Technical Note Lincoln Laboratory. A Noise Rejection Filter for Waveguide Carrying High Power 3PY. W. J. Getsinger. 3 August 196!

Size: px

Start display at page:

Download "Technical Note Lincoln Laboratory. A Noise Rejection Filter for Waveguide Carrying High Power 3PY. W. J. Getsinger. 3 August 196!"

Joel Wilkerson
5 years ago
Views:

3PY U A 470 LIST DIVISION- Technical Nte

> Prepared under Electrnic Systems Divisin

1 3PY U A 470 LIST DIVISION- Technical Nte W. J. Getsinger A Nise Rejectin Filter fr Waveguide Carrying High Pwer 3 August 196!> Prepared under Electrnic Systems Divisin Cntract AF 19(628)-5167 by Lincln Labratry MASSACHUSETTS INSTITUTE OF TECHNOLOGY Lexingtn, Massachusetts A ft

2 The wrk reprted in this dcument was perfrmed at Lincln Labratry, a center fr research perated by Massachusetts Institute f Technlgy, with the supprt f the U.S. Air Frce under Cntract AF 19(628>5167.

3 TS4 MASSACHUSETTS INSTITUTE OF TECHNOLOGY LINCOLN LABORATORY A NOISE REJECTION FILTER FOR WAVEGUIDE CARRYING HIGH POWER W. J. GETSINGER Grup 46 TECHNICAL NOTE AUGUST 1965 LEXINGTON MASSACHUSETTS

4 ABSTRACT A nvel design fr a waveguide band-rejectin filter carrying high pwer in its pass band is presented. The resnant cavities f the filter are munted in pairs n the narrw walls f the guide and cupled t the guide thrugh rectangular penings the full height f the guide. Thus there are n edges r radii n which electric field cncentratins can ccur. An X-band mdel having six cavity pairs had a rejectin greater than 80 db ver k^> Mcps in its 6tp band with an insertin lss under 0.05 db and VSWR under 1.02 in its pass band. With water cling, the filter carried ver 300 lew CW withut breakdwn. Accepted fr the Air Frce Stanley J. Wisniewslci Lt Clnel, USAF Chief, Lincln Labratry Office ii

A NOISE REJECTION FILTER FOR WAVEGUIDE CARRYING HIGH POWER I. INTRODUCTION The M.I.T. Lincln Labratry Haystack Facility transmits 100 kw CW at 7750 Mcps while receiving at 8350 Mcps using the same antenna.

This filter is placed in the transmitter waveguide between the generatr and the diplexing arrangement.

5 A NOISE REJECTION FILTER FOR WAVEGUIDE CARRYING HIGH POWER I. INTRODUCTION The M.I.T. Lincln Labratry Haystack Facility transmits 100 kw CW at 7750 Mcps while receiving at 8350 Mcps using the same antenna. A reject filter is used t reduce substantially the part f the transmitter nise that lies within the receiver pass band and cuples int the receiver. This filter is placed in the transmitter waveguide between the generatr and the diplexing arrangement. It strngly attenuates energy arund the receive frequency while passing energy at the transmit frequency with very little attenuatin r reflectin. Filters placed in very high pwer lines are typically f the absrbing, rather than reflecting, type in rder t avid large electric fields that might induce vltage breakdwn in the structure and t prvide a lad in which the undesired energy can be dissipated. Hwever, fr Haystack it was decided t use a reflecting-type filter fr the fllwing reasns: 1. The resnatrs used wuld be timed t the receive frequency and thus wuld allw nly small vltages at the transmit frequency. 2. The nvel design prpsed had n edges r radii n which electric field intensificatin culd ccur. 3. A terminated high-pwer circulatr lcated at the generatr was available t absrb reflected energy. k-. A reflectin filter culd be made t have higher stp-band lss and lwer pass-band lss than an absrptin filter. 5. A reflectin filter culd be made smaller in vlume and weight than an absrptin filter. 6. The perfrmance f a reflectin filter is mre accurately predictable than is the perfrmance f an absrptin filter.

A reflecting nise-rejectin filter develpment prgram was begun that re- sulted in an interim filter design having a rejectin band abut 8050 Mcps and the final filter design with a rejectin band abut

The 805O Mcps filter has three pairs f resnatrs while the 835O Mcps filter has six pairs, since it was required t have a wider stp band than the 805O Mcps filter.

