NATIONAL RADIO ASTRONOMY OBSERVATORY

CHARLOTTESVILLE, VIRGINIA

ELECTRONICS DIVISION INTERNAL REPORT NO. 280

AN ADJUSTABLE SHORT-CIRCUIT FOR MILLIMETER WAVEGUIDES

A. R. KERR

JULY 1988

NUMBER OF COPIES: 150

AN ADJUSTABLE SHORT-CIRCUIT FOR MILLIMETER WAVEGUIDES

A. R. Kerr

Introduction

Adjustable waveguide short-circuits are used as tuners in manymillimeter wave applications, and are often lossy and unrepeatable.Commercial millimeter-wave components have even been made with waveguideshort-circuits consisting simply of loose-fitting metal strips held inplace by a set-screw. Figure 1 shows the measured return loss of such ashort at Ku-band; it is unacceptably lossy over the whole waveguide band(between the markers).

This report describes a non-contacting short-circuit for use in full-height WR-4 waveguide (170-260 CHz). Its loss is low, and the phase of itsreflection coefficient varies smoothly across the full waveguide band. Thedesign can be scaled for use in any millimeter waveguide band.

A. Requirements for a Good Adjustable Waveguide Short-Circuit

The essential requirements for an adjustable waveguide short-circuit are (i) low loss, (ii) smooth variation of the phase of thereflection coefficient with frequency, (iii) smooth variation of phase asthe short-circuit is moved, and (iv) good repeatability.

B. Contacting Versus Non-Contacting Shorts

In the standard millimeter waveguide bands, non-contacting short-circuits are commercially available. These usually consist of a metaldumbbell, often teflon coated, which rotates on the end of a long micrometershaft. Centering of the dumbbell in the waveguide is not easy to ensure.It has been common experience that such shorts can be lossy in some partof their frequency range. From the measurements described here, the mostlikely cause of this loss appears to be off-center mounting of the dumbbell.

In the microwave bands, adjustable shorts are commerciallyavailable which use plungers with rectangular or circular cross-sections.As will be shown below, rectangular non-contacting plungers are prone toin-band resonances if they are not precisely centered in the waveguide, whichmakes them unsuitable for millimeter wavelengths.

In reduced height millimeter waveguide, non-contacting dumbbellshort-circuits are difficult to make because of the small dimensions andclearances involved. For example, a WR-10 (75-110 GHz) mixer may have itswaveguide height reduced to 1/4 of full height -- i.e., to 0.0125". Non-contacting plungers of rectangular cross-section, centered in the waveguideby a layer of mylar adhesive tape [1], have achieved some success, but ourexperience indicates that carefully made contacting shorts are superior inreduced height waveguide.

Contacting short circuits have two main loss mechanisms: resistiveloss at the contact points, and loss due to power leakage into the gapbehind the contact points. This latter loss is caused by the contactfingers (or strips) acting as antennas to couple power from the incidentwave into the slender waveguides formed between the body of the short-circuit and the upper and lower waveguide (broad) walls. This might beovercome by using multiple quarter-wave spaced contacts; however, contactingshort-circuits cannot readily be made with multiple sections. Wear canlimit the lifetime of contacting short-circuits. We have found this canbe minimized by carefully rounding the tips of contact fingers and usingheavy gold plating on the sliding short and on the inside of the waveguide[2].

This report will not deal further with contacting short-circuitsnor with short-circuits in reduced-height waveguide.

C. The Quarter-Wave Plunger as a Waveguide Short-Circuit

Within their normal operating bands, rectangular waveguidespropagate only the dominant TE E) mode. Introducing a conducting plungerinto the waveguide creates a gap which supports the coaxial TE ll -like mode(like two rectangular TE E) modes wrapped around the plunger as shown inFigure 2(a)), and the same mode rotated 90° -- Figure 2(b). Also, thecoaxial TEM mode is supported -- Figure 2(c). Depending on the plungerdimensions and the presence of any dielectric material in the gap aroundthe plunger, higher-order gap modes may also be supported within thenormal waveguide band.

