I-X —~, LA RCI )6:QRr''Vg OF Nr-ipsTEGNf CA, Teebrcl Memoradu NSo. 4 2262&mi.674V age 17 S4et'hes acmaylgepha o h'aW F I a ik ad 19, shoulda be t SpoCeb

ENGINEERING RESEARCH INSTITUTE THE UNIVERSITY OF MICHIGAN ANN ARBOR THE DESIGN AND MEASURED CHARACTERISTICS OF 450 MC POWER DISTRIBUTED AMPLIFIERS Technical Memorandum No. 48 Electronic Defense Group Department of Electrical Engineering D. Hamburg Approved by",_ B. F. Barton Project. 2262 TASK ORDER NO. EDG-1 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 NOo 194B July 1957

TABLE OF CONTENTS Page LIST OF ILLUSTRATIONS i i ABSTRACT v 1, INTRODUCTION 1 2. GRID AND PLATE LINE CONSIDERATIONS 1 2e1 Number of Tubes for a Flat Gain Characteristic 4 2.2 Terminating Networks 4 2.3 Some Notes on the Physical Layout of the Amplifiers 10 2.4 Tuning Procedure 11. 2.5 Power Measurements 15 3, CONCLUSIONS 23 REFERENCES 27 DISTRIBUTION LIST 28 ii

LIST OF ILLUSTRATIONS Page Figure 1 Constant-k Line 2 Figure 2 Variation of Equivalent Shunt Resistance of a 4X150A Caused by Lead Inductance and Transient Time 5 Figure 3 Constant-k Prototype and M-Derived Terminating Sections 5 Figure 4 Grid Line m-derived Terminating Section 6 Figure 5 Plate Line m-derived Terminating Section 6 Figure 6 Maximally Flat, Low Pass, One-Ohm-to-One Ohm Matching Network with a 1 rad/sec Bandwidth 7 Figure 7 Maximally Flat, Low Pass, 50-Ohm-to-50 Ohm Matching Network with a 1 rad/sec Bandwidth 8 Figure 8 Maximally Flat, Low Pass 50-Ohm-to-50 Ohm Matching Network with a 500 mc Bandwidth 8 Figure 9 Maximally Flat, Bandpass, 50-50 Ohm Matching Network with a 500 mc Bandwidth 8 Figure 10 Alternate Forms of a Maximally Flat, Bandpass, 50 Ohm-to-60 Ohm Network with a 500 mc Bandwidth 9 Figure 11 Maximally Flat, Bandpass, 50 Ohm-to-30 Ohm Matching Network with a 500 mc Bandwidth, 10 Figure 12 Photograph of 4X150A Distributed Amplifier 12 Figure 13 Close-up Photograph of a Portion of the Plate Line 12 Figure 14 Photograph of Underside of Amplifier Chassis) Showing Grid, Filament, and Screen Lines 13 Figure 15 Close-up Photograph of a Portion of the Grid Line 13 Figure"16 Block Diagram Used in Plate Line Trimming 15 Figure 17 Block Diagram Used in Power Measurements 16 Figure 18 High Frequency Portion of Low Level Gain vs Frequency Characteristic of 4X150A Distributed Amplifier 17 iiil

LIST OF ILLUSTRATIONS (continued) Page Figure 18 High Frequency Portion of Low Level Gain vs Frequency Characteristic of 4X150A Distributed Amplifier 17 Figure 19 High Frequency Portion of Low Level Gain vs Frequency Characteristics of 4X250B Distributed Amplifier 17 Figure 20 Peak Output' Power vs,Input Power for 4X150A Distributed Amplifier with Duty Cycle as a Parameter 19 Figure 21 Peak Power Gain vs Input Power for 4XI50A Distributed Amplifier with Duty Cycle as Parameter 20 Figure 22 Peak Output Power vs Input Power for 4x250B Distributed Amplifier with Duty Cycle as Parameter 21 Figure 23 Peak Power Gain vs Input Power for 4x250B Distributed Amplifier with Duty Cycle as Parameter 22 Figure 24 Peak Power Gain vs Input Power for' 4i150B Distributed Amplifier at 270 mc with;- Fixed Screen Voltage and 1/4 Duty Cycle 24 Figure 25 Peak Power Gain vs Input Power for 4x250A Distributed Amplifier at 270 mc with-Fixed Screen Voltage and Duty Cycle 25 Figure 26 Schematic Diagram of High Frequency Distributed Amplifier 26 iv

ABSTRACT This report presents the results of an investigation by the Electronic Defense Group into means of increasing the frequency capabilities of power distributed amplifiers. The design, physical construction, and electrical characteristics of a 6-tube 4X150A distributed amplifier and a 6tube 4X250B distributed amplifier are described. These amplifiers demonstrate experimentally the validity of the theory of distributed amplifiers using dummy constant-k line sections between tubes. The amplifiers have upper cutoff frequencies in the neighborhood of 450 mc and useful output power capabilities. The output power capability is, however, a function of the duty cycle at which the amplifier is operated.