6 A reflecting nise-rejectin filter develpment prgram was begun that re- sulted in an interim filter design having a rejectin band abut 8050 Mcps and the final filter design with a rejectin band abut 835O Mcps. II. DESCRIPTION Figures l(a) and (b) shw the nise-reject filters fr 8050 and 835O Mcps. The 805O Mcps filter has three pairs f resnatrs while the 835O Mcps filter has six pairs, since it was required t have a wider stp band than the 805O Mcps filter. The waveguide size is WR-137 with inside dimensins f x inches. The flanges are stainless steel CPR-137F» which are preferable t standard UG-3M*- flanges fr high-pwer wrk. The pipes sldered t the brad walls f the filters are fr the cling water needed at high pwer levels. The resnatrs are cupled t the waveguide thrugh penings in the narrw wall f the waveguide. These penings run the full height f the guide s that upper and lwer walls f the waveguide and resnatrs are flush and cntinuus, with n edges r prtrusins that might cause electric field cncentratins. If the resnatrs had been placed n the brad walls f the waveguide, as is usually dne with reject filters, the edge f the pening between resnatr and waveguide wuld have been in the regin f greatest electric field intensity. The electric field wuld be further intensified alng this edge, enhancing vltage breakdwn. The higher-rder mdes generated by resnatrs n the narrw walls f the guide are nt greatly attenuated with distance alng the guide because the cutff frequencies f these mdes are determined by the wide dimensin f the guide. In cntrast, the cutff frequencies f higher-rder mdes generated by resnatrs munted in the brad walls f the guide are determined by the narrw dimensin f the guide and s are mre strngly attenuated with distance. In any case, these higher-rder mdes prvide undesired alternative cupling paths between adjacent resnatrs, bradening the stp bandwidth and reducing the maximum attenuatin. Thus it is desired t suppress higher-rder mdes as much as pssible.

The symmetry f this arrangement prevents generatin f the TE p0 mde. The next least-attenuated mde is the TE_ Q mde.

7 The least attenuated higher-rder mde fr the reject filters shwn in Figs. 1(a) and (b) is the TE?Q mde. T prevent this mde frm affecting filter perfrmance, it was fund necessary t place identical resnatrs identically tuned n either side f the waveguide at the same terminal plane. The symmetry f this arrangement prevents generatin f the TE p0 mde. The next least-attenuated mde is the TE_ Q mde. In rder t avid adjacent cavity cupling by this mde, it was fund necessary t space the resnatrs by five quarters f a guide wavelength rather than the usual three quarters. Small inductive irises can be seen Just within the flanges f the tw filters. These irises match ut the residual reflectin f the filter t give a very lw VSWR in the pass band. The large screws shwn extending frm the cavities are fr fine tuning. They allw each pair f resnatrs t be set t the prper resnant frequency and t be put in electrical balance. N attempt t adjust these screws shuld be made in the field. III. DESIGN PROCEDURE The fllwing specificatins were assigned t the 835O Mcps nise-reject filter. Waveguide WR-I37 Rejectin 60 db min ver 835O + 15 Mcps Pass-band lss 0.1 db max at 775O Mcps Pass-band VSWR < 1.15 Pwer handling 100 lew CW min Operating temperature F Simple calculatins shwed that the allwed temperature variatin culd change the center frequency f the filter by almst 7 Mcps. The 60 db rejectin bandwidth was increased t allw fr this variatin and 3 Mcps mre were added t allw fr tuning tlerance. Thus a 60 db rejectin bandwidth f UO Mcps was used in the design f the filter.