If the rectangular plunger is perfectly centered in the waveguide,the incident rectangular TE10 mode excites the coaxial TEll gap mode and alsohigher order (evanescent) gap modes with anti-symmetry about the virtualshort-circuit plane in Figure 2(a). A non-contacting short-circuit can bemade using a series of high and low characteristic impedance quarter-wavelong rectangular plungers. The length of the sections is based on theguide wavelength of the coaxial TE ll -like gap mode (Figure 2(a)), and shouldbe corrected for the fringing fields associated with the evanescent gapmodes.

If the plunger is off center vertically, the incident mode willalso excite the coaxial TEM mode and higher-order (evanescent) gap modeswith both vertical and horizontal symmetry. If the plunger is diagonallyoff center, then the 90°-rotated coaxial TE ll -like gap mode (Figure 2(b))will be excited together with the corresponding set of 90°-rotated higher-order modes.

The sharp resonances in many of the measurements below areattributed to resonances between pairs of undesired modes all of which arereactively terminated at the far end of the plunger. It appears to begenerally true that plungers which are diagonally off center have more in-band resonances than those which are vertically or laterally off center.This is a result of coupling between the incident TE E) waveguide mode and

2

two sets of gap modes, one rotated 900 relative to the other in the

diagonally off-centered case.

The above discussion applies equally to rectangular and circularplungers in rectangular or circular waveguides.

It may be tempting to fill the gap around a waveguide plungerwith a dielectric material to improve the performance of the short-circuitand to help center the plunger. However, there are some disadvantages tothis approach: (i) The presence of the dielectric lowers the cut-offfrequencies of the higher-order gap modes, which increases the likelihoodof in-band resonances. (ii) In practice, some clearance is needed betweenthe dielectric and the waveguide walls. Lateral movement of the plungerwithin this air gap may cause greater changes in loss or phase than wouldoccur in the absence of dielectric. (iii) Fabrication of millimeterwaveguide short-circuits with dielectric supports is a non-trivial additionalcomplication. For these reasons, and because it is possible to make agood short-circuit without dielectric supports, we do not considerdielectric-filled short-circuits further here.

II. Measurements

In this and the following sections we describe measurements on anumber of Ku-band waveguide plungers and dumbbells. These measurementswere made on a HP 8510 microwave network analyzer over the frequency range10-20 GHz. In all figures in this report showing measured results, twofrequency markers are used to indicate the edges of K u-band (12.4-18 GHz).Except where explicitly noted, plungers or dumbbells are arbitrarilylocated along the waveguide and are not necessarily positioned at themeasurement reference plane.

A. Rectangular Plungers

Initially it seemed desirable to use a plunger which would blockas much as possible of the waveguide cross-section. It soon became apparent,however, that a plunger considerably narrower than the waveguide was muchless susceptible to in-band resonances. Figure 3 shows the magnitude andphase of S il for a typical rectangular plunger in Ku-band waveguide(12.4-18.0 GHz). In Figure 3(a) the plunger is carefully centered betweenpolystyrene foam supports (e r = 1.0) and no undesired in-band resonancesare evident. In Figure 3(b) the plunger is slightly off center and a sharpresonance is visible near 17.4 GHz. A slight rotation of the plunger(about the waveguide axis) excited a resonance near 13.6 GHz as shown inFigure 3(c). A rotated and off-center plunger exhibits both theseresonances, as shown in Figure 3(d). The height of this plunger was chosento give a 0.002" clearance to the waveguide broad walls when scaled to WR-4 waveguide (170-260 GHz).

To improve the 0.4 dB return loss of this plunger, an additionalidentical section was added at various distances from the first section.

Typical results are shown in Figure 4 for three spacings. A return loss< 0.05 dB is achievable across most of the band, but always with at leastone resonance. (With perfect centering of the plungers, resonances couldprobably have been avoided.)

It seems clear from these and other experiments that it would bedifficult to make an acceptable millimeter wave short-circuit using arectangular plunger. Mechanical tolerances would not allow sufficientlyprecise location of the plunger to ensure resonance-free operation.