The University of Michigan ~ Engineering Research Institute THE DESIGN AND MEASURED CHARACTERISTICS OF 450 MC POWER DISTRIBUTED AMPLIFIERS 1. INTRODUCTION Power amplification in the frequency range above that in which conventional power amplifiers are able to operate is a subject of continuing interest in the countermeasures research area' Conventional distributed amplifiers using presently available tube types are limited to upper cut-off frequencies in the neighborhood of 300 mc. A novel method of overcoming the upper cut-off frequency limitations of conventional distributed amplifier circuits through the use of dummy grid and plate line sections was investigated' theoretically by Dr. P. H. Rogers shortly before he terminated his EDG activities, The present paper gives the design and operating characteristics of a 4X150A distributed amplifier and a 4X250B distributed amplifier which were designed using the principles outlined in Reference 1. 2. GRID AND PLATE LINE CONSIDERATIONS The problem of raising the upper cut-off frequency of a distributed amplifier resolves itself into that of raising the upper cut-off frequency of the grid and plate lines with fixed tube capacitances and inductances appearing as elements of the lines. The plate lead inductance of the tubes discussed in this report (4X150A and 4X250B) is sufficiently low so that in the operating frequency range it can be assumed that the tubes present to the plate line only shunt capacities equal to the output capacities of the tubes. These tubes present to the grid line a shunt reactance which can be satisfactorily

The University of Michigan ~ Engineering Research Institute approximated over the operating frequency range with a circuit made up of -the series combination of grid lead inductance and input capacitance. The ch.aracteristics of the m-derived lumped-constant lines used in the grid., circuit of -the amplifiers to be described, and in which the fixed tube cons-tants dlescribed above appear, are anacly-zed in Reference...l' The analysis suggests that one should u.se negative mutual coils in the grid line to-cancel or, if it were possible, to even overcompensate for -the grid lead inductance. Actually, with maximum attainable coefficients of coupling (approximately o.6) for the small air core coils used, the maximum bandwidth is obtained by just cancelling the gri lead in. c d c-;.::Linc. This cancellation would result in a constant-k line s.cih s is shown in Figure 1. In this diagram, Ct is the tube input L +,Vi L+M Ld L+M L+M L+M -M= -Lg -M=-Lg d= Ct Cd=Ct Cd= Ct _L Cd=Ct ct Ct Fig. 1 Constant-k Line capacitance, L is the grid lead inductance, L is the self-inductance of the g coils which are coupled so as -to reflect a negative mutual inductance equa~ l in n ma.gnitude to L into the grid lead, C. is the capacitance of'the clammy cons'tantg k sections, and L is the inductance of the dummy constant-k sections. If kt~~~~~~~~~.L~~~~~~~t the d.h.tP.}.t')jy se.:i,:._n:as we-re n,.'.k-lesent the capacitance per section would be Ct and the inductance per section would be 2(L + M), resulting in a cut-off frequency of = _ 2 (1)

The University of Michigan ~ Engineering Research Institute For the circuit of Fig. 1 where dummy sections are used, however, the inductance per section is L + M and the cut-off frequency becomes Cc 2 (2) (/L + M)Ct Thus, through the use of dummy sections one can increase the upper cut-off frequency by a factor of 2.f2 for a fixed magnitude of the shunt capacitances. The grid lead inductance L of either the 4X150A or 4X250B tubes is approximately 0.007 microhenries. With a coefficient of coupling between the L's of 0.6, then L L = k - 0.012 microhenries (3) The input capacitance of either a 4X150A or 4X250B tube with a hot filament is in the neighborhood of approximately 20 micromicrofaradso The maximum realizable cut-off frequency with these tube parameters is approximately 515 mc. The amplifiers described later were constructed using the above values as starting points, giving for the grid lines a nominal characteristic impedance of 30 2. In the plate line there is no problem of cancelling lead inductance for the bandwidths achieved, and the realization of this line is simple. The main requirement for the plate line is that for a given grid line the phase shift between plates be the same as the phase shift between respective grids. This means that the cut-off frequncy of the plate line should equal the cutoff frequency of the grid line, and, in addition, the number of constant-k sections between tubes should be the same in both the grid and plate lines. The output capacity of either the 4X150A or 4X250B tubes with shielding is approximately 10 micromicrofarads. The cut-off frequency of the plate line is given by c (4) c L C