8 Figure f Ref. 1 shwed that a 0.01 db ripple Tchebyscheff filter f six resnatrs wuld prvide 63 db rejectin ver a ko Mcps band at 8350 Mcps, and that this filter wuld have a ttal stp bandwidth f 128 Mcps. This relatively narrw stp bandwidth (l.5#) insured that the pass band was well remved frm the lssy range near the stp-band edge and allwed use f the apprximate design prcedure fr filters with very narrw stp band6 described in Chapter 12 f Ref. 1. The equivalent circuit fr the filter is shwn in Fig. 2. All the resnatrs are tuned t the stp-band center frequency. It might be bserved that the resnatrs are shwn in the equivalent circuit as series-resnant circuits shunting the transmissin line, althugh the actual resnatrs are antiresnant structures attached t the side f the waveguide. This is nt incnsistent because the effect is the same pen circuiting the waveguide at its side is the same as shrt circuiting it at the center, when viewed frm sme distance alng the guide. In fact, as clsely as it was practical t measure, it was fund that the axial centerline f the cavity cincided with the terminal plane f the effective shrt circuit caused by the cavity at its resnant frequency. This is in basic agreement with the electrmagnetic bundary-value slutin fr a similar structure given in Sec. 6.6 f Ref. 2. The structure f Ref. 2 is an iris-cupled H-plane T junctin. It differs frm the structure used in the filter in that it has a side arm n nly ne side f the main guide and because it uses very thin irises. Als, it is based n small apertures and des nt apply near the secnd mde cutff frequency, where the filter perates. These differences make the circuit f Ref. 2 nly qualitatively useful in determining resnatr and cupling iris sizes. The lw-pass prttype and prcedure f Ref. 1 were used with the pertinent values given abve t find the slpe parameters fr the resnatrs. Only three slpe parameters are required because the structure is symmetrical abut its center. The required slpe parameters, nrmalized t the guide characteristic impedance, are x L = 87.65, ^ = 45.65, x = 40.5O.

In the interests f mechanical simplicity, such an impedance step was nt incrprated in the actual structure.

9 These slpe parameters are defined n a frequency, rather than reciprcal guide wavelength basis. The theretical design calls fr the characteristic impedance, Z,, f the waveguide between resnatrs t be slightly less than the generatr and lad impedances, Z. In the interests f mechanical simplicity, such an impedance step was nt incrprated in the actual structure. This missin causes negligible effect in the stp band and a slightly larger maximum reflectin than predicted in the pass band. Hwever, pass-band perfrmance was nt expected t agree with simple thery because the resnatr spacing f 5A guide wavelengths at the stp-band center frequency was a pr apprximatin in the pass band t the frequency independent 90 line lengths assumed by the thery. Fr this reasn it was necessary t use small Irises at the ends f the filter t btain lw VSWR in the pass band. The simple resnatrs shwn in Fig. 2 are gd apprximatins t the actual situatin in a narrw band arund their resnant frequencies, but nly when bth cavities f a pair are identical. A mre accurate equivalent circuit that allws fr small differences between the cavities f a pair is shwn in Fig. 3«It can be seen that if the elements f subscript 1 are identical t thse f subscript 2, the circuit f Fig. 3 reduces t that f a single series resnatr. Hwever, if any differences exist between element values f subscripts 1 and 2, then the circuit has tw impedance zers separated by a ple. The mutual capacitance C is needed t accunt fr the fact that the tw zers d nt cnverge t the same frequency as ne resnatr f the pair is tuned tward the frequency f the ther but remain rather widely separated. Frequency differences f abut 150 t 400 Mcps, increasing with decreasing slpe parameter, were bserved fr the cavity pairs investigated fr the filters described herein. Electrical balance f the tw resnatrs ccurs when the lwer-frequency zer cmbines with the ple and bth disappear. If the frequencies f the zers are bserved by mnitring transmissin lss f a swept signal, the lwer frequency zer will disappear nly if R, = Rpj therwise, the transmissin lss at the lwer frequency will simply pass thrugh a minimum. It is pssible t use this phenmenn t determine when a pair f cavities is electrically balanced. Als, it is imprtant t knw where the secndary

$/C is given by the equatin C l/ C - I if \2 te 1-1 (i) where f and f. are the upper and lwer bserved frequencies f minimum transmissin.$ Since the mutual cupling is presumably mstly by way f the TE_ Q mde, it shuld be pssible t use this infrmatin t cmpute the prprtins f energy stred in the waveguide in the TE_ 0 mde and the energy

Since the mutual cupling is presumably mstly by way f the TE_ Q mde, it shuld be pssible t use this infrmatin t cmpute the prprtins f energy stred in the waveguide in the TE_ 0 mde and the energy