B. Circular Plungers

Our initial measurements on circular plungers were on a HP ModelP920B Ku-band unit. This has a copper dumbbell which rotates on the endof a micrometer shaft as shown in Figure 5. Measurements of S 11 for the HPshort are shown in Figure 6(a). The return loss is < 0.02 dB (themeasurement noise limit) across the band, and the phase variation is wellbehaved. The effective short-circuit plane is located 0.015" behind theend of the dumbbell (i.e., away from the source) as indicated in Figure6(b). The effect of moving the short towards the waveguide broad wall isshown in Figures 7(a) and (b). It is clear that only relatively largemovements away from the waveguide centerline cause a resonance to moveinto the waveguide band.

The HP design is in most respects suitable for scaling tomillimeter wavelengths. However, the 0.006" clearance between the dumbbelland the waveguide wall is too small; in WR-4 waveguide (170-260 GHz) thecorresponding clearance would be 0.0004" which is impractical. For thisreason we tested a series of dumbbells (Figure 8) with progressively smallerdiameters but having other dimensions the same as the HP design. Thesedumbbells were supported in a section of waveguide by polystyrene foam.Figure 9 shows S 11 for the dumbbell with the original HP dimensions, andFigures 10(a)-(c) show S 11 and 1S211 for the smaller dumbbells.

From these measurements on simple three-section dumbbells, it wasclear that additional sections would be needed if the dumbbell diameterwere to be in a reasonable range for scaling to millimeter wavelengths.Five- and seven-section dumbbells of diameter 0.253" (Figure 11) weremeasured, with the results shown in Figure 12. The smaller return loss andslightly wider I S il l-bandwidth of the seven-section dumbbell make it agood choice for millimeter wave application. The 0.253" diameter and0.029" wall clearance of the Ku-band dumbbell correspond to 0.0175" diameterand 0.002" wall clearance when scaled to WR-4 waveguide (0.0430" x 0.0215").

III. The Millimeter Wave Design

The seven-section 0.253" diameter dumbbell was mounted in the end ofan aluminum holder of cross-section 0.614" x 0.303", i.e., 0.008" smallerin each dimension than the waveguide. This clearance was chosen to simulatea 0.0005" clearance in WR-4 waveguide. The end section of the dumbbell

4

was now entirely inside the holder. Figure 13 shows the measured S il . InFigure 13(a) the end of the dumbbell is located at the measurement referenceplane, while in Figure 13(b) the dumbbell is moved towards the source 0.090"to give the flattest phase variation across the band. From this, theshort-circuit can be characterized as a constant reflection coefficient ofunit magnitude and phase -130°, located 0.090" from the end of the dumbbell,or, equivalently, a constant normalized impedance of -j0.46 located at thesame place.

Figure 14 shows that bending the dumbbell substantially towards thebroad wall or narrow wall of the waveguide does not produce any in-bandresonances. Bending the dumbbell substantially towards the corner of thewaveguide, however, causes two very sharp in-band resonances as shown inFigure 15(a). These resonances can be suppressed, as shown in Figure 15(b),by putting lossy ferrite beads on the end of the holder. This is probablythe reason for the use of absorbing material in the HP design (Figure 5).Even in the present measurements with a greatly off-center dumbbell, theseresonances were weakly coupled to the TE 10 waveguide mode and are barelydiscernable on the phase plot. At millimeter wavelengths, the greaterconductor loss will tend to suppress these resonances (as did the absorbingbeads). Certainly, a dumbbell reasonably centered by eye should have aresonance-free response.

IV. iltaign_taK_EILA_Iigmellicit_i112:161014.1

A. Construction

Scaling the above design from Ku-band to WR-4, a factor of 14.5in frequency (and linear dimensions), results in a dumbbell structure withdiameters 0.0175" and 0.0086". To machine this on a lathe was judgedimpractical, so the short-circuit was assembled from simpler components --brass beads on a copper-plated steel rod, mounted on the end of a brasscarrier. To obtain the necessary precise alignment (± 0.0002"), an assemblyjig holds all these components in place on a hot-plate while low temperaturesolder is applied. After assembly the whole structure is gold plated.

