A HIGH GAIN VIDEO AMPLIFIER Thai. for the Degree of M. 3. MICHIGAN STATE COLLEGE Robert Orlando Bolster 1941 THEIl‘S I A I 3H GAIN VIDEO AMPLIFIER by Robert Orlando Bolster ~— Submitted to the Graduate School of Michigan State College of Agriculture and Applied Science in partial fulfilment of the requirements for the degree of MASTER CF SCIENCE Department of Electrical Engineering 1941 THESiS Acknowledgement 1 The author wishes to express nis gratitude to hr. F. V. Schultz, formerly of the Department of Electrical Engineering for his cooperation in the develOpment of this tiesis; and to Professor L. S. Foltz, Mr. B. K. Osborn, and Mr. R. Gibson for their kind assistance in checking the manuscript. 11.0.3. 134213 Introduction. Part I -- Theory of the Amplifier. 1. Problem of Constan Gain. 2. Problem of Constant Time-Delay. Part II -- Des ,n, Construction, and Cperation f . 1. Preliminary Design. . Construction and Operation of the First Circuit. 3. Construction and Operation of the Second Circuit. Conclusion. 'I Bibliogratiy. 41 50 52 Introduction The purpose of this thesis is to present the theory, design, construction, and Operation of a new type of video amplifier. In television equipment the performance of the video amplifier is always a question because of the tremendous band-width involved. For xcellent reproduction of the original image, using standard transmission, this band-width spreads from about 30 cycles to about 4,000,000 cycles per second. Actually the band allotted in the radio frequency spectrum for the televised image is 4,500,000 cycles per second. 'In order that the picture elements will form a normally shaded reproduction with a prOper distinction be- tween details, it is necessary that the gain and the time- delay of all amplifiers be as nearly constant as possible. The maintenance of a constant gain and time-delay across the 4 megacycle per second band is the foremost problem encountered in video amplifiers. The greatest evil in the problem of constant gain is the shunt capacitance of 'Ude video amplifier's plate circuit to ground. This shuint capacitance which limits the allowable plate load tC) a very low value is composed of the amplifier tube's cuitput capacitance, the input capacitance of the follow- jfhi tube, and the capacitance to ground of the plate 2 circuit elements and wiring. In this thesis an attempt is made to eliminate this shunt capacitance through the use of a reactance tube displaying negative capacitance. A discussion of the application of negative capacitance to the video amplifier is presented in Part I. The low frequency response of the video ampli- fier will not be considered in this thesis. Part I Theory of the Amplifier As brought out in the Introduction, the foremost problems encountered in the video amplifier are the maintenance of a constant gain and of a constant time delay across the tremendous band-width. These two factors will be treated separately in the following work; the problem of constant gain being discussed first. 1. Problem of Constant Gain. Concept of Negative Capacitance. Since the shunt capacitance to ground of the video amplifier's radio frequency plate circuit limits the allowable plate load resistance and thus the gain, a solution to this problem would be effected if this shunt capacitance were eliminated or neutralized. To accomplish this elimination or neutralization, the phe- nomenon of negative capacitance, as defined below, is introduced. With a sinusoidal voltage applied, a positive capacitance is characterized by a current vector which leads the voltage vector and a reactance which decreases with increasing frequency. A negative capacitance, as defined for this work, is characterized by a current 4 vector which lags the applied voltage vector and a.react- ance which also decreases with increasing frequency just as does a positive capacitance. Thus negative capacit- ance and inductance d ffer in the manner in which their reactances‘vary with frequency. There is a fundamental difference, however, be- tween negative capacitance and ordinary circuit elements; negative capacitance can appear only as a dynamic cir- cuit element. This fact at once suggests the use of a reactance tube network to obtain the desired negative capacitance. A reactance tube is merely a vacuum tube supplied with the correct alternating grid voltage to exhibit the Q: esired relationship between plate voltage and plate current. Thus the tube may exhibit positive and nega- tive resistance, inductance, or capacitance depending upon he arrangement of the grid circuit. Example of the Possibilities of the Reactance Tube. Consider the circuit of the elementary phase modulator of Figure l. Oscillafor f R Ohm? 6 CI LI Figure l I Ca 1-1 1 _ AAAA AA E?%f 4;. in 5 c is assumed to determine the oscillator frequency. \ o is a large condenser having a reactance which is in the operating range, very small in comparison with th resistance of R. Then the current through EC is in phase with e. if, in phase with e,, thus lags e by 900.. Now, if an increase in frequency occurs in the oscillator, e, decreases; and 1,, being directly prOportional to eg, also decreases. These conditions indicate that the tube is acting as a dynamic positive inductance since the current vector lags the applied voltage vector by 900, and the reactance increases (current decreases) with increasing frequency. Thus the frequency of oscillation actually depends not only upon L,C,, but also upon the reactance tube circuit. If now an audio frequency signal is applied to the suppressor grid through the micrOphone and trans- former, the amplitude of 1P varies in accordance with the audio signal voltage. As i, varies, so varies the inductance in parallel with Li; and, as the inductance varies, so varies the frequency. Thus the frequency of oscillation varies with the audio signal, resulting in a form of phase modulation. Incidentally, it will be noticed that the variation of the amplitude of ip with the audio signal is amplitude modulation which is necessary for the suc- cessful Operation of this circuit. f 0 Determination of Grid Circuit Arrangement. 3 The example presented above indicates tn (D method of connection of the reactance tube in the cir— cuit. This method is carried over directly into t 1 ne video amplifier circuit shown in Figure 2. For the pur- pose of simplificati n, triode tubes are shown here, al- though pentodes are used in practice because of their :h plate resistance and mutual conductance. 