The University of Michigan ~ Engineering Research Institute where L is the inductance of a plate line section and C is the output p 0 capacity of the tube. For a cut-off frequency of 515 mc and an output capacity of 10 micromicrofarads one finds that L - 0.0382 microhenries (5) P Again, the amplifiers described later were constructed using these values as starting points, giving for the plate lines a nominal characteristic impedance of 60 Q. 2.1 Number of Tubes for a Flat Gain Characteristic P. H. Rogers has shown that by the proper choice in the number of tubes, grid line losses due to tube loading can be made to compensate for the rising impedance characteristic of the mid-shunt grid and plate line seations so as to result in a flat over-all gain characteristic for an amplifier of a given bandwidth. This proper cholice is found from Eq 6. nw - G. (6) c G~. in where n is the number of tubes for a flat gain characteristic, oc is the cut-off frequency of the amplifier, C is the input capacity of the tubes, and g G. is the conductive component of input admittance of the tubes. For the in 4X150A, G. is determined from Figure 2. Using G. f2 /2260 and C = 20 in in g micromicrofarads, one finds that for c = 515 mc, n = 5.5. It was decided to c use 6 tubes in the 4X150A amplifier to be fabricated. Because of the similarity in input admittance characteristics between 4Xl50A and 4X250B tubes, 6 tubes were also planned for a 4X250B distributed amplifier. 2.2 Terminating Networks m-derived half-sections were used to match the constant-@ lines to the terminating loads. The resulting circuit is equivalent to Figure 3. __ ___ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ___ 4b __ _ _ _ _ _ _ _ _ __ _ _ _ _ _ _ _ _

=9 1-L 1 31 61-IS-V 1 1 1' 4000 3000 RI N 2260 2000 f2 f = FREQUENCY IN uLJ HUNDREDS OF MEGACYCLES. W 800 Z) Soo 600 400 z 300 200 100 20 30 40 50 60 708090100 200 300 400 500 1000 FREQ. (MC.) - FIG. 2.VARIATION OF EQUIVALENT SHUNT RESISTANCE OF A 4X150A CAUSED BY LEAD INDUCTANCE AND TRANSIENT TIME. LK LK LK LK Li L2' LK 2CK, ^2CK 5 R=/ CK T2 FIG. 3. CONSTANT-K PROTOTYPE AND M-DERIVED TERMINATING SECTIONS. 5

The University of Michigan ~,Engineering Research Institute The elements of the m-derived section are determined by the elements of the constant-k prototype; L1 = mk (7) (1 - m ) (8) C2 = mCk (9) A value for m of 0.6 was chosen for these terminal half-sections, which results for the grid line in the element values of Fig. 4. [ 1. 30 fl Fig. 4 Grid Line m-derived Terminating Section For the plate line, the terminating half-section of Fig. 5 results. ____._ I.0. Fig. 5 Plate Line m-derived Terminating Section The basic design was modified to facilitate the solution of two practical problems. The first of these problems is associated with supplying dc power to the plates and grids of the tubes. The very desirable shunt feeding of these voltages is best accomplished if a shunt inductance is present in the lines. A second problem results from the undesirability of the magnitude of the grid line terminations (30 ohms) which resulted for the experimental amplifier. It is often desirable, both for experimental work and subsequent applications, to be able to use commercially available 50 ohm power terminations and measuring equipment. 6

The University of Michigan ~ Engineering Research Institute It is possible to solve both of the above problems over a bandpass frequency range by insertion of suitable bandpass matching networks between the plate and grid lines, and their several terminations. Since the unique feature of the amplifiers under consideration is their ability to operate in the frequency range 300-500 me, this frequency range was singled out for experimental investigation. Bandpass matching networks were therefore chosen with 3 db points at 100 mc and 600 me, resulting in very moderate mismatch in the 300-500 mc range. The mean frequency of the 100-600 mc band is approximately 245 me, and is therefore the "center frequency" of the networks. Background material necessary for the choice of suitable networks is widely available in the literature of modern network synthesis, and will not be presented here. However, the generation of the specific networks will be exemplified. Suitable bandpass networks can be derived from the low pass prototype network of Fig. 6. This three-pole one ohm-to-one-ohm matching network has a maximally-flat transmission characteristic, and a 3 db bandwidth of one radian per sec. A suitable network is obtained by performing impedance level, bandwidth, and lowpass to bandpass transformations on this prototype3. E2 T if if 1 Fig. 6 Maximally Flat, Low Pass, One Ohm-to-One Ohm Matching Network with a 1 rad/sec Bandwidth. Raising the impedance level of this circuit to 50 ohms yields the circuit of Fig. 7. This is accomplished by increasing all impedances by a factor of 50 at a fixed frequency. 7