The WR-4 short-circuit assembly is shown in Figure 16. Figures17 and 18 give details of the assembly fixtures.

B. Drive Mechanism

At short millimeter wavelengths it is desirable to control theposition of a short-circuit to micron accuracy, and for this reason the drivemechanism, shown in Figure 19, has a #000-120 lead screw (0.0083"/turn).A rotating nut moves the non-rotating lead screw which is a part of theshort-circuit assembly. The mechanism is compact enough to allow two tobe mounted on the back of a WR-4 mixer block. Friction is very low, andthe mechanism is suitable for cryogenic operation provided oil is removedfrom the ball races.

5

V. References

[1] M. K. Brewer and A. V. RAisamen, "Dual-Harmonic Millimeter WaveguideBackshorts: Theory, Design, and Test," IEEE Trans. MicrowaveTheory Tech., vol. MTT-30, no. 5, pp. 708-714, May 1982.

[2] V. Summers, "Chemical Lab Procedures, 1988," Chemical Lab ReportNo. 7, National Radio Astronomy Observatory, Charlottesville,VA 22903, Jan. 1988.

6

S 1 1 0 Ett 11-4-3*--

di

1111111111111111111111111111111111111MINIUM1111111

1111=1111111111111111111111111111111111-11111111111111111111111111111101

aiiinlaritIMMIliraillir11111f111111311111LEAM3

araminommumimmummotausimmonumm

4 't

— Ito*

to GA\ 2-c7- G-4-s 2 0

Fig. 1. S 11 magnitude and phase for a rectangular bar inside a WR-62waveguide (0.622" x 0.311"). The bar is 0.007" smaller than thewaveguide in both dimensions.

VirtualS/C plane

Coaxial TEn -like mode

(a)

Vir-tualSlC ptane

Coaxial TE11

-like mode rotated 90°

(b)

Coaxiat TEM mode

(c)

Fig. 2. E-field patterns of modes in the gap between a plunger and thewaveguide. The coaxial TE ll -like gap mode (a) is excited by theincident rectangular waveguide TElo mode. If the plunger islaterally off center, the 90°-rotated version of the same mode(b) is also excited. If the plunger is vertically off center,the coaxial TEM mode (c) is also excited.

8

10CAA

20

\AaS

e.

r- T--

----

7-7-

--7-

---7

7,__

T1

NM

MI

ME

7+

r___

III

UMW

IIIl

l I

NIB

INN

IMIE

in

1 .1

1111

MIM

EM

OM

RomP

ii _511

1

IM

EM

IIII

MM

IE

EM

I=

MU

M=

rnaaaii

iiaaaa

1111

1111

1 =11

1111

1111

11M

I111

111111

111111

111111

111111

1111

111111

1 MIN

EM

ININ

IME

MI

111111

111111

111=1

1111

0

1111

1111

1

ON

E' -

';--

4--

-''

iV

illi

nt

Ole

M111

1111

.1111

1=11

1111

1111

111

1.111

1111

1111

1111

.1111

1 KIM

1 =111

111111

111111

111111

111111

11111

NIIII

NNI.1

1111

1111

1111

1111

1111

1101

1.111

1111

1111

1111

111

111111

.111

1111

1••E

MU

M1111

111111

MB

MU

Mi_

sauto

mm

tim

aim

(a)

111111

111111

111111

11 M

IME

NM

I IM

RE

OPPIE

M111

116111

1,111

111111

111111

111111

111111

aium

man

sin

mE

n

;

111111

117111

111111

111111

111111

Iiiiri

alm

bi•N

EEM

PLP

111111

111111

111111

111111

111111

111111

111111

111111

111111

111111

11111

111111

111111

111111

111111

111111

111111

111111

111111

111111

111111

111111

111111

1111

111111

111111

111111

111111

111111

11111

111111

111111

111111

111111

=111

1111

1111

1111

111111

111111

111111

111111

111111

1111

0 c.4

48

0Zo t

o20

to

6;41

'2. 0

(c)

Fig.

3.