10 frequencies are, since they may lie at a pass-band frequency f interest and cause a lss peak there. If the cavities are nearly the same and tuned nearly alike, it can be shwn that the rati f C./C is given by the equatin C l/ C - I if \2 te 1-1 (i) where f and f. are the upper and lwer bserved frequencies f minimum transmissin. Since the mutual cupling is presumably mstly by way f the TE_ Q mde, it shuld be pssible t use this infrmatin t cmpute the prprtins f energy stred in the waveguide in the TE_ 0 mde and the energy stred within the resnatrs, but this will nt be dne here. With the establishment f the physical cnfiguratin fr the resnatrs, the frequencies invlved, and slpe parameters required, a test Jig f adjustable iris pening and cavity length was made up. The cavity width was fixed at 1.00 inch because this made the cavity nearly square, which is a minimum-lss prprtin. The test Jig lked much like the filter f Fig. l(a), but with nly a single pair f resnatrs (Drawing A-7^082). The test prcedure cnsisted f selecting a value f iris pening, then shimming t find the cavity length that resnated with this iris pening at the stp-band center frequency. Resnance was cnsidered t be the frequency f f maximum transmissin lss. A lumped-element equivalent circuit fr the test Jig, near cavity resnance is shwn in Fig. k. The vltage V f Fig. h is prprtinal t the vltage acrss a matched detectr n the utput side f the test Jig. Analysis f the circuit f Fig. 1* shws that R / Z " I L /20 \i / -1 / (2) where L is the measured transmissin lss in decibels at resnance, Next, the bandwidth, Af, was measured between tw pints having sme cn- venient level f insertin lss L in decibels, typically 3 r 6 db. Then the slpe parameter, x, f the resnatr pair was fund frm

(R + Z /2) 2 _- R p 2 inl/10 JW (3) 10 If R < < Z and bandwidth between the 3 <lb pints is used, this equatin can be apprximated by *fh "Is < k) Since the simple equivalent circuit des nt hld

made. On the basis f the abve-described measurements, graphs were prepared f cavity-length, x/z, R/Z, and pass-band VSWR as functins f iris pening.

11 (R + Z /2) 2 _- R p 2 inl/10 JW (3) 10 If R < < Z and bandwidth between the 3 <lb pints is used, this equatin can be apprximated by *fh "Is < k) Since the simple equivalent circuit des nt hld quantitatively in the pass band, it was necessary t measure the VSWR f each cavity pair at the pass-band center frequency s that a reasnable estimate f the filter reflec- tin in the pass band culd be made. On the basis f the abve-described measurements, graphs were prepared f cavity-length, x/z, R/Z, and pass-band VSWR as functins f iris pening. The first tw graphs were used t find the cavity lengths and iris penings apprpriate t the required values f the slpe parameters given in Sec. II. The graph f R/z culd have been used t predict the maximum attenuatin f the filter (assuming n adjacent cavity cupling via higher-rder mdes) but this was nt believed necessary. A calculated value fr the electrical spacing f the resnatrs at the pass-band center frequency and the apprpriate values f resnatr VSWR as taken frm the graph were used n a Smith chart t predict that the VSWR f the filter at the center f the pass band wuld be 1.24, befre cancelling with irises. The final internal dimensins f the 835O Mcps reject filter are shwn in Fig. 5 (Drawing S-17159). Iris thickness is inches in every case. The cavities were made an additinal inch lng t allw fr tlerances and tuning. IV. MEASUREMENT, TUNING AND PERFORMANCE The basic scheme fr tuning and measurement f rejectin invlved a swept- signal generatr feeding tw channels, ne f which held the filter and the

ther f which allwed the insertin f calibrated attenuatin up t 90 db. Essential t the peratin f this technique was a lw-nise TWT amplifier used Just befre detectin.

The limiting actin was needed t prevent blcking and disabling the detectr and fllwing circuitry. The tuning prcedure began with all resnatrs detuned by the tuning screws.

12 ther f which allwed the insertin f calibrated attenuatin up t 90 db. Essential t the peratin f this technique was a lw-nise TWT amplifier used Just befre detectin. This amplifier allwed sufficient sensitivity t make accurate measurements in the stp band and acted as its wn limiter when verladed by the much strnger signal in the pass band. The limiting actin was needed t prevent blcking and disabling the detectr and fllwing circuitry. The tuning prcedure began with all resnatrs detuned by the tuning screws. The first pair f resnatrs was tuned fr maximum attenuatin at the stp-band center frequency, 8350 Mcps. Subsequent pairs f resnatrs were brught in t the same frequency, with the requirement that the stp band remain centered n 835O Mcps after tuning. Then the screws f each pair were adjusted the same amunt but in ppsite sense t reduce secndary resnances in the pass band withut changing the frequency f maximum rejectin. Finally, tuning fr each resnatr pair was tuched up t minimize any discernible ripple acrss the stp band. The cut-ff f the tuned filter was very sharp. When the scillscpe was adjusted t shw bth the pass-band respnse and the zer-signal base line, the transitin between pass and stp bands appeared as an almst vertical line. Figure 6 is a plt f the attenuatin between 55 and 80 db as a functin f frequency. The slight step between 60 and 70 db n the lwer-frequency side is caused by measurement errr, nt by the filter. It can be seen that the 60-db bandwidth is abut 52 Mcps, rather than the predicted value f k-0 Mcps. With the clearer visin f hindsight, the cause f this discrepancy was traced t an errneus measurement r test-jig fault fr the test iris f largest pening. The displacement f this end pint n the graph f x/z vs. iris pening was interpreted as curvature rather than errr. The result was that the fur center cavity pairs have larger iris penings than they shuld fr the prpsed design. It had been demnstrated previusly with the 805O Mcps reject filter that the design technique gave gd agreement with thery, and s n appreciable discrepancy was cnsidered likely n that cunt.. Since the errr was in the preferred directin, n attempt was made t rewrk the filter. 8