The University of Michigan ~ Engineering Research Institute 100 h 50 _.02 f 02 f 50 O Fig. 7 Maximally Flat, Low Pass, 50 Ohm to-50 Ohm Matching Network with a 1 rad/sec Bandwidth. Increasing the bandwidth of the network of Fig. 7 from 1 rad/sec 6 to 500 mc by dividing all reactive elements by 27 x 500 x 106 results in the circuit of Fig, 8..0318 Lh 50 Q 6 36 rf 36 f 50 Q Fig. 8 Maximally Flat, Low Pass 50 hm-to-50 Ohm Matching Network with a 500 mc Bandwidth. The required bandpass equivalent of the circuit of Fig. 8 with the same 3 db response bandwidth is determined by parallel-resonating all capacitors and series-resonating all inductances at 245 mc, This results in the circuit of Fig. 9..0318 kih 13 yif 50 1 6.36 4f o066 ih 6.36 4f't o066 Oh 500o Fig. 9 Maximally Flat, Bandpass, 50 Om-to=50 0Ob Matching Network with a 500 me Bandwidth. 8

The University of Michigan ~ Engineering Research Institute Finally, for the plate line where the characteristic impedance is 60 ohms, the network should provide an impedance transformation between the 60 ohms and the desired 50 ohm impedance level of the termination. This is accomplished by inserting an ideal transformer into the network in accordance with the principles of Ref. 3. It can be shown that the desired ideal transformer should have a turns ratio equal to the s4uare root of the desired impedance transformation, or.\60/50. The resulting network is shown in Fig lOa1; which has an alternative realization shown in Fig. lOb..0318 ph 13 Off 1: 6 so50 I5636L~ I066 IhT3 h'tf.079 ph Ideal Fig. lOa.0318 ph 12.1 Attf 50.3.066 ph 1.15 o f 6 - Fig. 10b Fig. 10 Alternate Forms of a Maximally Flat, Bandpassn 50 Ohr' c>-6O Ohm Ne+twork with a 500 Mc BEandw~idth. 9.

The University of Michigan ~ Engineering. Research Institute Now) the lower end of the 0.079 microhenry coil is by-passel co gro-:nc:i wi. th a feedsthrough capacitor and' B+ fed to tLhe plate line throu. gh h.is coil. A similar design procedure was used for the grid line, except that here the grid line characteristic impedance is approximately 30 ohms so that the terminating network in this case is designed to match from this impedance to a 50 ohm input cable. The circuit diagram of the grid line matching network is shown in Fig. 11. I I."L 20G;~ E-h.. 17! t1a.f -.01]91 Ah 50Q.o66 h 21 f f 1.6f aK97 30W' 1 7 1000 [IlIzf Fig. 11 Maximally Flat, Bandpass, 50,Ohm-to0-30 Ohm Matching Network with a 500 mc Bandwidth. The above described grid line matching networks were used in both the 4X150A and 4X250B distributed amplifiers. However, the above plate line matching network was used in only the 4X150A distributed amplifier; The mismatch loss between a 60 ohm line and a 50 ohm load is only 0.04 db, and is negligible. The 4X250B plates were shunt-fed through an inductance across the reverse termination, the low RF voltage end being bypassed with 1000 imtf The inductance was chosen to resonate at 100 mc with the stray capacitance present at the 50 ohm reverse termination. Satisfactory frequency characteristics were realized with both types of feed...2.3 Some Note'o:.,n t1ve Physi' clLayouti.o f: te. A..Ariplifiers In the construction of the amplifiers based upon the above considerations, preliminary models of grid lines fell far short of realizing the 10