Sil magnitude and phase for a single rectangular plunger 0.311"

x 0.253" x 0.117" long in a terminated Ku-band waveguide (0.622"

x 0.311"). In Fig. 3(a) the plunger is centered between

polystyrene foam supports and no undesired in-band resonances are

evident. In Fig. 3(b) the plunger is slightly off center and a

sharp resonance is visible near 17.4 GHz. In Fig. 3(c) the

plunger is slightly rotated (about the waveguide axis) and a

resonance appears near 13.6 GHz. A rotated off-center plunger

exhibits both resonances, as shown in Fig. 3(d).

1111111111111111111111111111111/11wilitit==d ekTi,wz11111111111111 111111111111111.1111In= 11111111111111111111111M111111.111111111111111111

11111111111111111111111111111111

Q9 0,141 .

MN1111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111•111111111111 11.0!1111111111111111111111111

to G-H-s zo 2.o

(a)

1 1 1 1 1

'485 11111

IIVIIIIIM111111

11111111111 1111111111111111111111.1111111111.11111

1111111111111111111111111111111111111111111111111111111111111111111111111111

la G-41 Zolo G44 Zo

(6)

c:\ v-t

1 I T

1.1111111111N111111111111111111111111.11111111111111111111111111111111MI1111111111.1111111111PA11111111111111111111111Millii

m•IT NMI=111111=111111=

111111111111-1111Z to

(C)

Fig. 4. Magnitudes of S 11 and S21 for two plungers with variousseparations. Each plunger is identical to the one in Fig.Overall lengths of the two-plunger assembly are: (a) 0.385",(b) 0.485", and (c) 0.600".

10

Micrometer shaftunplated steel

Fig. 5. Details of a HP Model P920B Ku-band sliding short. The copperdumbbell rotates on the end of the micrometer shaft. Absorbingmaterial is placed behind the dumbbell. Clearance between thedumbbell and the waveguide broad walls is 0.006" above and below.

11

S‘l rViase•••••••■•••••■••■••

to G41 2O 10

(6)

(a)

41%0'

— ma*

k•••••••

Fig. 6. (a) S 11 for the HP short with the end of the dumbbell located atthe measurement reference plane. The return loss is < 0.02 dB (themeasurement noise limit) across the band, and the phase variationis well behaved. (b) S 11 with the short positioned to give theflattest phase variation (by eye). This occurs when the end ofthe dumbbell is located 0.015" in front of the reference plane(i.e., towards the source).

12

1111111111MINIAIMI1.11111=111111.1111 i4°2.

IMIIIIIIMIIIIIIIIIMIlli11111111111!IIMIIIIIIIIIMMILI

imillitu illfiiillidliiililkillisitAll 1111

1111 li 111M11111111111111111M11111111111111111111111MINIMMIll ! IIIMINIIIIIIIIIMIIIIII

I s 11111111.11111.1111110111111111111111111111111111111111111111 11 IIKIIIIMINIIIIIIII0

41:to°

0

—1%0.

ri 11111111MININUIMIEMIIIIINIMININICI

•11111111111111MINM

ry. 1 c) .r1 T

itimionge i 1111.1111•11111111

1! , . , 1 1N, s IIllIIau1 1 1 i .,

. . ii, • . : •o•iumumil, 11111111411111111

0

t it o

a

—PIO°

Sli

to

Gil VNACkey,■• Sit pLa.s e

to 6-4-r '2._e to

(6)

Fig. 7. The effect of moving the short towards the waveguide broad wall.(a) The dumbbell is 0.002" off center (0.004" gap to the waveguidewall). (b) The dumbbell is 0.005" off center (0.001" gap).

13

Material* Brass Dimensions in Inches.

Tolerances* ±0.001"

D Values* 0.299'0.282'0.253'0.224'0.195'

Fig. 8. Dumbbells with different diameters usedfor measurements in Figs. 9 and 10.

s (i

ammommuluma1111111111110111111111111111111111111111111111111111111111111111MAIMMOUPPIPS

0.z 11011111111111J1111111111111111111111111•11110,4 1111.111111111111UNIMONI

111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111M111111111111111111111111111111111111111111111111.11111111111111111111111111111111111111111111111111111111111111111111111111111.11EMPINB.jm

-oriii11111111111111 K11111111111111111111111111111111111111111

LaSe

—(%0D

113 Zo to

Fig. 9. S11 for the dumbbell with the original HP dimensions 0.299" dia.).

