Beruflich Dokumente
Kultur Dokumente
305
MAIN LINE
1 1 1 I I
spect to the hole, whereas, the magnetic coupling will The dependence of coupling upon hole size is shown
induce waves in the auxiliary line which will be anti- in Fig. 4. The coupling was found to vary in the theo-
symmetric, as shown in Fig. 2(b). Thus, it can be seen retical manner when the coupling hole is smaller than
that the combined electric field in the auxiliary line will 2a/)X = 0.1. Larger holes produce less coupling than would
be stronger on one side of the hole than on the other. It is be expected from Bethe's theory.
to be observed that the strength of the electric coupling
should be independent of the angle between the trans- -1- -1z I4 4
I.
- -
Jo 1.60l
Bethe (1) has shown that electric coupling and mag- 1.1I
cn
I
m.10
netic coupling for a cricular hole have identical depend- N .09
.08 E
ence upon wavelength. In coaxial lines, in the dominant N .07
u
strength of the coupling, however, will vary and is given 0 -10 -20 -30 -40 -50
.03
-60
by COUPLING DR -
Power in main line X2/ Fig. 4-Dependence of the coupling for the Bethe-Hole coupler
upon the hole diameter. Wavelength was held constant during
this test. 2a=0.14X; the coupler was not compensated.
where C is a constant which depends upon the diameter
and the impedance of the coaxial lines, a is the radius II. LARGE COUPLING HOLES
of the coupling hole, and X is the wavelengtlh.
Experimental confirmation of the above statements is A. General Remarks
shown in Figs. 3 and 4. Fig. 3 shows angular dependence In an effort to study the effects of close coupling, a
of the electric and magnetic coupling through a small directional coupler was built using a 0.875-inch inside
hole (diameter of the coupling hole was equal to 0.14X). diameter coaxial line, and a coupling hole of 1.098
The two types of coupling were separated by placing inches in diameter. (The couplig hole is larger than the
shorts X/4 and X/2 away from the center of the hole. inside diameter of the tubing.) The ratio of the diameter
I.0
of the hole to the wavelength used was (2a/X) = 0.28,
THEORETICAL which is obviously outside the scope of Bethe's pre-
0
U 0* EXPERIMENTAL
dictions. Indeed, an experimental study of this coupler
showed a number of significant deviations from the sim-
<j 0.6
oC1SINE E vs.e ple theory discussed above.
04.
004 1 0 o 0
0 D0 - In the tests of the large hole coupler, it was found that,
(a) SWR (in voltage) of 1.38 was produced by the
~-0.2 presence of the hole in the main line. It was found, as
4
could have been-expected, that discontinuity could be
represented by a series inductance.
0
CONE0.1 0.2 0.3 0.4 0.5 0.6 0.7 08 0.9 1.0
(b) The angle at which directivity was best was no
Fig. 3-Dependence of the electric coupling E and the magnetic longer 600 but approximately 55°.
coupling H upon the angle between the two lines. (c) The directivity was very poor.
Since the discontinuity was of a series inductance
In the first of these cases, the magnetic field at the hole type, it was immediately apparent that the simplest
is zero, and the coupling is due to the electric field only. way to eliminate the effect of the discontinuity was to
In the second case, the electric field at the hole is zero, construct a T-section filter, making use of the series in-
and the coupling is due entirely to the magnetic field. ductance and adding an appropriate shunt capacity. If
Fig. 3 shows that the electric field is independent of the the characteristic impedance of this filter section is made
angle between the coaxial lines, whereas, the magnetic equal to the impedance of the line, then the reflections
coupling varies as cosine of the angle between the lines. from the hole should be eliminated. It is not imme-
Within small experimental errors, Fig. 3 demonstrates diately obvious that such compensation would also be
the validity of the elementary theory. beneficial in the case of the other two difficulties. How-
1950 Ginzton and Goodwin: Coaxial Bethe-Hole Directional Couplers 307
ever, it is plausible to think that since all of these diffi- C. Directivity of the Compensated Directional Coupler as
culties are due to the fringing fields, and if one of the a Function of Frequency
troubles is eliminated, then the others may be improved Fig. 6 shows the directivity of the compensated 1.098-
as well. While this reasoning is not on safe grounds, ex- inch hole directional coupler. It is to be observed that
periments have confirmed the general concepts. for wavelengths greater than 10 cm, the angle for opti-
B. Optimum Directivity Angle
mum directivity is independent of frequency, while the
relative directivity falls from its very high value at 10
There are many forms that an appropriate shunt cm and asymptotically approaches some moderately
capacity might take. For simplicity, a cylindrical probe high value.' The region above 10 cm represents the us-
was chosen to test the ideas described above. It consisted able range of this particular coupler.
of a 4-inch machine screw placed diametrically opposite
the hole in each of the two lines. The penetration of this u)
probe was varied, and the effect upon the angle of opti- w
uj 74 - -j
TO
n 3s
od
w
a 45 t _
62 cr
MV A :
25 _
4-28 TO. PROBE PENETRATION INTO EACH LINE -INCHES Fig. 6-The directivity and optimum angle for a properly com-
.140 .160 .180 .200 .220 .240 .260 .280 .3tO pensated large hole coupler as a function of wavelength. The
65 diameter of the hole=1.098 inch; inside diameter of coaxial
line=0.875 inch.
0~~~~~
>. 4/
SLIDING SHORT
-16 X0.1 800.2 0.3 70 d 04 , 605 6 so 0.7 40.0 0.8 0.9
! UNDER TEST
Cose
0~~~~
8/
a: SCALE OF e IN DEGREES.
680 7'0 60 5'0 4'0
where T= RC and p=jw. It is desired to evaluate the This will prove (3) and (4) valid for all values of n
coefficients a1, a2, * an-1 and bI, b2,
-
bn2 as func- greater than zero.
tions of n.
1 ,TFg 1
Fig. I Fig. 2
If we adopt the convention of letting am,n be the mth In Fig. 2 the box represents the n-mesh network
coefficient for the case of n meshes, it can be shown that whose transfer function and admittance are given by
the general a coefficient is given by equations (1) and (2). The driving point admittance at
point A is G G,+ Cp. The driving point impedance of
a =-(n m)! +
the (n+ 1) meshes is
=
(n - m)!(2m)!
1 1 + RGn+ Tp
*
Decimal classification: R143.2 XR386.2. Original manuscript re- Zn+l = R + G - -
ceived by the Institute, May 26, 1949; revised manuscript received, Gn+Cp
November 9, 1949.
t Fairchild Engine and Airplane Corporation, Farmingdale, N. Y. and the admittance of the (n+ 1) meshes is