7 - - I T I ' _[ i» ‘ . ' Ce "9 —w——4ww R215 1 (P I 0-————~ws ‘D — — 1’ 1» C ‘T- ”—43 A 1, 1: S 1: itk' fl I {'3 C'T ' d ‘ - - i _._‘ 8* Figure 2 7 In the above figure C, is the shunt capacitance of the hirh potential radio frequency circuit to ground. ‘5‘ ~../ It is desired to neutralize Cs with the reactance tube shown to the right of Cs. In order that this may be accomplished, the vector diagram of Figure 3 must hold for all frequencies. Figure 3 Rememberin: that ic increases with increasing frequency ' 3 (assuming e constant), the problem now is to find a three terminal network, A, which will give the vector 7 relation above and at the same time insure that i, and LG increase with increasing frequency in the same manner. The solution to this problem is found by means illustrated in Figure 4. . cI=CbnwhwnE I lie: . .i . c H (a) j c, swig—Jew. t. +6 ’ f I 3 Giana/l PH (,3) e A ”it eg=Jew£§R |e,| +6 1 1 $ awn” PM ' F 1 "°"° e R e,-£.- '-’--- '-°— led +L (c) ‘3 C -R ’wC’ eoC‘R H" l :9 Tm. , , I l. """ I"? (C1) e 3 69:557. ”=‘J'g leg +L 1 :5 I? wmmfi P} i 7 7 ‘WVVO T R levl ((-3) e ’ eg=£xjwt = jig—t: . l 99 L R R I‘Pl +6 § :nwfl' _ F ‘f’ I «Weir I"! c i c, _ _ o ; Figure 4 8 From Figure 4 it is apparent that the only possible form of the three terminal network, A, is that which consists of a large resistance in series with a radio frequency transformer so connected that the polar- ity Oi the secondary is reversed from that of the primary. Thus Figure 2 becomes Figure 5. “ax: J go .1 '13: T vvvv" I : J5 «P—v Go 1‘ 1 I” [s 3 S- Figure 5 Examination of Grid Circuit Radio Frequency Transformer. It is of interest to make a close inspection of the transformer of Figure 4 (f) and Figure 5. Since the apparent reactance of the transformer on the primary side must be small, the inductance of L, and L, must be small. For this reason the values of distributed capac— itance across L, and L‘ become of importance because both primary and secondary must nave resonant frequencies well above 4.5 megacycles in order that the apparent re- actance will be inductive throughout the band-width. If the capacitance between windings is neglected here, the circuit under consideration is shown in Figure 6 (a). KO Intel-oi Figure 6 (a) . O Figure 6 (b) The equivalent circuit of Figure 6 (a) is shown in Figure 6 (b) with the assumed positive direc- tions of currents as shown. Now the gain of the circuit can be calculated as follows: a.) fl".....=-—91’ . . q, but 3 _H_j_£h‘ 3 ”Q and i-c’, 9”"ch- so that _j_‘L (2) aa1n=_°1=__f_vc+= fig, 8’ _J'_£!:£:l_ (‘1-‘I}Q me, Since the currents in parallel branches divide inversely as the respective impedances, J 4w" [Jill (Ll—H) —Iw (3) t. _ L. JwM+J"-LwL,-JwM —j _ t. a” —- l - a JW(‘1"N) —I.¢- I ”‘1‘ 555 , ,2." [u (tr-M) 45:] and . . _ . ”(A-”)4. Jule-4;} “" " - Q —J—— I we. Substituting equations (3) and (4) in (2) (5) To find the actual value of e,, e, must be evaluated. ‘I‘h ('7’) Gal? :3: C ’4 C3 «fave-4Q}!- L, . th'jwah-J%)-4L%E] . -l;-c-" ”00,-,” + jug—ii e,=" o“ _” _J._L -J;%.41m44_$0.+‘mJ4L'(z' ) ‘35. ' i“¢¢‘fzfis F I a L 1 ' '—£;a'(w’0weflgéo'+'E-) = ‘2 .§ + it +w‘(M‘—l,l.,) — iéc - e=—j z, Qw‘W-LLJH. ’ mac. .9... _l-L + my”. -4“)-— erefore, from equations (1) 39: -J. «TI—cc, and (6) 59H wC,€¢ [iii-+Q Lc_+w 0wl_ L,l')- -—1§?-:] Thus 11 If M, L,, La, 0,, and C2 are assigned represent- ative values, the expressions for gain and e, may be simplified to quite an extent. Let L,= Le: 80 microhenries 1‘-'I= K VL, L1 = 0.2x80 =16 microhenries C,= 5 micromicrofarads “'lO micromicrofarads (dist°ibuted capacitance of coil plus input capacitance of tube.) Then the gain is af‘ A‘x 25—Ji—zz4ai gain = _ _ _ In IO a at?!“ -c¢oo)xn "Mama" 3° Thus, providing the anti-resonance frequencies of L,C, and L“). are well above 4.5 megacycles, the gain may be represented by (8) gain = 1:,- L: Ifiow e, =_ j a" “no": / C X :0 [low ,4 x10‘ __ 6’“ ”IX/0-!" 0.0: XI" lit will be seen that the last term in the denominator of ‘tkie bracketed factor is predominant. ‘hus it is pos- £3i‘ble to write 0 . ‘9" I . (‘3?) =~—‘ x - Jui'ff .x J —- 9’ «no, __I_ " a‘QQE Sisssuming the same conditions as given under equation (8). *_.x_~_a__ . __ . 12 Equation (9) indicates that e, is shifted from i, by a positive 900. Actually this shift may be made the desired negative 900 simply by choosing the correct connection of the secondary. Also, equation (9) shows that e, is directly preportional to frequency as has been shown necessary in Figure 4. To date it has been tacitly assumed that the reactance tube's plate current is in phaSe with the grid voltage. This is strictly true only when the load is purely resistive. Since the load in this case is highly capacitive, the alternative is to have a plate resistance sufficiently high to overshadow the capacitive reactance. Thus the reactance tube must be a pentode. The Value of the Negative Capacitance Now the alternating current circuit of the video ww. H L .amplifier may be drawn as in igure (7). . V; 1! We, I ‘ R 0"! ______. D 1» T ”a; e EEK: 3E3 Figure 7. :111 Figure (7) 2 represents the reactance tube plate Circuit. 3,. includes Re, R , and R of Figure 5. C, is 4.. ”fie? shunt capacitance. 13 Since the reactance tube is a pentode, it may, in general, be regarded as a constant current generator. Whence 4=9m%0 Substituting the value of e grom equation (9) (is) 4=—;ww'm army" Now 4;: (5+ Q's-F ‘g or (11) " = _e_.. __c_ —Q_ . R;- + ._J'_l_ +J°————R ‘wck «mnmfiv It is desirable that the last two terms of equa- tion (ll) should be equal and opposite so that if‘will be in phase with e. If this condition holds, then the apparent negative capacitance must be __ ____ finPV, (12) c— T alue of neg- Ffimmlequation (l2 Eitive capacitance is dependent directly upon the mutual mutual induc- ) it may be seen that the tube and the The value of R is more or Ccanductance of the reactance teuice of the grid transformer. lJesss fixed in the neighborhood of 25,000 ohms as R must always be much greater than the equivalent reactance of 4-." O O bflE? radio frequency transformer primary. Effect of the Distributed Capacitance between Windings of Grid Circuit Transformer It is now of interest to investigate the effect 14 of the distributed capacitance between the windings of the reactance tube grid transformer. Since the lower ends of the two windings are tied together,.the maximum voltage between coils will be at the free ends; so that the distributed capacitance may be considered as being lumped across the free ends of the transformer. This is shown in Figure (8) (a) and (b) ' T? c — j 1 ~— Qt .1.an (a) (b) Figure 8 -1, l The circuit of Figure (8) b is not easily solved for e9»; therefore use is made of the equivalent ”section of the transfor111er. The conversion from; T to7r Inay be made according to equation (13) with reference to Figure 9 . a, =Jw(L.