A VSWR f 1.32 was measured in the pass band (7750 Mcps) rather than the predicted value f 1.2^. The difference was attributed t the same errr discussed abve.

Figure 7 Is & plt f the VSWR f the iris-cmpensated filter frm 77^2.5 Mcps t 7762.5 Mcps. The VSWR is under 1.02 ver mst f this frequency range.

13 A VSWR f 1.32 was measured in the pass band (7750 Mcps) rather than the predicted value f 1.2^. The difference was attributed t the same errr discussed abve. The size and lcatin f matching irises at each end f the structure were fund by an impedance-measuring technique related t the apprach used by Kajfez fr matching waveguide tw prts. Figure 7 Is & plt f the VSWR f the iris-cmpensated filter frm 77^2.5 Mcps t Mcps. The VSWR is under 1.02 ver mst f this frequency range. Attenuatin measurements in the pass band f the filter are shwn in Fig. 8. Insertin lss f 0.05 db r less was fund by audi-substitutin attenuatin measurements under lw pwer and crrbrated by calrimetric attenuatin measurements under high-pwer cnditins. High pwer in the pass-band frequency range was applied t the filter by incrprating it in a resnant ring circuit. The pwer was increased in steps frm 50 kw t 305 lew CW. The filter carried the maximum pwer, 305 kw CW, that the ring culd deliver. After ten minute's peratin at this level withut breakdwn, the high-pwer test was ended. During the test cling water was run at the rate f ne galln per minute thrugh the pipes n bth brad walls f the filter. At 305 kw CW, the maximum temperature f any cavity was 180 F, the minimum at any cavity 120 F, while the maximum temperature alng the center f the waveguide was 155 F and the minimum 130 F. At the rated pwer f 100 kw CW, the maximum cavity temperature was 100 F and the minimum 75 F, while the average temperature n the waveguide was 95 F. During the test the input water temperature varied between 70 and 80 F, and the ambient temperature was 75 F. The lw-pwer measurements f stp-band rejectin and pass-band attenuatin and VSWR were made bth befre and after the high-pwer test. The nly change that ccurred was a small imprvement in the VSWR after the high-pwer test. The VSWR plt f Fig. 7 is based n measurements taken after the high-pwer test.

V. CONCLUSION A nvel design fr a reflecting-type, band-rejectin filter fr highpwer waveguide has been presented and shwn t perate essentially as predicted.

14 V. CONCLUSION A nvel design fr a reflecting-type, band-rejectin filter fr highpwer waveguide has been presented and shwn t perate essentially as predicted. It has reasnable size, large stp-band rejectin, lw pass-band lss and reflectin, and carries large amunts f KF pwer withut breakdwn r excessive heating. WJG:mfm 10

15 P I Fig. 1 (a) Three cavity pair 8O5O Mcps reject filter 11

16 -r 00 CM I s 00 u CO DO 12

17 .0 -t M CO ^MrHI a;.-i H D»"3 fa M (O TÜÜLrHI M IO -JÄlrHH O I O w ß X H CD u 3 Ü u N Owl JÜLHH" w (SI ^JWLHi H En 13

18 ^w ic. 2 Fig. 3 Equivalent circuit fr ne resnatr pair Fig. h Equivalent circuit fr test Jig Ik

19 -f r- 00 i vo -r i _ I D CO -in-». m CVJ Cd CD ö ±1 T CM CO +^> UJ z 0) " "8 _) B (0 ac III UJ m ft I Z ü UJ ^ u z X ~ t- CO UJ rr CD < t- c/> z O CD CO z UJ < _) < <H CO Ö H (0 a 1 n r d i- -p UJ c 3= s 2 >- CO PL. H 15