14

mu

mm

um

mu

mn

mo

mm

inam

mu

mm

umm

umm

uum

mm

mum

maim

mm

umm

umm

umm

um

mu

mu

mm

ame

MM

IZO

MW

W=

Iiii

kM

OIM

MO

MM

MA

MM

MM

UM

MU

MU

MO

IME

MM

UM

M

—>

2. .1

ASC

/

O

211e

•

MR

MU

MM

EM

Ma01

0,171

71111

1111

1111

1111

1111

MUri k

MM

EM

MM

UM

WM 11

11M

UM

MIU

MM

EIN

E M

I

mum

mam

mul

mu

mm

um

mu

mm

mum

mim

mum

mm

um

mim

mum

mis

mom

mom

mom

mis

mom

mm

inu

mm

um

nn

om

min

imW

=r4

=M

Ela

iirl

aM

MU

MII

MM

U6t

is

vLa

StO

tfe

ctse

234'

mm

am

mu

mm

mo

mm

amm

um

mm

mum

milm

Tum

mW

WW

WW

MO

MM

WA

MM

MU

NK

AM

AM

ME

IMM

EN

WM

t

111111

111111

111111

111111--

111111

111111

11111

1111

1111

1111

MIM

IIIIM

I MM

I11

1111

.1111

--111

1111

1111

1111

1111

1111

111 -1

1111

111-

1111

1111

111111

111111

111111

111111

111111

111111

1111

11111 -

-1111

1111

1111

111

1110

1111

1 =1.

1111

0111

111

111111

1 E3111

111111

111111

11111

zo

MM

EM

MU

MM

OM

MM

EM

MIB

MW

MM

ME

MM

UM

ME

MU

MM

MU

MM

EM

MM

IOM

MU

MM

MM

ME

MM

UM

MM

UM

MM

MU

MM

WM

MM

OV

IMIN

WS

'm

ilm

munsw

am

i mm

um

mim

mu

0 — 17

t.

1.0

SAt

tO-

lt "

C

.i11

1

2es"

mom

mum

mim

m=

mu

um

uu

=M

IND

WA

IOim

mu

mm

ing

unu

mm

um

man

AM

MIU

MM

UM

W

mo

mm

ism

om

mm

um

munim

mm

um

um

mim

ma

mm

um

mil

mm

mam

muum

mU

MM

MIS

MIM

Mm

om

mom

mum

mu

mm

um

ms

nai

mm

mm

men

um

mu

mm

Fig

. 1

0.

Su

an

d I

S21

1 fo

r th

e d

um

bb

ells

wit

h d

iam

eter

s:(a

) 0.2

82",

(b)

0.2

54",

and (

c)

0.2

24".

0.124= 0.253 0.124IL 5—SECTION

D = 0.253 0.1240.124 0.124

Materials Brass Dimensions in inches.

Tolerances' *0.002' except on D.

0.187i2:187-..F0.187 12187k0.1877—SECTION

12.... 0.187

.F

1134.-0.010.1 87.187 0.187

0.187

Fig. 11. Five- and seven-section dumbbells of diameter 0.253". The 0.253"diameter and 0.029" wall clearance of the K u-band dumbbellcorrespond to 0.0175" diameter and 0.002" wall clearance whenscaled to WR-4 waveguide (0.0430" x 0.0215").

16

<2.