-~) 3 Woo (L. W) 1:55: 3% sin” age, iEiA 1 1! t Figure 9 214+elai+hll - a = ( ) 2A 2’ ( l 3 )_ b _ zeil+§§+631 ( ) 2,... a, ‘ (O) z _a,a.+i.a.+z.as . ¢._ '3: Thu s Now the equivalent circuit of Figure 8 (a) may be drawn as shown in Figure 10. C: I" A z *4 Figure 10 Let (15) A =l.,I..-H‘- ‘ ‘ 1 0 I r' N Now the values of tne parallel impedances, Z_, é , and av . . . Z , shown in Figure 10 may be found as follows: ._""337 )(a'~-,,—-£7,-) E -J';-é’-+J'“"TA”- or (a) 2’- ij Ay—fW—«uen (16) (b) 2”=,'_“},"1«,5A (c) Zié: 1+€;€_ ‘34 . The gain of the transformer circuit may now be evaluated. 0 U a! qu' ( 17) Gain :5! = 16 ‘. _‘. z'+z'“ _ .[z,-~’A 4+ C,][£.-M—~‘c,A] a - I 11' - 6 [L'_ M—uanAJEV— «1‘63A J Substituting 1, in equation (17) M - 49,49 A ‘0 - “’A(CD *6) (1Q) __. ”’JGé-‘rfi? l;- an+ Ca)“, Q-fl") From equation (17) Gain = e,=e gain . But __ kz’(z”+z'" _ iy-zH-z‘v- z" 01” J.“ A [ I I + I b—M-w’QA L,-M-~'C.A m-acm f 6:1 9 l I / qr + . + L.-M-au 64 £,-M—w’(',.4 m-MQA Simplifying this expression for e, there results ' J.“ [‘1 “”760 “3)4] 3:49 , o / -£. ~'(C.+¢:) -A.w‘(c'.+¢3) + w‘A(<'.<'.+¢'.¢3 +c.c:)+ iflu'c: Multiplying by the gain as given in equation (18), Jen/Va} (10 e = ' ,) ’ 1-4,w'(c,.c,)-L,.t(c,+c.)+«4(L,L.-n'Xc.c.+c.6.«schmw: If C3= 0, equations (18) and (19) become M (20) Gain: 1: " ”'62 (‘0‘! 47") 17 JauMhi and c =: ' l-w‘zc C, " u"! C! + 04C,C((‘,£g -M') Dividing numerator and denominator by -u’C,C, o ‘.'N I ”CC l 18_ I - l ' 2 a: + a; w (‘0‘! M) -WC. Equations (20) and (21) are identical with equations (3) and (7). Now, as before, representative values of M,L,, L‘,C,, 0,, and Q, are assumed as follows: L,= 2 = 80 microhenries M = K L, 12:0"? x 80 =16 microhenries C,=:5 micromicrofarads Ca: 10 micromicrofarads 03: 5 micromicrofarads Substituting the above values in equation (18), I: xro"- Jar 10""(6‘440 - £J‘)x/a"" BQMF‘LuJQRdeMQZfim-3adXEM' Gain: (22) 33.1: —fl . Under the same condition . . 4 e9: qux/(l/o . Therefore, (25) €9=jw g'y M where the sign of e, may be either positive or negative according to the connection of the transformer secondary. 18 The effect of the distributed capacitance between windings of the transformer is thus negligible if the value of the capacitance is of the same order of magni- tude as the distributed capacitance of one winding alone. 2. Problem of Constant Time-Delay. A constant time delay for varying frequency means that the phase shift must be directly proportional to frequency. The phase shift of an amplifier is inherently 1800 so that the phase shift which must be proportional to frequency is given by equation (24). (22+) I.=¢ -— 1800 WW I I.) I” .f‘g IV) A ALLA EgFHQ+JXp e Figure 11 Using the equivalent circuit of Figure 11, n‘ fibfgg a d _ e="£._ / a. ’p +20 so that /’£;:=./7@Lfifind ‘6’ n+2. I’P 4' 30+). ‘0 ’3 U1 OE} 90 H. :3 la" n Now the negative sign attached to e, takes care of the inherent 180O shift, so that the added phase shift, IL , is given by the phase angle of the gain. The tangent of ¢h Hmy'be found by separating the gain into its real .and imaginary parts as below: 19 main = /(?o+.l.x¢) = / (fig?) {'39:} “Pf/(10"; Max. ’P.m. V r, +Po {'ch (r, +P‘f4' X02 Therefore " Xorp O~ a: (M . -6) f PJ'f-EVp-I-X,‘ A lifier tube used is a pentode with a ..L plate resistance much mrcater than the plate load, t is restricted to small angles, Now, by inspecting the equivalent load shown in igure 12, the values of series reactance and resistance, _.. °__’_:: 3 3! v_R_ Figure 12 thus 2.: ,’ .fi— — + :2: ”9- iwymf7 = ”pf-3w ”r‘*(Qt-9MM) fiflikeuaflhf-wamfwgf. Substituting in equation (28), 20 ) hz-me-em/v) (29‘ R Therefore the phase shift of tne amolifier is always preportional to tne frequency, or the time-delay is al— ways constant. 21 Part II Design, Construction, and Operation of the Amplifier In this section are described the preliminary design, the construction and operation of the first cir- cuit, and the construction and Operation of the second circuit. 1. Preliminary Design. In the design of the amplifier, a number of dis- tinct problems were faced. These problems and their so- lutions will be presented in as IOgical an order as possible. The circuit of the amplifier as first preposed is illustrated in Figure 13. is I“; T 1! I I ‘1’“ I - I .»-- flav- . 3-17— Figure 13 In this Figure I is the amplifier or gain tube, a pentode. I? is the reactance tube, a pentode- u, is the shunt capacitance to ground. C. is the blocking condenser. 22 BL is the resistor through which direct current is supplied. A is the coupling network of the reactance tube grid. e&, is the input voltage. e.‘ is the output voltage. The 6AC7/1852 television pentode as T, seemed the best tube available. It is commonly used in video amplification circuits because of its high plate resis- tance and extremely high mutual conductance. The out- put capacitance of this tube, a factor contributing to Cs, is but 5 micromicrofarads. The normal Operating conditions of this tube as a clas A amplifier are given below. Heater Voltage 6.3 volts Plate Voltage 300 volts Suppressor Voltage 0 volts Screen Voltage 150 volts Cathode Bias Resistor 160 ohms minimum Plate Resistance (Approx.)0.75 megohms Pransconductance 9000 micromhos Plate Current 10 Milliamperes Screen Current 2.5 Hilliamperes Next the tube to be used as I? was determined. Since the grid voltage to this tube must necessarily be low because of the connecting network, the tube must have a.high mutual conductance. Now a consideration of 23 the magnitude of the current taken oy Cs at 4. 