The University of Michigan ~ Engineering Research Institute theoretical cut-off frequency. After a large number of possible difficulties were inve-stigate-d, with no appreciable success, t was fi.nally hypothesized that perhaps the chassis itself could not be considered a zero impedance structure and that the line currents returning through the chassis inductance could cause sufficiently large voltage drops to affect the behavior of the line To test this hypothesis, a six tube line was constructed with four dummy sections per tube (so as to make the chassis.inductance per section very small)O In addition9 the line was built circular so as to bring forward and reverse terminations in juxtaposition and located in the center of the ring. Circular construction resulted in a short chassis path through which line currents flowed back to the input, and thereby minimized the return impedance for line currents. The circular construction with a relatively large number of d=umy sections resulted in frequency characteristics not far removed from the theoretical. Photographs of the physical construction are shown in Figs. 1!2 15. These photographs were made of the 4X'50A amplifier~. However, the 4X250B amplifier is almost physically identical. The unique tube socket arrangement might be noted froam Figo. 15. This arrangement was used to decrease the lead inductance and stray capacitance below that realizable with conventional tube sockets. All grounded pin receptacles were soldered directly to the chassis. 204 T'rning Procedure Theoretically, if one could fabricate all the elements of the grid and plate lines and the terminating networks so as to be pure capacitances and pure inductances with the designed values then by defi.nitior no adjustment of the elements would be necessary. However9 of course, this is not possible' and with the large number of constantlk sections resulting from the use of the dummy sections and the requirment that the signals arrive in phase at the 11

LS-91-L 03ti III-ld-V Z9?Z...................................................... ggg................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................... I.................................................................................................................................................'I............................................................................. I.,.....",....................................................................................................................................................................................................................................................I............................................................................................................................................................................................................................................................................................I.......................................I - -.1.........................................................................................................................................................................I.................................................................II.............. I —I.............................................................................................................................................................................................................................................................................................................................................................................-..............................................................................................................................................................................................................................................................................................................-...............................................................................I..............................................................................................................................................................................................................................................................................................................................................................................- I..''...........................................................I............-..........................................................I......................................................................... -....................................................................................................... I....,..",...,..."......,.,.....,.......,....,.,..........,...,.........................:::....................................................................................I.............-................................................................................................................................................... FIG. 12. PHOTOGRAPH OF 4XI50A D1517RIBUTED AMPLIFIER. THE. STRUCTURE MOUNTED ABOVE THE, TUBE, IS AN EXHAUST TYPE BLOWER SYSTEM. THE GRID LINE IS ON TH EUNDERSIDE OF THE%' C11ASSIS, WHILE THE Pt-ATE LINE. IS VISIBLE ON TOP OF THE CHASSIS................................................................................................................................................................................................................................ I....................-................................................................. -.1.............. I..............I I.........................................................................................................................111.1........................................................................................................................................... I..............I.................................-........................................................................................................................................I........................................................ 11................................................................................................................................ II.................I........................................ I...............................II....................................................................................................I11.............................................................. -............-.......................................................................................-......................................................................... I'll, 11..........................................1.11......................................................................................................................................................................................................................................................................................................................................................................................................................................................................................II.............................................................................. 1.1- 1- 1- 1.11.1............................................................................................................ I........................................................................................................................................................................................................................................................................................................................................................................................................... S..............................................................................................................................................................................................................................................................................................................................................................................................................................................................................I......................................................................I....................I........... I..,...........-............................................................................................. 11 I........................I I..............- I I I.:.:,.:: -:-:-:.:::..-..................................................................................................................................................................................................................................................................................I.................................................................................................................................................................................................................................................. I.,.............................................................................................................................................................................................................................1.1..........................................................................................................................................................................................................................................................................................................................................................................I..............................

'i~~~~~~~~~~~~~~~~~k F.. P TP. 5....CO..T.. M........................EN..................................................................E........ 111.......................................... N M............ ~ ~ ~ ~.................. FIG. 14. PHOTOGRAPH OF UNDERSIDE FIG. 15. CLOSE-UP PHOTOGRAPH OF~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~....... OF AMPLIFIER CHASSIS, SHOWING A PORTION OF THE GRID LINE.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~....... GRID, FILAMENT, AND SCREEN LINES~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.......

The University of Michigan ~ Engineering Research Institute plate of each tube so as to reinforce, it is necessary that the relative phase difference between respective points on the grid and- plate -lines be very small. If this is not true, then gain characteristics with extremely large ripple result. Therefore, it was found necessary after final assembly to trim the plate line so as to cause this line to have the same phase characteristics as the grid line. This triiiing results in a smoothing of the gain characteristic. Trimming was accomplished using the block diagram shown in Fig. 16. The procedure was as follows: 1. Disconnect all but the screen of tube number 1 and connect the plate line termination to the plate of this tube. Trim the tube Noo 1 plate line constant-k inductances by moving the shorting bars shown in Fig. 13 so as to produce a smooth one tube gain characteristic. Since only one tube is present, grid line losses will not now compensate for the rising mid shunt impedance characteristic at the high frequency end. Hence, a rising gain characteristic should result with one tube. 2. Connect the screen of tube No. 2 and move the plate line termination to the plate of tube-.NQ. 2. Repeat the above prow cedure, trimming now only the plate line constant-k inductances between tubes 1 and 2. 3. Continue the above procedure until the complete plate line is adjusted. Note that as more and more tubes are added, the gain characteristic will flatten at the high end. The above procedure has been used successfully on both amplifiers. It has been our experience that indiscriminate adjustment is not a convergent process. However, minor judicious trimming by an experienced person after 14