‘1,N

kte

.SI

%III

I111

111111

111111

1111M

111111

111U1

111111

111111

111MM

AIM

Ill

111!11

111111

111111

MM

EIIM

II11

1 111

111 M

1111

1111

1111

klIPIli

lailiN

1111

111

111111

111111

111111

111111

11U

IIU

INIU

IMII

IIII

IMII

III

el 1111

111111

111111

111111

1.1111

11111M

ausu

nmam

pras

111111

111111

1111M

111111

111111

111111

11111

— —

IIII

1111

111111

1110 7

111111

111111

111111

111111

111111

111

11111r

All11

1111

1111

1111

1111

1111

1111

111114

111111

1 M11

1111

1111

1111

1 111

1111

1111

1111

1111

1111

1111

1111

110111

11111111

11111111

11111111

11111 1

11

tirin

g:P

IM

•

0

o •

2_

cV-e

o t

+

2. Z+

0

tt,17S

oe

(;:

tA20 1

0

111111

111111

111111

1i

i111

111111

111111

111111

0211 =

1111

111111

M11111

11111 M

1111

1111

1141.

111111

111111

111110

111111

1111

111111

1111

1M11

1111

1111

11111

11111N

IVAI

IIIIII

IME

ME

MP

111111

,4111.

1111

1111

1111

111P

211

111111

111111

111111

111111

111111

1111

115

21su

mm

umm

aim

s111

111111

111111

111M

1111

1111

1111

11111

111111

111111

111111

111111

-111

11111

111111

111111

111111

111111

11111

0111

111111

111111

111111

111111

1E11

1VA

111111

111111

111111

111111

111111

111111

11211

0.1111

111111

111111

111111

.11

1N

111 .111

11k2

111111

111111

111111

111110

1111

1020

Li"

C)

Fig. 12. Sli and

I S2

11

for 0.253" diameter d

um

bb

ell

s w

ith

(a) fi

ve s

ecti

on

s, and (b) seven

sections. The smaller return loss and s

lightl

y w

ider Sii

-bandwidth of the seven-

section d

um

bb

ell

mak

e it a good choice for millimeter wave applications.

r

to.:\ wm..n

1 r fir!

es,

1Acc e

+±,705*

- Ott)

(a)

t o G44 7.0 10 GAt zo

Fig. 13. S 11 magnitude and phase for the 0.253" diameter seven-sectiondumbbell supported in a metal holder (see text). In (a) the endof the dumbbell is located at the measurement reference plane,while in (b) the dumbbell is moved towards the source 0.090" togive the flattest (by eye) phase variation across the band.

• 8

.‘ Wt. a. C‘t

oit*-

V ",„

i ' I 1 1 11111111111111111111111111111111111111111ib......Lit.u...—,_.

14231iralm

(111111.9 111111111 1111111111111111111111111Entlelk111111111111111111111111111111111•1111IIIIIIMIIMIIIIIHMIIIIInummwsmnmum111ISIIIIIIMMIMIIMNIIIM1113111111111111111111111111111111111111111111111111111111111111111111

IIIII

H,1111.11111110111.01 ,

litithiNAIIIIIIIIIMMINV1111111111111111111111111111111111111111111111111NOMMIMII11111111111111111111111111 N11111111

ilif

IliMU

64-11 2.0(a)

tr“\ cs, .

1 1

Aer

11111111111111111111111111F11111Pm1111111111111111111111PLTIMM

1111111110111111111Malkir I1/111111111111111/11111111111/ I I

11111111111111111111111111111111111111111111111111111111111111111111111111

11111111111111111111111111111111

111111111111111111111111111IMMINMENIMIMMIN1111111111111MMINIMIILUMMOINEMIMMI1111111111111M1111111111111111111111111111111111111111111111111irmummumonaut111111111111111111111111111111

20 m(b)

111111111111111111NOMIN11111111111IIMP:a;mainiumbitaisibkir. 711111111111 U111111•1111111111111111 I1111111111M11111111111111111111111.1 I I I

111111M111111111111 N1111111 I I1111111111111111111111111111111111111

tr: f

11111111111111111111aaaa1111111111111111111111 .1111111111L.

11111111111111M.11111111111111116711011111.111111111111111111111ah;assammmur

1111 111111UNICIIIIINBEIMENIMITIMBEI

•11M1111111E11111011111111111111111111111111111111111111=11

sto

4 1%e

rzo *

641 -s

Fig. 14. The effect of bending the dumbbell in its holder so the end is:(a) 0.023" off center towards the waveguide broad-wall (0.006"gap), (b) 0.029" off center towards the waveguide broad-wall(i.e., just touching the wall), and (c) 0.029" off center towardsthe waveguide narrow-wall.