5 me3a- cycles per second will indicate the necessary current capacity of the reactance tube. Let Cs==40 micromicro fare is eu¢=l5 volts. Then ics=e£ =eaths=15 x 2fl'x 4.5 x 10‘ X 40 X 10”: s ::17 milliamperes. hultiplyin3 by VF;— (30) 14m; = 24 milliamperes. Because of the reouirements of high alternatin? - I, O plate current, and high mutual conductance, and iirh llate resistance (discussed in Part I); the tube cnosen as Ik was the 6A37, a television power pentode. The output capacitance of this tube is 12 :icromicrofarads. A bad Iea ture of thi tube is its hi3n input capacitance (D of 12 micromicrofarads. This value, however, is typical of the television pentodes. The normal operating condit— ions of the SAG? as a class A amplifier are given below. Hea ater volta3e 6.3 volts Plate voltage 300 volts Suppressor voltage 0 volts Screen voltage 300 volts Grid volta3e -10.5 volts Plate current 25 milliamp oeres Screen'current 6.5 zilliamperes Plate resistance 0.1 me3ohm Transconductance 7700 micromhos Iext encountered was the problem of plate feed fcxr the 6A0? and €A37. Tne combined direct plate CLLPIQHC of the two tubes was 35 milliamperes. It was dfiBSiIed to make the load resistance as high as possible irl order that most of the alternatin3 plate current ‘wcmnld be shunted through the reactance tube's grid cir— Cilit. Because of the requirement that the current crfrough this grid circuit be in ph—se with the volta3e axxross it, the value of the resistor here was fixed in J t'us neighborhood of 25,000 onms. Thus the plate feed rwesistor had to be at least 40,000 ohms. A 40,000 Ohms Iwesistor carryin: 35 milliamperes would have had a fkirect current volta3e across it of (31) a“ =IR=0.035x4u,0-00=l400 volts. ’3 )1 )- 1... ‘us the power supply would have had to supply 35 TI1111.1.iamperes at 1700 volts direct current. This was a ‘:r U aetical impossibility. Incidentally, the power dis— Siqoated by the above resistor would have been P = 1% = < I =1400x0.035 =49 watts <313th ugh no tests were run, it was feared that a resis- tor capable of dissipating 2+9 watts would have had an axiti-resonant frequency below 4.5 megacycles per second ‘NJe to distributed inductance and capacitance. Obvious- 13r this type of parallel feed was ruled out. 25 At this point it was recalled that pentodes have a.‘very high alternating current plate resistance. Thus .if‘ a pentode could be selected which was capable of EDEASSIHQ 35 milliamperes, the plate feed problem was scfilved. The 6F6} power pentode was selected. the output czaqoacitance of this tube is less than the equivalent 6F6 lneatal tube, and its other characteristics meet the Iwecpdirements as is demonstrated by its normal Operating ccndditions as a class A, amplifier. Heater voltage 6.3 volts Plate voltage 250 volts Screen voltage 250 volts Cathode resistor #10 volts Plate current 34 milliamperes Screen current 6.5 milliamperes Plate resistance 80,000 ohms Since the available power supply had a 500 volt Okrtput, it was decided to operate the 6AC7/lb52 and 61¥37 at a plate voltage of 250 volts. At this voltage, tine operating conditions as class A amplifiers are as follows: 6AC7/1852 6AG7 Heater voltage 6.3 6.3 volts Plate voltage 250 250 volts Screen voltage 150 140 volts Grid voltage -l.67 -3.0 volts 26 6A37/1852 6A3? Plate Current 8.0 35.0 milliamperes Screen current 1.72 6.7 milliamperes Plate resistance 0.7 0.0667 megohms C) C.) Transconductance 9'0 10,000 micromhos The values of plate resistance given in the above tiable were found experimentally and the values of trans- ccmnductance were determined from data given in the RCA Skeceiving Tube Kanual, 1940. Next undertahen was the design of perhaps the Inc>st important single part of the amplifier, that of the rweactance tube's grid transformer. As shown in Part I, if‘ the anti-resonant frequency is above 4.5 megacycles f the transformer is 0 pear second, the output voltage dfirectly proportional to the mutual inductance, M. Thus ift was desired to measure the anit-resonant frequency of Idsimary and secondary, and the mutual inductance. The circuit used to measure the anti-resonant I'1'“equencies of the coils is snown in Fixure l4. .1 R” I 5"9’", ZZZ—m 6J7 r Generoz‘or " v V” 84 ll u “ '.‘_’ 5x—rfl Figure 14 27 In Figure 14 the coil to be tested was connected (1 across the terminals 1—2, and a ~ignal which varied in frequency from 100 Kilocycles per second to 31 mega- cycles per second was impressed upon the grid of the (7\ J7. The lowest frequen y at which a peak was recorded on the vacuum tube voltmeter indicated the anti-resonant frequency of the coil in question. The resistor R I(25,000 ohms) was included to prevent the output capac- 1 itance of 6J7 from playinr a part in determining tne anti-resonant frequency. Thus only the input capacitance of the vacuum tube voltmeter influenced the anti—resonant frequency. This input capacitance was measured by determin- ing the change in frequency of an oscillator when the vacuum tube voltmeter was first connected across the frequency determining circuit and then removed. Since the capacitance across the above circuit with the VTVM removed must be known in this method, a General Radio Precision Type Jariable Jondenser was used. The precis- ion condenser was set at a value large enough that the distributed capacitance of the coil and the electrode cacacitance of the oscillator might be neglected. Then hr“ (32) C,=C. G: —l where C, is the unknown capacitance C. is the capacitance of the precision condenser, 28 's the oscillator frequency with C, in l the Circuit f. is the oscillator frequency with Cg removed, The resonant frequencies were determined using a radio receiver with a beat frequency oscillator. By this method Cx==l2.0 micromicrofarads. This value of Oz Jas checked by using the substitution method at 1000 cycles per second. here a null was found on a General Radio bridge using only the precision condenser mentioned above. Next the vacuum tube voltmeter was connected in parallel with the precision condenser and the precision condenser adjusted again to the same null. The change in the precision condenser setting gave directly Cfi=11.8 micromicrcfarads. Since this value was necessarily measured with the circuit cold and with the grid resistor disconnected, it is presumably slightly low. The value of the mutual inductance was deter- mined by measuring the primary, secondary, and total inductance with a 1000cycle per second bridge. At 1000 cycles per second the effect of the distributed capac- itance was negligible. The total inductance was measured series aiding aid series opposing as a check. From the values of total inductance (33) M=t Ef’g‘i’fl') '29 The positive sign was used when the coils were connected series aiding and