operation suddenly become, on reversal, highly stressed were then removed and placed in an oven .

at 165°C for
cathode-insulator junctions (see lower portion of the graph approximately 200 hours with an applied direct stress of
of Fig. 2). Such high-field regions are capable of initiating 1 kV. The samples were then removed and rechecked in
electron emission from the cathode-insulator junction with a vacuum at 80-90°C for direct holdoff voltage. When
relatively small applied voltage. Until these sodium-depleted checked with the applied voltage in the same direction as the
layers become replenished by the reversal of the ionic polarising voltage, the first breakdowns occurred between
current flow or are short circuited by conducting-surface 25 and 40 kV, conditioning up to similar levels to those
previously obtained. With the applied voltage in the
opposite direction to the polarising voltage, the first break-
1OOO down occurred between 3 and 7 kV, and the sample showed
t=O only slow conditioning to higher voltages. At voltages of
about 10 kV, it was found possible to obtain large stable
800 prebreakdown currents after some conditioning (see Fig. 3).
If the applied voltage was removed for, say, 1 min and then
suddenly reapplied, the prebreakdown current did not
600
appear immediately, but built up to a steady value after a
delay of the order of 0-5 s. This is consistent with the time
400 estimated for the re-establishment of the high field across the
depletion layer under these conditions. The stable pre-
breakdown currents obtained at low reverse voltage were
200 accompanied by optical surface phenomena in the form
of fan-shaped glows propagating from discrete emission
points at the cathode-insulation junction.
g1
v sample length (1-05cm)^ Practical considerations: In practical high-voltage vacuum
-200 insulation, various forms of electrostatic screening are used
to suppress electron emission from the cathode-insulation
junctions by reducing the local electric field. Where
-400 insulators are subject to prolonged but reversible d.c. stress
(particularly at high temperature), the formation of high-
-600 resistance ion depletion layers in the insulator may consider-
ably reduce the effectiveness of such screening. This problem
may be satisfactorily overcome by impeding the electrolysis
-800 in the ceramic by the use of a lower alkali-oxide content and
the blocking of alkali-ion movement by the introduction of
more tightly bound less mobile ions into the composition.
-1000
H. GIBSON 20th June 1971

Fig. 2 Voltage distribution as measured using potentiometer GEC Nelson Research Centre
and probe along a 2 cm-diameter 1 cm-long h.v. porcelain Beaconhill, Stafford, England
test sample
Temperature = 150°C, applied voltage 1 +957 Vfor t = 0 to 337 h and—957 V References
for t = 337 to 669 h
1 LINDSAY, E. w., and BERBERICH, L. J.: 'Electrical properties in ceramics
as influenced by temperature', Trans. Amer. Inst. Elect. Engrs.,
flashover products (e.g. cathode metal-vapour deposits), 1948, 67, pp. 734-742
2 VALEYEV, KH. s., GAREVSKii, v. N., and KOSTYUKOV, N. s.: 'The
the voltage holdoff performance of the insulator will remain variation of the electrical strength of high voltage porcelain under
relatively poor. Experiments on small porcelain samples sustained high direct voltage', Elektrichestvo, 1963, 1, pp. 59-61
3 HAWLEY, R.: 'Solid insulators in vacuum: A review', Vacuum, 1968,
have confirmed this behaviour. In one experiment, six 18, pp.385-390
2 cm-diameter 1 cm-long porcelain cylinders were prepared
with fired-on platinum electrodes on their end faces, and the
insulating surface ground to give a well defined and unshielded
cathode-insulator junction. These samples were placed
between plane electrodes in a high-vacuum chamber and the
direct-voltage holdoff values measured at a temperature of
8O-9O°C. The first breakdowns occurred at 18-34kV, and
the samples rapidly conditioned up to 32-46 kV. The samples
THEORY OF V.O.R. ANTENNA RADIATION
bulk insulator
current (ionic)
PATTERNS
direction of
Indexing terms: Antenna radiation patterns, Antenna theory
polarisation
The radiation field produced by v.o.r. antennas consisting of
Alford loops located above a circular conducting ground
\surface plane is investigated theoretically and experimentally.
prebreakdown Geometrical theory of diffraction is applied to obtain the
theoretical expressions for the far field produced by such an
current (electronic) antenna. Within the range of approximation, the agreement
between theory and experiment has been found to be very good.

