. Diamond and M. Winsnes, "Use of Ferromagnetic and Ferroelectric Materials in the Tuning of RF Components" Quarterly Progress Report No 9, Task Order EDG-4, Electronic Defense Group, University of Michigan, Oct. 1953 5. L. W. Orr, L. Beavis, R. Bradley, H. Diamond and M. Winsnes, "Wide Range Tuning Methods and Techniques Applicable to Search Receivers" Quarterly Progress Report No 10, Task Order EDG-4, Electronic Defense Group, University of Michigan, Jan. 1954 6. H. Diamond and L. W. Orr, "Interim Report on Ferroelectric Materials" Technical Report No 31, Electronic Defense Group, University of Michigan, (to be issued). 7. E. Gelbard, "Magnetic Properties of Ferrite Materials" Tele-Tech, May 1952 8. E. Albers-Schoenberg, "Ferromagnetic Oxide Bodies —A Counterpart to the Ceramic Dielectric" Ceramic Age, Oct. 1950 9. A. L. Kaufman, "Circuit Design with Controllable Inductors" Electronic Design, v. 2, n. 4, p. 12, April 1954. 10. J. L. Snoek,"New Developments in Ferromagnetic Materials" Elsevier Publishing Co., New York, 1947. 11. L. W. Orr, H. Diamond and M. Winsnes,"Use of Ferromagnetic and Ferroelectric Materials in the Tuning of RF Components" Quarterly Progress Report No 8 Part I, Task Order EDG-4, Electronic Defense Group, University of Michigan, July 1953 12. C. G. Sontheimer, "Application of High Frequency Saturable Reactors" Proceedings of the National Electronics Conference, Vol 9, Feb. 1954 13. H. W. Batton, R. A. Jorgensen, A. B. Macnee and W. W. Peterson, "The Response of a Panoramic Receiver to CW and Pulse Signals" Technical Report No 3, Electronic Defense Group, University of Michigan, June 1952 27

DISTRIBUTION LIST 1 copy Director, Electronic Research Laboratories Stanford University Stanford, California Attn: Dean Fred Terman 1 copy Commanding Officer Signal Corps Electronic Warfare Center Fort Monmouth, New Jersey 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 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 - Signal Property Officer Inspect at Destination File No. 25052-PH-51-91(1443) 28

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 10 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 29