Lg-tl-8 8V OI -IS-V -t9z GR TYPE 1263 A AMPLITUDE REGULATING SUPPLY a GR TYPE 1750A SWEEP DRIVE MECHANISM GRID REVERSE TERMINATION TERMINATION (50s) (5O(2) __ -. F DISTRIBUTED AM- o GR TYPE 1209B PLIFIER BEING SCOPE UNIT OC. (250- I DJUSTED 900 MC) INPUT VERT. HORIZ GR TYPE OUTPUT 874 VR MIXER PLATE TERMINATION a GR TYPE 874 VR MIXER FIG. 16 BLOCK DIAGRAM USED IN PLATE LINE TRIMMING 15

The University of Michigan ~ Engineering Research Institute using the above procedure can further serve to reduce remaining ripple Photographs of final low-level power gain characteristics for the 4X150A and 4X250B distributed amplifiers are shown respectively a-s Figso 18 and 19. From Fig. 18 it can be seen that the 4X150A amplifier has usable gain to above 450 mc (reasonably close to the theoretical). However, from Figo 19 it can be seen that the 4X250B. amplifier has usable gain to only 420 mc. This resulted because the 4X250B tubes available at the time of construction had 20 percent higher input capacitance than the 4X150A tubeso Therefore one would expect an appreciable improvement through the use of presently available low capacity 4X250B tubes. 2'. 5 Pow'er Measurements The power characteristics of the two amplifiersdescribed were obtained using the block diagram shown in Fig. 17. GRID REVERSE TE'MIAMTION TERMINATION POWER(50 DaR) ISTRIBUTED (so ) _ OSCILLATOR AMPLIFIER GIR 874 GA BEING TETED ADJUSTABLE I ATTENUATOR.' WATT - TERMlRATI 0-500 MC. METER (5,0 ) LOW-PASS FILTE.R BOLOMETER MOUNT, Fig. 17 Block Diagram Used in Power Measurementsn

FIG. 18. HIGH FREQUENCY PORTION Of LOW LEVEL GSAIN VS FREQUENCY CHARACTERISTIC OF 4X150A DISTRIBUTED AMPLIFIER. FIG. 19. HIGH FREQUENCY PORTION OF LOW LEVEL GAIN VS FREQUENCY CHARACTERISTICS OF 4X250B DISTRIBUTED AMPLIFIER.'7

The University of Michigan ~ Engineering Research Institute One would not expect the power gain versus frequency characteristic to differ appreciably from the low-level gain vs frequency characteristic. Thus basic power measurements were planned at one convenient frequency with spot checks made at other frequencies to test agreement with the low-level curves. The major parameter of interest -in the power measurements was the duty cycle'. No attempts were made to operate CW. The 4X150A amplifier was operated with 600 volts on the plates, sufficient screen voltage as a function of driving power to result in rated dissipation and a grid gate voltage swing from -80 volts to -20 volts. Previous experience has indicated that these conditions are near optimum. The 4X250B amplifier was operated with 750 volts on the plates, sufficient screen voltage'as a function of driving power to result in rated dissipation, and the same gating voltages as used with the 4X150A amplifiero it is felt that these operating conditions were again near optimum. Figures 20 and 21 are the measured power characteristics of the 4X150A amplifier. Figures 22 and 23 are the measured power characteristics of the 4X250B amplifier. The aforementioned frequency spot checks are shown as partial curves in the -above mentioned figures If one correlates these curves with the low-level gain characteristic, one finds good agreement. The frequency of 270 me at which the extensive power measurements were made was dictated by the inability of available driving oscillator to operate at high power levels near the upper frequency end of the amplifier response. The measurements at 270 me on the 4X250B amplifier occurred near a peak of the gain curve, whereas the measurements at 270 me on the 4X150 amplifier occurred at a mean point on its gain curve, It is estimated from the power curves that the 4XX50B has a mean gain approximately 3 db greater than the 4X150A amplifier, due to its ability to operate within rated dissipation at higher operating voltages. Dues then, to a peak of the 4X250 response atl