Introduction: The existing very-high-frequency omnirange

(v.o.r.) antenna system consists of four Alford loops suitably
-8 -6 -4 4 6
placed in a plane parallel to and above a circular conducting
I,u.A ground plane or counterpoise.1 Depending on the environ-
surface mental conditions, the entire antenna assembly is located at a
prebreakdown convenient height above the ground. The v.o.r. system
current (electronic) operates in the 100-120 MHz frequency range, and the
counterpoise has a diameter, typically, of 52 feet and is
located 75 feet above the ground. In one mode of operation,
Fig. 3 Vacuum breakdown tests on 2 cm-diameter 1 cm-long called the sideband mode, at any instant of time, only two
porcelain sample (typical result) of the Alford loops are excited with equal amplitude but
418 ELECTRONICS LETTERS 29th July 1971 Vol. 7 No. 15
opposite phase, so that the pattern of the antenna in the The symmetry of the system dictates that the far field
horizontal plane is a figure of eight. The elevation-plane produced by the antenna shown in Fig. 1 will also be
pattern has a maximum in the direction 30-40° above the polarised in the <j) direction. In the present case, the two
horizon and has a minimum in the axial direction. Alford loops are separated by a small distance d and are
Although this antenna has been in use for v.h.f. omnirange
for many years and its performance is approximately known
from experimental considerations, there has been no
satisfactory theory for their patterns. In this letter, a theory
for the free-space radiation patterns of the v.o.r. antenna
has been developed by utilising concepts of geometrical
theory of diffraction. The theory is then compared with the
measured patterns for different sizes of the ground plane.
The agreement between theory and experiment has been
found to be very good.
Theoretical analysis: For the purpose of theoretical analysis,
the v.o.r.-loop counterpoise antenna is replaced by a point
source with appropriate far-field variation placed above
the counterpoise, as shown in Fig. 1. The free-space radiation
180
Fig. 3 Far-field elevation pattern of Alford-loop counterpoise
antenna
kh= 2-75, kd= 0-92, kA = 17-92,/= 1080 MHz
o o o theoretical
experimental

excited with signals having equal amplitude but opposite

phase. Thus f{9, ft) can be explicitly written as
/(0 s 0) = 2isin(fcdsin0cos0 (2)
With this value of f(6, ft), the horizontal plane (9 = TT/2)
Fig. 1 Theoretical model for the Alford-loop counterpoise pattern of the antenna will be afigureof eight having maxima
antenna along the x direction. We do not go into the details of the
method here, since it has been discussed elsewhere2 for the
field of the point source (in the absence of the counterpoise) special case f{6, ft) = 1. Following the method discussed in
is represented by Reference 2, it can be shown that the complete expression
for the far field in the vertical plane and valid in the region
4p no, a sine- (1) 0 < 9 < nis given by
ka exp {i(fcR-?r/
where EA ~ (3)
R, 6 and </> = usual spherical co-ordinates of the far- R
field point with origin at the point where
source
rj0 = intrinsic impedance of free space V2
k = 2n\X = free-space propagation constant \cos9\ sin <f>0/2 ... . r 0
(4)
70 = amplitude of current in each Alford loop V(nkr0 sin 9) lkr
<>)L

a = equivalent radius of each Alford loop

F ° ( 0 ) = exp {ikr0 sin (9- fa)} f e x p (in - ? - I dt
f(6, ft) = pattern function which is determined by -CO \ 1
the method of excitation and orientation
of the two Alford loops.
- exp {ikr0 sin (0 + ^0)} exp I in — I dt (5)
The time dependence exp (-jcot) is assumed.
Note that the field in eqn. 1 is polarised in the <j> direction.
exp Q(TT/2 - kA sin 9)} f(90, ft) cos* 4>Q -f(8, ft) sin* 9
V (1 - sin 9) cos 4>Q - sin 9
expp'fc/lsinfl) f(90, ft) cos* <f>0 .
cos ^O + sin 9

270
p2 = (7)