19

I

I

WIth

1 =

; q .

YA0.5I I

1111111110.1111114111111111111M11111 I

elIMAImilithaiiiiiitamiCal l 1. I1111111111.111111111111111111.1ir

1111111111111111111111111111111111111111111111111111111111111 •11111111111111111111111111111 •

-11111111111111111111111111111111111111111111111=1111111111111111111111111111111111111111111111111111111111M111111111111111111111

uiaaaaa si1111111M11111111111M.

1111111111111111111=1111111E1111111111111=111111111N111111111111111111111111111111M1

t .s

o e

Zo 2.0

(PO

111M1111111111111=11111.11111111111111111111111111111111

ILIMM111111111111111.11M11111111111111111111111111111N11111

1)(Aa

+•,,t)•

10 zc t 644 2, 0

OD'

Fig. 15. (a) The effect of bending the dumbbell in its holder so the endis 0.029" off center at 45° to the waveguide axes. (b) Thesame, but with absorbing beads attached to the end of the holder.

Shafts 0.009' steel, copperplated 200 u-In.

Quantity' 4 0.113 *0.002'

Beads Brass,Quantity! 12.Tolerance, *0.0002"

0.0175

0.0357

(a)

0.0129

T-0.0175

Arommorrwris, r 0.009

■IMINIMIPIPM11111111•1011111•111111

0.0129 -4 --I 0,0129

Fig. 16. (a) The WR-4 short-circuit assembly including the rectangularholder, which centers the dumbbell in the waveguide, and the#000-120 lead-screw. (b) The copper plated steel shaft andbrass beads.

21

2.00

MAIN BLOCKCross—section

0.50 SPACER

Fig. 17. Assembly fixture for the WR-4 short-circuit - see Fig. 18 fordetails.

22

Hole Al Tap #4-40 x 0.25' dp..

Holes 131 For 3/32' pins.

Holes CI Tap #2-56 through.

0.1501

0.2500.100

.11111•11110.71

-4.1 0.300 0.80

Quantity' 1.

1.00

2.00

0.250

--T*11111•1•1•1111M111111111MONIMMINIVM11111111111111111110

Cross-section

17If)e20.50

awe

Materials Aluminum.

Material' Aluminum.

I 0.0253

(a)

SPACER

LL Material* Murmur% 0.0122 *0.0002.

Holes Es ror 3/3irHoles Di 412-56 clear csbore for socket head.Hole Ft Tap 80-80 through.

(b)Fig. 18. Details of the WR-4 short-circuit assembly fixture

shown in Fig. 17. (Continued on next page.)

23

0100

0.0129

r I...-. 0.0129

0.0129

-41.- 0.007

0.053Oran 4••111.

0.01E90.0129

0.353

110•80 clear.

0.06311.41.

1080

•••10,

Tolerances, X.XXX ±0.002"X.XXXX ±0.0002'

0.0205' slot machined with guideand main block assembled, butwithout spacer

(c)

Material' ALUMIFILIPI Tolerances' XXXX *0.001X.XXXX ±0.0003

Quantity' 1,

I

--- can -I 1 0.061

(d)

24

(a)

Materials SS ilusettyl

0.0 1 r_

oar

0.685

0262

(b)

0255

Material Steel or SS. Quantity' 2 Tolterances• *0.002' exceptbeorMg hole 0.250 +.001/-.000

> U76

1

Th=

Breather hole, drill 067 (0.0324,

JACKSHORT MIME SHAFT I. NUT

(c)

* Note' Shaft dia. to frt PIC E4-5 bearing.

Fig. 19. Drive mechanism for the WR-4 short-circuit. (a) assembly, (b)body, and (c) drive nut. The drive shaft and nut (c) run in twoball races mounted in the body (b) which is attached to the mixer(shown dotted in (a)).

25