20 CO p a; 10 - in CO. >- z UJ a a: ft 2 & CO U a c p fo Ü d) v bo ID CM CO (qp) N0i±03r3a 16

21 u V B Iß ft ft OO CO >> I TJ I H UMSA 17

22 u iß R n CO Ml Q- O 2 c B»**» 0 O 4 >- B U u z 2 O UJ -Ö 3 s as UJ I t t & t! t CO i (qp) SSCH NOI1M3SNI 10

REFERENCES 1. G. L. Matthaei, L. Yung, and E.M.T. Jnes, Micrwave Filters, Impedance-Matching Netwrks, and Cupling Structures, (McGraw-Hill Bk Cmpany, New Yrk, 196*0 2. N. Marcuvitz, Ed.

23 REFERENCES 1. G. L. Matthaei, L. Yung, and E.M.T. Jnes, Micrwave Filters, Impedance-Matching Netwrks, and Cupling Structures, (McGraw-Hill Bk Cmpany, New Yrk, 196*0 2. N. Marcuvitz, Ed., Waveguide Handbk, Rad. Lab. Series 10, (McGraw- Hill Bk Cmpany, New Yrk, 1951)- 3. D. Kajfez, "Wide-Band Matching f Lssless Waveguide Tw-Prts," IRE Trans, n Micrwave Thery and Techniques, MTT-10, l~fk (May, 1962) 19

24 DISTRIBUTION W. J. Getsinger (6) L. W. Bwles F. J. Dminick C. E. Muehe W. A. Andrews M. L. Stne S. S. Cupli C. W. Jnes E. P. McCurley J. S. Arthur L. Rainville A. F. Standing R. R. Silva 20

UNCLASSIFIED Security Classificatin DOCUMENT CONTROL DATA - R&D (Security classificatin f title, bdy f abstract and indexing anntatin must be entered when the verall reprt is classified) 1.

GROUP Nne A Nise Rejectin Filter fr Waveguide Carrying High Pwer DESCRIPTIVE NOTES (Type f reprt and inclusive dates) Technical Nte AUTHOR(S) (Last name, first name, initial) Getsinger, William J.

25 UNCLASSIFIED Security Classificatin DOCUMENT CONTROL DATA - R&D (Security classificatin f title, bdy f abstract and indexing anntatin must be entered when the verall reprt is classified) 1. ORIGINATING ACTIVITY (Crprate authr) 3. REPORT TITLE Lincln Labratry, M.I.T. 2fl. REPORT SECURITY CLASSIFICATION Unclassified 2b. GROUP Nne A Nise Rejectin Filter fr Waveguide Carrying High Pwer DESCRIPTIVE NOTES (Type f reprt and inclusive dates) Technical Nte AUTHOR(S) (Last name, first name, initial) Getsinger, William J. REPORT DATE 3 August a. TOTAL NO. OF PAGES 23 7b. NO. OF REFS d. CONTRACT OR GRANT NO. AF 19(628)-5167 PROJECT NO. Nne AVAILABILITY/LIMITATION NOTICES 9a. ORIGINATOR'S REPORT NUMBER(S> Technical Nte OTHER REPORT NOISI (Any ther numbers that may be assigned this reprt) ESD-TDR Nne It. SUPPLEMENTARY NOTES 12. SPONSORING MILITARY ACTIVITY Nne Air Frce Systems Cmmand, USAF 13. ABSTRACT A nvel design fr a waveguide band-rejectin filter carrying high pwer in its pass band is presented. The resnant cavities f the filter are munted in pairs n the narrw walls f the guide and cupled t the guide thrugh rectangular penings the full height f the guide. Thus there are n edges r radii n which electric field cncentratins can ccur. An X-band mdel having six cavity pairs had a rejectin greater than 80 db ver 45 Mcps in its stp band with an insertin lss under 0.05db and VSWR under 1.02 in its pass band. With water cling, the filter carried ver 300 kw CW withut breakdwn. 14. KEY WORDS filters high-pwer waveguide X-band bandpass filter reflectin filter nise rejectin filter Haystack 21 UNCLASSIFIED Security Classificatin

EE 3323 Electromagnetics Laboratory

EE 3323 Electromagnetics Laboratory EE 3323 Electrmagnetics Labratry Experiment #1 Waveguides and Waveguide Measurements 1. Objective The bjective f Experiment #1 is t investigate waveguides and their use in micrwave systems. Yu will use