the negative sign when the coils were connected series Opposing. In equation (35) M is the mutual inductance L7 is the total inductance L, is the primary inductance L3 s the secondary inductance. H. In Figure 15 are tabulated the values determined from the use of equation (33) and th circuitcf Figure 14. The resonant frequencies given are those obtained :ith the vacuum tube voltmeter in the circuit. The input capacitance of the vacuum tube voltmeter, l2 micromicro- farads, has been subtracted from the total capacitance to give the distributed capacitances listed. For the secondary windings the actual resonant frequencies were idearly the same, when the coils were in the circuit, as those given in Figure 15 because the input capacitance of the TAG? tube is very nearly the same as that of the ”vacuum.tube voltmeter. Coil 0’} Number it * (MmC©fl()O“fiJ>UHpRDH 9-31- 10* 12 12% 14 14* 15 lc-x- 5 17 18 19 20 w inding; Numser of Turns Pri. Sec. o.u. o.u. c.s. 21 5 o.u. o.u. 100 o.u. 100 v.o.u. 100 v. o.u. 100 c.s. 58 o.u. 100 100 o.u. 100 100 o.u. 50 5O o.u. 5O 5O v.o.u. 0 4O v.o.u. 40 4O o.u. 40 4O o.u. 4O 4O c.s. 100 c.s. 100 c.s. 75 c.s. 75 c.s. 50 o.u. 75 o.u. 55 o.u. 90 150 o.u. 80 100 I‘1 1/4 \H 0 Dimensions in Inches Pri. r2 V4 1 7/16 (/525/15 1/4 3/4 1/4 5/8 3/8 3/8 5/0 3/8 “/1 3/4 3/8 (/10 5/15 7/l5 5/16 7/16 3/8 13/323/8 15/525/ 7/Q // V 7/16 11/3 21/4 7/16 11/321/4 3/8 ,4 3/8 15/521/4 15/521/4 13/325/16 15/525/155 7/16 7/16 3/8 3/8 ;; 3/8 (-7 3/8 3/8 1/2 7/15 1/2 7/16 3/16 5/15 1/4 Sec. Sep. r1 r2 3/4 3'4 3/52 7/16 11/32 7/16 11/32 1 13/321/4 1/4 3/8 13/321/4 3/163/8 13/325/16 3/153/815/525/15 3/1611/3213/323/16 3/1611/3213/323/16 1/163/8 9/16 5/16 1/163/8 1/2 1/4 Figure 15 (concluded pn next page) In Figure 15 * means powdered iron core o.u. C.S. means close solenoid means 0p en universal o.u. means close universal v.o.u. means very Pri. 8C. 0‘) means primary means secondary r1 is inner radius open universal 51 r2 is outer radius 1 is length of winding Sep. means separation between windings. Coil Resonant Inductance, hutual Distributed 2km Frequency, hicrohenries Inductance, Capacitance, Megacycles hicrohenries Kicromicro- ' farads Pri. Sec. Fri. Sec. Fri. Sec. 1 0.19 17,100 28.1 2 5.94 35 5.2 80A 6.75 15.4 51 10 6.0 2.1 5 5.76 10 6.2 4 5.29 125 6.7 5 5.76 100 5.9 6 4.28 88 5.7 6* 5.65 112 5.2 7 6.25 39 4.3 b 5.25 5.25 155 155 21 6.0 6.0 8* 2.46 2.55 250 255 57 6.2 6.5 9 4.48 4.50 85 85 14 5.2 5.1 9* 4.47 4.56 69 72 19 6.4 3.5 10 5.07 5.07 72 72 29 1.7 1.7 10* 4.0 4.19 64 60 25 12.7 11.9 12 4.62 4.67 87 8 57 1.7 1.8 12* 5.86 5.75 65 69 17 14.2 14.5 1 5.75 82 12.0 14* 2.4 125 25.5 15 4.5 70 5.9 15* 2.8 95 25.9 16 5.6 57 2.1 17 3.8 98 5.4 18 5.1 70 1.9 .19 5.78 2.85 110 177 42 4.1 5.6 2%) 5.97 5.57 100 150 40 4.1 5.2 Figure 15 (concluded) A very striking thing will be seen in the tabu- latitnis of primary and secondary inductances. The in- sertdxni of a.powdered iron core increased the inductance i1s Triving 75 turns or better. This is the normal 0 of‘c: resnxlt. For'coils having 50 turns or less, however, the 52 innmtmwm decrea ase d with the insertion of the powdered irm1ccna The core, being a semi—conductor, increased finadishdbuted capacitance ufiicien 1y that the res- cmant:frequency of t.e coil cecreased. There is no awarente., anation of this phenomenon. Coils numser 12 and 20 were wound on concentric paper Hues which were tightly fitted so that the value of:1fipel.inductance might be varied simply by sliding the inner tube up and down in the outer'tube. Coil No. 10 proved to be the only coil which had an anti-resonant eh enough that equations (8), (9), and (12) ap plied acress the complete band. 2. Co nstruc m1 n and Cperation of the First Circuit. Based upon the prelim'nary design of the pre- ced n2 section, the circuit of Figure 16 was constructed upon a metal chassis. 61"66 , 6A 67 .1— ; ' 31m ,4. .4 Jfifiomac “ Figure 16 In Figure 16 Et= l megohm Rg= 160 onms R3=460 ohms R,=25,000 ohms 30 ohm s :0 h II .002 mica-I- .ol+ 25 electrlytic (microfarads) .002 u1ica+.01 paper-+1 paper (microfarads) C,=C‘ =C’= 0‘: .01 paper RFC = Radio Frequency Choke - 8 millihenries T as determined from Figure 15. The heater circuits are not shown in Figure 16. F60 was 250 direct current I 0\ Since the cathode of the volts above ground potential, it was necessary to use .a separate 6.3 volt transformer with the secondary at th same direct potential as th 6F6} cathode. Here another .protflfinn‘was encountered in the form of the cathode- heaimnc capacitance. The impedance of the transformer secondary to ground was very small so that the cathode- heater capacitance, which was thought to be large, practically shortcircuited the 6A07/1852 alternating plate current to ground. The approximate value of this capacitance was calculated from the‘ formula for the capacitance for two infinite concentric cylinders. Figure 17 (34) ks x /o_ MM: (1— V—_ _ — ’7 ’7 Here C = capacitance in micromicrofarads -'0.025 centimeters :5 l r, = 0 .l centimeters C’) H 3.0 centimeters J. dielectric constant. a H A = [cl/r." _ U 0” ”‘67:? 1.2 k micromicrofarads. (35) quuation (55) gave such a low value of the heater- cathcmua capacitance that it was measured by the sub- stituidxna method and found to be 9 micromicrofarads. Ilnis thugs problem was satisfactorily solved by the f‘fi’fi 4-2. ma 4-. . o rvm. Okar candle e c nnection Oi :i-ure 18. I?E=£5Q¢%K7afins. ”5' Figure 18 7' .7_1'LS—\‘3M run“ O (u. v 35 he gain of the circuit of Figure 16 was of thecnfiarof 4 at 100 Kilocycles per second. Under nmnalcnerating conditions _, . I 2‘ 4! e (‘0) main ._ 2’ ‘ — 9'" e, ‘ 9m 2‘. _ or E "‘ 6m}. 