ZL-8-L A/M gOt-19-V z9Zz o 270MC A- ----- 375 MC O -' -- 430MC'-DUTY CYCLE 100;DUTY CYCLE X, 80. DUTY CYCLE 60 0.?- 4 040 4 7-DUTY CYCLE 4'DUTY CYCLE 0. 0 0 10 20 30 40 INPUT POWER (WATTS) FIG. 20 PEAK OUTPUT POWER vs. INPUT POWER FOR 4X150A DISTRIBUTED AMPLIFIER WITH DUTY CYCLE AS A PARAMETER 19

LS-8-L KM 90~-II-Y Z9aZ o 270 MC...6 L la-375 MC 6 O - 430 MC t DUTY CYCLE.0 I I 1 ~~ DOTUYCYCLE 0, oL I I I \1 2UTV CVCLE ~ i DUTY CYCLE 00- 10; 20 30 j ~~DUTY CYCLE O 10 20 30 40 INPUT POWER (WATTS) FIG. 21 PEAK POWER GAIN VS. INPUT POWER FOR 4XI5OA DISTRIBUTED AMPLIFIER WITH DUTY CYCLE AS PARAMETER 20

LS-8-L M1H o ~ 270 MC A_____-330 MC - 0- —.415 MC 400 - DUTY CYCLE t DUTY CYCLE 0 200 I DUTY CYCLE 0100 DUTY CYCLE,<u~~ — ~I~ ~DUTY CYCLE 0 0 10 20 30 40 INPUT POWER (WATTS) FIG. 22 PEAK OUTPUT POWER vrs. INPUT POWER FOR 4 x 250B DISTRIBUTED AMPLIFIER WITH DUTY CYCLE AS PARAMETER 21

LS-8-L -, 80~-19-V Z9ZZ o 270 MC -—. —- 330 MC D- -... 415 MC I0 | |" 1 3DUTY CYCLE 7-.7 - DUTY CYCLE z6'""''~ ~DUTY CYCLE zw~~ 6~~.> | 4 >, DUTY CYCLE I N |P \ 4TR UEW_ DUTY C YCLE (WT 0 10 20 30 40 INPUT POWER (WATTS) FIG. 23 PEAK POWER GAIN vs. INPUT POWER FOR 4x250B DISTRIBUTED AMPLIFIER WITH DUTY CYCLE AS PARAMETER 22

The University of Michigan ~ Engineering Research Institute 270 mce the power gain at 270 mc can be seen from the curves to be substantiall above the power gain at the other frequencies checked. It will be noted that the variation in power gain with frequency at 1/4 duty cycle is much less for the 4X150A amplifier than for the 4X250B amplifier, as would be expected from comparing the low level gain characteristics. In equipment applications it might not be feasible to adjust the screen voltage so as to maintain rated tube dissipation as the input power is varied and thus a power gain curve was obtained for eacfh amplifier using fixed screen voltage with rated dissipation occuring at maximum drive available. These curves are shown plotted in Figs. 24 and 25. A complete schematic diagram of the high frequency distributed amplifier is shown in Fig. 26. 3. CONCLUSIONS The amplifiers described in this paper are seen to ihave usable gain and power characteristics, and in addition possess bandwidthsmuch superior to the bandwidths of conventional distributed amplifiers. The wider bandwidth obtained required lower impedance levels in both the grid and plate lines, and thus the gain and power characteristics are inferior to ~those of lower frequency models. This is especially true with the 450 mc 4X150A amplifier, and at low duty cycles there might even be some question of the usability of this amplifier. The usability of the 4X250B amplifier, however, is not so questionable, because of its ability to operate at higher voltages and currents with a corresponding increase in gain due to increased gm. 23 _

LS-8 —L AN 602-19-tV Z92E 0 10 20 30 40 INPUT POWER (WATTS) FIG. 24. PEAK POWER GAIN vs. INPUT POWER FOR 4 x 150B DISTRIBUTED AMPLIFIER AT 270 MC WITH FIXED SCREEN VOLTAGE AND 4DUTY CYCLE.......,~~~~~~

LS-8-L ByH 01~-19-V;9Zz 8 0 10 20 30 40 INPUT POWER (WATTS) FIG. 25 PEAK POWER GAIN vs. INPUT POWER FOR 4 x 250A DISTRIBUTED AMPLIFIER AT 270 MC WITH FIXED SCREEN VOLTAGE AND DUTY CYCLE 25