It should be remembered that, in obtaining the above

expressions, the basic assumptions of geometrical theory of
diffraction have been used. In the present case, it mainly
implies that eqn. 3 is valid for kA> \. Eqns. 3-7 give the
180 far field produced by a loop counterpoise antenna that is
Fig. 2 Far-field elevation pattern of Alford-loop counterpoise
being used in the existing v.o.r. system. To the best of this
antenna
author's knowledge, the expressions given above are new and
kh = 2-75, kd= 0-92, kA = 6-32,/= 1080 MHz appear here for the first time. In the following Section, we
o o o theoretical make a comparison between the results obtained by numerical
experimental
ELECTRONICS LETTERS 29th July 1971 Vol. 7 No. 15 419
computation of eqn. 3 for <]> = 0° and the corresponding pattern near 0 ~ nil for kA = 51-69 are attributed to the
measured results. outside pattern range and the feed system. The measured
results shown in Figs. 2 and 3 were taken with an antenna
Comparison between theory and experiment: The measured pattern range inside a large anechoic chamber. The agree-
free-space far-field elevation patterns of Al ford-loop counter- ment between theory and experiment for kA = 6-32
poise antennas are shown in Figs. 2, 3, and 4 for three (2/1 = 22 in) is not as good, although it may be considered
selected values of the counterpoise diameter. All the patterns to be fair over most of the region (see Fig. 2). The reason
have been measured in the xz plane (Fig. 1) and at for this is attributed to the fact that, owing to the small size
1080 MHz. The corresponding theoretical patterns obtained of the counterpoise, the ray-optics approximation becomes
by numerical computation of eqns. 3-7 are also shown in poor.
Figs. 2—4 for comparison. The normalised dimensions of the
antenna given in Fig. 3 correspond to those of an actual Conclusion: The elevation-plane radiation field produced
v.o.r. antenna at 108 MHz. The agreement between theory by existing v.o.r. antennas has been obtained theoretically
and experiment for kA = 17-92 {1A = 5-2 ft) and kA = 51-69 by applying the geometrical theory of diffraction. Within
(2A = 15 ft) may be considered to be very good. The minor the range of approximation used in the theory, the agreement
lobes in the pattern in directions 9 > n(2 and the kink in the between theory and experiment has been found to be very
good.
e=o Acknowledgment: The author wishes to thank J. E. Ferris
for his help in taking the measurements. The research
reported in theis paper was sponsored by the US Federal
Aviation Administration, Washington DC, under contract
FA 69 WA 2085, project 33O-OO1-O3N.

D. L. SENGUPTA 28th June 1971

Department of Electrical Engineering
270
University of Michigan
Ann Arbor, Mich. 48105, USA
180' References
Fig. 4 Far-field elevation pattern of Alford-loop counterpoise
antenna 1 ANDERSON, s. R.: 'V.H.F. omnirange accuracy improvements', IEEE
kh = 2-75, kd = 0-92, kA Trans., 1965, ANE-12, pp. 26-35
51-69,/= 1080 MHz 2 SENGUPTA, D. L., and WESTON, v. H.: 'Investigation of the parasitic
o o o theoretical
experimental loop counterpoise antenna', ibid., 1969, AP-17, pp. 180-191

SYNTHESIS OF RC NETWORKS CONTAINING As can be seen from Fig. 2, the nodes of the flowgraph
OPERATIONAL AMPLIFIERS USING for such a network and the operational amplifiers in the
network are in one to one correspondence.
FLOWGRAPH TECHNIQUES Furthermore, a branch gain in the flowgraph is equal to an
Indexing terms: Operational amplifiers, Lumped-parameter admittance in the network. The synthesis technique is based
networks, Network synthesis, Transfer functions essentially on these properties. In this letter, we develop a
By using Coates's flowgraphs, a new approach to operational- method for synthetising a biquadratic voltage transfer
amplifier-/? C-network synthesis is given. The general function. However, any transfer function of order n (n > 2)
biquadratic voltage transfer function is realised by this method.
Furthermore, it is shown by an illustrative example that can be obtained by the cascade connection of biquadratic
obtaining any other desirable transfer function directly from networks, which is preferable from the sensitivity point of
this realisation is possible.
view.8
The application of graph theory to the analysis and synthesis
of networks has yielded many interesting and significant
results in recent years. In most topological synthesis
techniques,1-2 the signal flowgraph3'4 has been used as a V, •
tool, and the advantages of flowgraphs51 6 appear to have
been overlooked. Recently, Robinson and Methiwalla2 have V 2 +o-
given a realisation procedure for a voltage transfer function
by the use of signal flowgraphs. They start with a signal -o*Vn
flowgraph and, by the use of several transformations, they
obtain a suitable one which realises the given function.
In this letter, we give a realisation procedure based on
flowgraphs which is straightforward and does not give
negative element values. 'n •
First, we shall discuss a method of analysis of operational-
amplifier networks, which will be used for the synthesis
procedure. A typical building block of an operational-
amplifier-RC network and the corresponding flowgraph
I i
are shown in Figs. 1A and 1B, respectively. The flowgraph is Fig. 1 A Single-input operational-amplifier configuration
obtained from the following equation, which can easily be
written by the inspection of the building block:

Yf Vo + Yi V], + ... + Yn Vn = 0 (1)

As an example, consider the network shown in Fig. 2A, which
is known as Lovering's network.7 The corresponding flow-
graph can be drawn by inspection of the network and is
shown in Fig. 2B. The voltage transfer function F 0 /F t can then
be obtained by Coates's formula6 as follows:
Vo ) B o ( -)l ) 1 YO OYAA YB-YA
(2)
- I) 2 Yo Yc+ ( - I) 1 Yo YD YC- YD
Fig. 1 B Flowgraph for Fig. 7A
420 EELCTRONICS LETTERS 29th July 1971 Vol. 7 No. 15