1 1-9- - 44" ohms 1‘ in 0.009 ‘1 . As cahnflsted from the circuit constants, £1 should have Thus it was evident been of the order of 8000 ohms. that Operating conditions were not normal and.that steps must be taken to raise the gain. All of the following work was done with the 6A37 grid grounded. First the removal of the 6F63 cathode by-pass condensers raised the gain from 4.5 to 10.5 at 100 Iext the 6AC7/1852 cathode is to 180 ohms changing fir» resistor was changed from 160 oh“ kilocycles per second. the gain from 10.5 to 15.0. The screen of the 6F6G was he\cathode with a 0.01 microfarad next by-passed to jdaper'cxmmdenser, and a 10 millihenr* radio frequency choke was inserted in the 13+ lead to the screen. The in at 100 kilocycles was now 36.8. * naoacitance, C, , by; (TD measured 5a Ntwv a measurement of the shun at; 10C) cyrfiles per second by the substitution method gave 1J2} nricrmnnicrcfarads with the 6F6G screen by-pass and 17,000 micromicrofarads with condenser disconnected, These measurements were the by-pass condenser connected. adieri wfitri the vacuum tube voltmeter connected to tre output. thlly a 25,000 ohm resistor was substituted fbr he aflbe screen choke and, with a 1700 ohm resistor in fiw>6F&}p1ate lead, the gain was measured at 51.8. Since the 6AC7/l5’32 catho- L); e resistor had not bcmipasafl.to this point, a certain amount of negative fbedbwfliimd been present. When this cathode resistor waslnupassed by a 1.0 microfarad paper condenser, the gain rose to 71. The load circuit of the 6AC7/lo52 was of the fi. type shown in rigure 19. Figure 19 Here R = load resistance 0,: s1 unt capacitance z; 123 micromicrofarads. Now 2 -ch. _ R-J'wC;# l. = -' _I+w"€a'9' R Jag; and the absolute value of 2‘ is (37) I£J=VI +30.” . 37 Rearranging leg 38 R== . ( > VI W's" a? Making use of equation (36) __ 60",! __ 7’ _ , 1 lzl — "rm — m -— 7’8900111118 . Substituting; this value cf |£2| , 03:123 micromicrofarads, f' . . and w=2fl‘x 10 radians per second in equation (38) gives (39) R: 9,950 ohms. The load resistance was made up of the 6F6& screen resistor, 6F60 plate resistance, 6A3? grid resistor, and 6A3? plate resistance all acting in parallel and was approximately equal to 9,920 ohms. Thus the amplifier was operating normally at this point. Now, with the 6A3 grid grounded so that no negative capacitance was acting, a set of readings of gadji‘versus frequency was taken and recorded in Figu1wa:20. Next the 6A3? grid was connected to the seccnmhary of radio frequency transformer No. 12 so that the Iwnmctance tube was functioning as a negative capac- A second set of readings was now taken and also itance. These data were then plotted in recorded in Figure 20. Since the low frequency response is not be- i Fiaure 21. l \J cedisidered here, it was not thought adviseable to 21 lfixgaritlmic frequency scale. 38 Frequency in Xegacycles. Soil #12 out. Soil #12 in. 0.1 70 72 0.5 45 45 1.0 31 32 1.5 25 27 2.0 23 25 2.5 19 23 3.0 15 20 3.5 14.3 17 4.0 13.8 17 4.5 l2. 16 Figure 20 It will be seen from Figure 21 that some nega- tive capacitance was acting. When the data on coil No. 12 was substituted in equation (7) of Part I, it was found that the value of negative capacitance at 4.5 meg- a cycles per second was theoretically 407 micromicro- farads if i, of equation (7) was considered to be equal to %% as given in Figure 5. Actually i, was not equal to} but was much smaller due to the reflected impedance of the transformer secondary which was approaching anti- resonance. Thus the acting negative capacitance must be cdetermincd from a rearrangement of equation (37). (4o) c=——i—Ufi"l£a . 3° «diflfl’ Now, if ., AV LT :‘...A‘-N' I. . I a I . 'o A . v s u . o If . I. a ..l. I . . D Willi)..- I, l i ‘11 . Illttan II)..- h‘tIfi‘I A’ I pu‘.» J-QIII-I¥‘I"O ... A! ..l; - .1- V- «1‘ I. 1.0!. 1‘: JO. 0. ‘. ll} ’1‘ V. -Ix'nl.v I II ‘¢l I‘ll‘ ICU .10 II I.“ .' ' . I .0 .II If. a a . . I . I u . . . I I o I .. u . . . . fl . . .. .. .. . .. . . . . . . ~ . .. . . . . . . . .I I I . I Q I I I n I I c I I I I I v I ’ . . o I . o . . . I I I . . . . . I o . I . . . I I I I I I . . . I I ~ I . . I . — I. o . I q . ~. I . v . . I I . . . . v I I . a I . I .D . . . . . .4 . . . . I I . I ‘o.. - . .l~v I I. . .... .I I 4. . I .. I I . .01 .i I u. a ...o I . u . u I. I . I I . I . I I. I . I . . I . I m». ¢.m ._MH»W..m V. .+.¢ Gu;¢ ..V:,.mmfi;uqm&3\a§ sysfiamxx ._ H. . ..H ..ny | » I .. I . . . . . r . . I. ~I . . . I I o I .. I . I I . I I I . . . . . . . I . as .Ih .. Wm _. . _ _ nx . 5 . . . . ( v . I I . P 5. .4 b . a. . . . IL. . . I ... . . . I . ... I I I o . I I .. . .. I I . I I I._ -. I . 0 .— . . . _ o . I . . . I . . a . .. . .7 . I I . .. I ~ .I I ~. . u . I o . I . . . o I I I I I Q I I I u I I a v I I . . . I I . I . . i I . . . «I 4 . u . . I I. ‘ {I . . (I b . u 4 1 . II I . .I . . . o I 1 LII . . . I O u I I I II 9! I . I 7. . ¢~ ! V.l. I .l I... . la 3.. I c cla.. . I I I c I .05 u . .. . . . . . I I . .. . . . . I o .. . . . . . . .v W . ....A I . I .. I I. . . . . I . .| VA ' I 1‘ t I . a I I o .I . II‘olv . . I .. I ... . .I . . . . .. c A I . I I a O U .. . . A... o .. . . . . . . I . . .a.. I. I . I . .t o . I Q a I Q I ... . III Iv . o ..I o I O I II Il‘u» I IQJC’IC \. . II| O Q o . Id: '\ v. .I .. . P I...“ v . ¢ _ I . I . -. .‘. .09.". I I . . I . I ~ . . I. .. I O I. I. . . I. I I. . u . . . II ... ..I. I. w vc. I I. I9.- I I I . v r.. 0.. I I . . . ... . I .. . I. . .... .. .. ~ - ' ’ I 0 ‘ I II “-I' II? # I I I v I to... . .. . I. I L ‘JI O‘Or I. I o . I I . . 0. ~vI . . A .I I. .I . I . . I I- a . . . . I .. . . I v . . I . I v. .o I. .. I . . I. I I _ . a . .. .. .. . o . .. I . o o I ~ I . . . I O _ I INII.M . a .V. . u ..- - - I .- l . I .1 .. .. . O..'l I‘ O: {.80. .‘ ..I 'W II. I OI. . VI . ...IOO ax to 05-1,. 00.19 . . 0 In! 3-...ooo+’.'. I. . ..-. ..II 0 . 0|.0‘I 'lOl‘- O‘l‘u 'h. CID!" ~ I I I I o . . I . I A. . I...I I 'o.~ . . I. . I. . v . . I . I . . . .I I . .. ... I I .I H . I I I .. “\n . I I I o. . . . . I. o . I. ..‘I g . I . I . . I . .. h .I . ... I. I . I v I. . .. I . . . I. .. . oI . . . I I . . . ... . .. .I I . . . I u A I . . . I I. I . I . I . . . . . . . . . . ...”... fig. .b.«_.Q.. QQNS . . . .. . . l o . I . . 0 I . . .. l . . . . . I I . I O. . v I I .0. AW 3 I .t I .0 I I I.'.'I I I 9 II I o p.yl.l_v . I 4 .v.0 .0..I.. .|.IJOL v . . . 0.. . . I I. .I' .. . . I. . u . . I . . o. e I . I I . . . . . . . . . e. . . . . . v . .. . . . . I .. .I. I . I o . . . ... u . .. a . .. I I I . I I. . . . I .. ... I . o a .II . .7 I I I I . . o .. v.0. o...-v . .. I I. . I _ . _ I l . I o I . I I. .. . . . . . .. I I o . — . .0. I o . . . . 1.. lo I . -.l w .. .001. I l O..|’ G .3 -I.‘ II. 1": .-v ‘V 10" ‘..'-b . . I ..... . I. . . .. _ . . . . ..I‘ ...;I I 1 (I..- I U‘ . . \ I I . . I I . ... .. I . . o . I. . I . . . . . . . I 0.067 . OI ”Q I o ... to .. .... .o 0 . v.6 tl.’ .1 O... IQiVI; '1. .. I I .I . §.O..I . I . . . . fl . . n I - u. .I I . . I II. . I I I . I . . d I . . I I ~.. I 5.0;. . . . I . fi . . . .- v . o O I I I I4 I .I . I C I I . . I o .I o .o I‘vII. . . o . . ~ . v I .. . _ _ I I 1 .. A .. ... . I I _ . .. I. . . I. . I. .. . . . .. . ... . .. . . .. . . . I.s. .. 