3 6-I C6-2 C-3 C36-4 C6-5 06-6 C6-7 C6-8 C6-9 C6-IO C6-lI C6-12 36-13 C6-14 C6-15 C6-16 36-17TC-3I C6-19 3-20 06-21 C6-22 {C623 C6-24 C 1-3 I L5-2 V1-1 I V-2 VI-3 Vl-4 Vl-5 Vl_6 C - ELR 2-1 R2- 0 R 2-3 R2-4 R 52-5 R L 6-1 L 6-2 L6-3 L 6-4 L 6-5 R I- R -2 R1 -3 R 1-4 R 1-5 R -6 T4-1 J4-2 T 4-3 F4-4 T4-5 C4-6 TC4-7 C4-8 - - tC4-9 C4-10 TC4-11 C4-12 CI-l L 2-1 L3-1 LR4-1 L4-2 L4-3 L4-4 4- L4-6 L4-7 L4-8 L4-9 L4-10 L4-11 L4-12 L4-13 L4-14 L4-15 4-16 L4-17 L3-2 L2-2 C 1-2 83- 100, 2W, ALLEN RRADLEY LO *'-0A-5 CO4-I CI I L1I L3-1 C3-2 C3-3 CL3-4 C3-5 C3-6 C3-7 C3-8 C3-9 L3- 0 C3-111 C3-12 C3-13 C3-14 C3-15 C3-16 C3-17 3-18 C3-19 3-20 C3-21 3-22 C3-23 C3-24 FIG 26. HIGH FREQUENCY DISTRIBUTED AMPLIFIER Cl- IOOOpLf, ERIE CB LI- 0.01015 Rh COIL C2- 6 /zf, CERAMIC TRIMMER, CENTRALAB L2- 0.0057 /h COIL C3- 19/z/f, 15/z/f ERIE CB PADDED WITH 3/z/f CENTRALAB L3- 0.0095Lh COIL C4- 1000 /A/f, CENTRALAB STANDOFF L4- 0.019 TCh COIL C5- 3.15 ft, CERAMIC TRIMMER, CENTRALAB L5- BROAD BAND RF CHOKE: 5 feet No. 28 GAGE WIRE LOOSELY WOUND TO RESONATE ABOVE 500 MC C6- 11 pFf, lO/z/f NOMINAL ERIE CB L6- RFC L7- RFC RI- 22Q, 2W, ALLEN BRADLEY L8-.019/zh COIL RZ- 470 S, 2 W, ALLEN BRADLEY L9- 0.0095 /h COIL R3- IO-, 2W, ALLEN BRADLEY L1O-0.016 Oh COIL LII- 0.032 /h COIL VI- 4X150A, EIMAC L12- CENTER TAPPED 0.038 /h COIL COEFFICIENT OF COUPLING = 0.6 0.007/h NEGATIVE MUTUAL REFLECTED INTO CENTER TAP

The University of Michigan * Engineering Research Institute REFERENCES 1. P. H. Rogers, "Some Useful Techniques for Overcoming the Frequency Limitations of Conventional Distributed Amplifiers," Electronic'Defense GrOup Technical Memorandum Noo 32, University of Michigan, Ann Arbor, Michigan, November 1956o 2. P. H. Rogers, "Large Signal Analysis of Distributed Amplifiers," ElectrOnic Defense Group Technical Report No. 52, University of Michigan, Ann Arbor, Michigan, July 1955. 3. B. F. Barton,"The Design of Efficient Coupling Networks," Electronic Defense Group Technical Report No. 44, University of Michigan, Ann Arbor, Michigan, March 1955o 27

DISTRIBUTION LIST 1 Copy Document Room Stanford Electronic Laboratories Stanford University Stanford, California 1 Copy Commanding General Army Electronic Proving Ground Fort Huachuca, Arizona Attn~ Director, Electronic Warfare Department 1 Copy Chief, Research and Development Division Office of the Chief Signal Officer Department of the Army Washington 25, Do C. Attn' SIGEB 1 Copy Chief, Plans and Operations Division Office of the Chief Signal Officer Washington 25, Do C. Attn: STGEW 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-EW 1 Copy Commanding Officer White Sands Signal Corps Agency White Sands Proving Ground Las Cruces, New Mexico Attn~ SIGWS-FC 1 Copy Commanding Officer Signal Corps Electronics Research'Unit 9560th USASRU Mountain View, CalifOrnia

65 Copies Transportation Officer, SCEL Evans Signal Laboratory Buiilding No, 42, Belmar, New Jersey FOR B SCEL Accountable Officer Inspect at Destination File No 22824-PH-54-91(1701) 1 Copy J. A. Boyd Engineering Research Institute University of Michigan Ann Arbor, Michigan 1 Copy Document Room Willow Run Laboratories University of Michigan Willow Run, Michigan 12 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

UNIVERSITY OF MICHIGAN I1011111 0 2 71111111111111111111111 11 I3 9015 03027 7001