9 . .. . . .. I I . o o ... _ ‘v. . I. . I . . I . I II. v. . O . I I O .I o n I I I . II. I . I I. I I .. . o. .v 't: . v . I I .I I . . . . . . I v w . . . . d. . -. I .1va . . w .. . .. .. . .. . . _ ... . ..... .... ... ‘ I I v .. I I...‘ . . . I. . a . o . . I . ..I ... I I o 0 w I In I. Iv I I\ p F. . I I. o s . v c H u . . O I c. I.~I _ I . v . 0‘ §U Ion . .I. _ I . . . 4.... . V . .I I. I O . ... . . . . I . _ a. . I . . . . . . . I . . . I. .2 ...? I . .. I . . . I I. . ~. I I . .. I . . . V I .. I I . I I .- ..-.... .. a .. a .. . .. . . . . . . . 1 . : . . .. . - t. . . . . . . . . . . . . . . ..L. . D] . . I . . . . . I. a. _ . . .. . . o o .. . o. . .. . I . .. . . w .. .. ...L . . .p- 0 . . . . . I 2. .. . ... . . ... . v.1... ...- . . . l. . . _ .. . _. . . . l . . . . . I .. . ‘ it: c . . . . . _. . g I III . I I. II. I..« I ...I I I I. I I I CI 9‘ I. 'o I .I. I I I ..I I I. o I o. 9L .I. . I a.... . I o . I I . . I. I.“ I . . o , _ . I — II I I . . I I ..O I OI. . . . . . I. o I. I . I on t . W. ”...- [90 a. I. n . I . .. . _ I .I . .. .o a. I . *. szl . . I . I I. I . I I II . o. ..b. 96 0910' .. . I I II . . I o. I I I .I I ' I .I . .I .v I . II 'I I I . I I A0 I I I II. 9“ .oc .~.I.. . . ~ : ..~. . I. . I I .* a... . . ‘fla. _ I. . A . a. ..... . . H . u 4.... _ ....u. . .. . . - . .... . . .. . . . . ..L .. ... _ . I I II. . . - (I I . . . I I H .. ..I. . . . . . . . . .. . . . .. . .. .'0.9.v fol .. ..w. . . ... o . . . w . . . . I I . . a... .I I. I. to II I I 9.. . o I I . a I A . . I .I I . .5 I a. I— ‘ .’ ply- . . . . . . I. . I . . . . . . . t .... I' y . I I . I I v . .. . . I .... . . . . . . . . .I . .. v1.01 . 0 I.‘ . o I I: . . .. . o. ...o I I I ‘t I ‘0' I I _ .. . ~5 I I o A . . I On . . . I .. . . . I I . o ...o I . I (.1 I o I .I . ... I II . I. I I I I . .0 o o _ . I. Ola-l . .. I. . . . . I I o I I .I — 4O C,=shunt capacitance with coil out = 123 micro- microfarads Cir—'shunt capacita-.ce with coil in Z,=shunt impedance with coil out at 4.5 mega- cycles per second Zz=shunt impedance with coil in at 4.5 mega- cycles per second R=9,950 01-1115, 3- 1 then g X 6. r-z‘ .2 5 (41) '—‘=: r a: '_L ? Now from equation (56) O or =2fl’.’ -—L‘-?-— = 1250 ohms 9m — max 2: '— 3% 1600 ohms. and 32: 123 VGAJ’f—a‘)‘; (2:: 95.6 micromicrofarads (mar-«2.7 $6 or C, -C,=123 - 95.6:27.4 micromicrofarads. Thus the negative capacitance acting in this case was in the neighborhood of 25 micromicrofarads. Since equa- tion (12) did not aaply in this case where the anti- resonant frequency of the coil was so near 4.5 mega- cycles, the value of the negative capacitance varied according; to equation (7) aurith frequency. I Coil No. 12 had the highest mutual inductance of any of the coils which anti-resonated above 4.5 41 megacycles per second, so that 2Y.4 micromicrofarads was the highest value of negative capacitance to be expected from this circuit. Evidently the only way to increase the gain of this circuit was to decrease the shunt capacitance of the circuit. A new circuit and construction layout was thus adOpted as described in the next section. 3. Construction and Operation of the Second Circuit. The main reason for constructing this second circuit was to decrease the shunt capacitance to ground. This meant that the whole circuit had to be compacted, and in particular the high potential radio frequency leads had to be made as short as possible. With this in mind the circuit of Figure 22 was constructed on a metal chassis. _ - - H J. CF]: E3: fiq' ‘— 5A6 7 1 fl? l i '1 1 . . Io; '5; 4_ e“ M67 [—23.— » >5" C7T F r ‘5";- v AA AAA v vv'v AAA“ "' AAA; ‘vvvv Q I AAA“ AAA‘A ‘ L. JR: E§—-5ufln&C-———fi Figure 22 42 In Fi ure 22, (H R,=l megohm Ra=12 7 Ohm S R5=203,500 ohms R§=S0,0SO ohm R,=100,000 ohms Rf=42,4OO ohms 33:43.2 ohms T as determined from Figure 15. Cf=Cf=l.O microfarad, paper C‘=C,= C4=C‘=C‘=O.l microfarad, paper. It will be noted that the separate 6AG7 grid resistor used in the circuit of Figure 16 was replaced in the circuit of Figure 22 by the alternating current plate resistance of the 6F6G. In order to do this a new set of operating conditions had to be developed such that the plate resistance of the 6F63 would be approximately 25,000 ohms, while the tube still passed‘ the 45 millianperes direct plate current of the 6A37/ 1852 and 5l37. Also it was necessary to keep the 6F6G screen resistor as large as possible because the con- denser, C, , of Figure 22 maintained the screen at the same radio frequency potential as the 6107/1852 plate. Experimentally the following operating conditions were found to satisfy the requirements stated above: 43 Heater voltage Plate voltage Grid voltage (control) Screen resistor 6.3 volts 250 volts 3 volts 50,000 ohms Plate current 35.5 milliamperes Grid current 5.6 milliamperes Screen current 5.9 milliamperes Plate resistance 25,000 ohms. To obtain the positive 9 volts on the 6F6G control grid, the voltage divider composed of R‘s-and R‘ in Figure 22 was devised. As a special precaution against shunt capacitance, the output lead from C‘ was brought out through a hole in a mica sheet centered in a one inch hole in the metal chassis. The shunt capacitance of the video amplifier now measured by the substitution method was 42.7 micro- microfarads. Added to this, of course, was the input capacitance of the vacuum tube voltmeter used to meas- ure the output. Thus the total shunt capacitance was 54.7 micromicrofarads as compared to 125 micromicrofar- ads for the circuit of Figure 16. No adjustment of the circuit constants was necessary because the direct current voltages were correct. With the 6A3 grid grounded the gain of the amplifier at 100 kilocycles per second was 109. When the 6A 7 3rid was connected to coil number 12, the circuit went into violent oscillation with an output voltage ”reater than 16 volts. On a receiver using a beat frequency oscillator, the oscillation was heard as a broad siznal at 2.3 megacycles. ihe low frequency involved and the broadness f the signal in- dicated that the primary of the radio frequency trans- former number 12 with its associated circuit, the 6F6G plate resistance and the shunt capacitance of the ariplifier, was anti-resonant. Various methods were employed to st0p the oscillation all of which so altered the circuit param- eters that the equations of Part I no longer applied. The same trouole was eicountered with coils number 8, 9, l9, and 20. No osci;lation occurred when the 6A3 7 grid was connected to coil number 10.. Readings of gain versus frequency, with the 6A? 7 ii rst connected to ground and ‘I then to coil number 10, were recorded in Figure 23. Frequency (megacycles)—Coil No. 10 out-Coil No 10 in 0.1 105.5 ' 107.5 0.5 45.5 ~ 58.5 1.0 24.5 32.5" 2.0 13.2 17.0 Figure 23 (concluded on next page) ‘T.«:_..sco "-7.?“ 45 ‘ Frequency (megacyclos)—3011 £0. 10 out-Coil No. 10 in 3.0 8.5 11.2 4.0 6.4 0.45 4.5 5.6 7.52 Figure 23 (nncluded) The data of Figure 23 is plotted in Figure 24. From the data of Figure 23 it will be seen that some negative capacitance was acting. Using equations (38) and (36) fimh R== 2*” veers-6w At 0.1 megacycle per second with Cf=54.7 micromicrofarads R =12 , 710 ohms . Let 0,: shunt capacitance with coil out = 54.7 micromicrofarads C,==shunt apacitance with coil in z'=shunt impedance with coil out at 4.5 mega- cycles per second ai==shunt impedance with coil in at 4.5 mega- cycies per second. From equation (36) - 5.6 _ I, ’D 1 z! - 0.“, — "vi—2 0-3318 — . a — Ci.‘ - ’\‘ Y' Z§-—-éefii —— o35 ohms. K I 1 O O C . ikdow utilizgnfi e_uat10n (41) 46 .0V —. vu‘.’ ...-i I:.. 1 but. .1 ., .1". ‘13 DE. . .4.k1 .1 .l‘ 1!. I}. v.1\l| I. 1-1. r .. OI Iv. III 0 . . It: 111. aq. .. I00 . in. 14...... A 1. - . . (1|! llub ‘ ' ...Jl‘o‘fi. J 1.11.1 .‘x l!‘1\1.. 1 1. ¢I1 111! ‘1‘ . . . .. . , . . . . _ , M . . _. . , . . . . , .— ., . _ . . . ., . . . . . L . . _ . ._ _ . . _ VN Nashua. . V . . .. H . . . . . _ 1 . . . .. . — . . _ . , . a n _ e .. o . . ._ ,. . nu 30¢qu .3. sutuavmkx . r a .. . .. . . . . _ . . . . . . 1.. 1 . . . . . . . .. . . . h . _ . . _ s- 1 . , . a . . . . . . . . . a c . . ,_ .1 . . . .. o .. . . . . w . . v. . . . . o . ..c a . 6 . 4. . . H h - . A i .0 1. IL»! u 177 . w - L 0 .7 - — p. . , .u . . ¢ A. . a ‘ . 1V» . a . . . . . . . . . . . e . I . . .. . w _ T , . . . . .. . “u .; .... . _ A H . 1 .. . .... — .. 1 . ... y . . . . _ . . . . , . 1 l . . .1 . A .. a 1. T . . l .. v . . - . . 1. 1.. v. .1 ..L . A 1 1:1. . p 9 . no 0 1. . L . . . .. .1 1 . . 'a u i .3 u , .. i. . ..oA. .... \ .11. l.Lu». ,- a 1. .. . a, . $ 0 A . . ...__ 1 ..-: 'l 1 1. - é +. ‘ t. n .I -l 1 a . I-. '- «1 I . —a ‘ t . . ---.4h<*—&-.AAL‘~..‘ - w o v. -4 .Io O I .. .. 9 . . ‘ ¢ ‘ —‘ D. .5 . _ .. . .. . . u .1 n. . h . . . . . . . . . . . . o a. . a .o . . .1 . . .a .. . I . 1 . v D . b . .4. 1 . . . 1‘ .... .. 1. . . . . . .. . . I s .... .. -~ 4... . w Ono . _ — v .‘ . _ . .. _ . . an. a. .. . . . .. 1 . n . . u. .. . 1 . . , ~ , . . u. . . v I; . a 1 L 1 v o .I. l .. ‘ .I . . .o A .1 .P ~. . u 1:. u. o .1.! u . - . a . .1. 31.. .t1. 1. -1. .v . o 1 . ... . . ., ._ . . . . .. , . l , . .1 . o . .. .. ~ . .H . u . . . . . . . c .r. , . , , . , .. 4 . .... . . . . v . . . . ..I, . . I. . e , \ . , . .... . .... 1 1 . .. .. . .. . . ., — 1. . . . .. . . 11 .. . . o .A , .. w v. a . 4 a . . . , . ... . l n .1 v1 . ,. .~ . . . . . . ,.. a . .v . .. . ., o a. . . .. .... ... r .. . . . . . . .... o. . o ‘ n F , . . . . . . .M . . _ so hawk. ,rkxiéi. .E..mm§§. . . . . . . . J1 . , , .. . . . . . a .. . , . . . . .. , 1 . _ . . . I '1 1 .. I ~ 11 . 9 £ . .. I ..., . . .. .h ... ¢.I . ...: ...v. . u . z . 1'1 . v..§.1 .wlt . c .t .1. ‘4. ., , vi: . _ « , . . . . .. . . . u . . u . . . . u . u . . 1 . . . . . . . . , . _ . _ ; . . . . .. . u o I . v. .. .‘ . . . . c . . . o l ..a.. u n u ‘ . . . . .0 . 1 . . 1 .2 . ~ . . . , . . q . a . . o . , 1 . . . . y . . . . . . .. T. . .. a . . 1 v .. a . . 1. U . . or u . . . .. . . ..¢ . _ . . . _. ‘ ..N . . . . 1 . ‘ _ . H . . . . _a g . . . c . . v . v . u . . . , . . . . . o . . o v _ .1 i . . .. _ 1. 9. 101 1 ‘11t . . . v . 1 .. 4 1 . . .. 1 1 . - a 1 M . . _ . . . . . _ . . .1. . . . fl . _ . ... . . a . . rm. . . . . . , . . . _ .. .. r ., -1 ~ 2 . . . , . . H . . ...- .—-.-| -ve‘.o - A L , . .. . . . . l . .. . ... M . . . . .. .. .. .o . . . . . . i. , o. v g y. ..~ . v . . . .. ~, . % oo‘ '4»; .. . '11 -.rv . . . . , .. .g I. . . q 1 u . c v. I . 4... v: . . __ _ n. 1 0Q . A. . . a . . 1 v. . . .. . .. _ . . . . ., u .. _ ... .5... .. . . . . _ _ ,. a . . a . . u . 1 . . s v . . .c .0 .. . . . A o o . .. v . . r .c . .I . .1 , .. . . . v . . .n . . . ,. _ .... , ._ . . . ... h. . . . .11! . 1. .1 . . . ..l ‘r. . l2: 1 .1 . .t 0 ..t . .. . ~. .. .I . . L.. I. 1. i 1 . . I" .p‘. 1 1 . . . 47 A" - 2" , 2.71 02-: c, a ', i=5é+.7 0 )5 (8”)! §~§§ =40. 8 R - a, 2: (12. 71) — (622)!— micromicrofarads. or (42) 02—0,=-54.7+ 40.8 =-—13.9 micromicrofarads. Equation (42) gives the value of the negative capacitance. Another way to find the negative capacitance is to make use of equation (12) of Part I. Thus .3 (43 11.. 9mM__o.o/x.29xu ——12.6 *V—_ R ’ 23,000 licromicrofarads. These two values of ne3ative cagacitance, found by two different methods, check closely. Fi3ures 25 and 26 are two views of the apparatus used in the experimental work. In FM 3ure 25 the unit at the left is the signal generator, the unit at the right the vacuum tube voltmeter, the amplifier is in the center foreground, and the power suppl ground. The larfe glass tube is the 6F6G, the small metal tube at the right is the 6AC7/1E52, and the metal c+ tube at the rear is he 6A37. To the left of the 6A37 n. F" (.0 cf- 0 O O H. ...: L3 1”) H (D H D; O Si) Fi3ure 26 sh :ws a close-up of the coil assembly with the shield can removed. The coil in position here “I." W “‘l - . 52.3 _ 33. \ m3?” 50 Conclusion It was found impossible to completely neutralize the, positive shunt capacitance of the video amplifier. fl The: concept 0: _._,,__, IJ-F—_ - ne3ative capacitance was introduced, hohwever, and this negative capacitance was actually usexi to neutralize a pa t of the shunt capacitance. The; actual value of negative capacitance is given by 5 animation (12) of Part I. This equation shows that, to imuzrease the negative capacitance, the mutual inductance snuat be increased. To increase the mutual inductance meajis to increase the primary and secondary inductances whicfli'means that both primary and secondary will anti- resmniate far below 4.5 megacycles per second. Thus the main obstruction to the successful opeuvation of the amplifier was the design of coils haviiig a high mutual inductance and still anti-resona— ‘bhtg ahove 4.5 megacycles per second. If the mutual irmhxztahce was increased too far, however, oscillation broke