OBJECT OF THE INVENTION
The present invention relates to a new structure for high-power and low
insertion losses microwave filters with symmetrical or asymmetrical transfer function
response to implement in rectangular waveguide H-plane configuration.
STATE OF THE ART
The increase in capacity, complexity, and RF power employed in satellite
communications and broadcast repeaters, has forced the use of sophisticated filter
transfer functions. Regarding to the out of band rejection a more and more
demanding specification is required at present for innovative applications. In the
same way, to save mass and volume it is mandatory not to use high degree filters to
fit these specifications. Therefore the inclusion of transmission zeros at real finites
frequencies is essential.
It is known from US5926079 of MOTOROLA a ceramic filter that introduces
finite frequency transmission zeroes in a filter's transfer function. Drawbacks in this
case are that only two transmission zeroes could be implemented, another one being
that two additional cavities must be added to the N cavities that implement the N
degree transfer function in order to implement the transmission zeroes, thus
increasing mass and dimensions.
It is known from US4360793 of RHODES and CAMERON an extracted pole
filter that allows to implement a transfer function with finite real-frequency
transmission zeroes, but it presents several drawbacks: main drawback is that phase
shifting waveguide sections have to be introduced, which makes a complex filter
layout and increases mass and size, another drawback being that only symmetrical
responses are possible, finally, mechanisms must be included in order to eliminate
degenerate modes that are present because of the electromagnetic mode of operation.
Furthermore another disadvantage is a costly manufacturing process.
The basic synthesis theory of filters with extracted poles for symmetrical
responses was developed in J. D. RHODES, R.J. CAMERON.: 'General extracted
pole synthesis technique with applications to low-loss TE011 mode filters', IEEE
Trans. Microwave Tech., Sep. 1980, vol. 28, n°9, pp. 1018-1028; and later on
generalized in R. J. Cameron.: 'General Prototype Network-Synthesis Methods for
Microwave Filters', ESA Journal, 1982, vol.6, pp. 193-206 for asymmetrical
responses. In J.R. Montejo-Garai.: 'Synthesis of N-Order Filters with N
Transmission Zeros at Real Frequencies by means of Extracted Poles , Electronics
Letters, Jan. 2003, vol.39, n°2,pp. 182-183, an extension is developed in order to
extract the maximum number of transmission zeros in N-degree filters with either
symmetrical or asymmetrical responses.
CHARACTERISATION OF THE INVENTION
The present invention seeks to overcome or reduce one or more of the above
problems by means of an electric transmission structure for implementing a transfer
function of N degree that can incorporate finite real-frequency transmission zeros,
the structure comprising: a main rectangular waveguide without change in height (H-plane
configuration) wherein the body of the structure comprises a plurality of
resonant cavities placed adjacent to each other and connected with inductive irises
and at least one cavity with a double electrical behaviour, i.e. operating as a resonant
cavity in transmission at central frequency of the passband and simultaneously
introducing a controlled transmission zero out of the passband.
A further object of the present invention is to provide a new structure for
cavity filters assuring a drastic reduction in mass and volume in comparison with the
all-pole transfer functions with the same rejection specification.
Another object of this invention is to provide a new cavity arrangement with a
double controlled electrical behaviour that allows to introduce transmission zeros at
finite real frequencies.
Yet another object of the invention is a synthesis technique for the synthesis
of N-degree filters with N-transmission zeros at real frequencies by means of
extracted poles.
In accordance with the invention there is provided an electric transmission
structure in rectangular waveguide for implementing a transfer function with
transmission zeros at finite real frequencies. This structure comprises a plurality of
resonant cavities placed adjacent to each other with inductive irises between adjacent
cavities, and at least one cavity with a double controlled electrical behaviour
connected to the "classical" resonant cavities by means of inductive irises.
The structure according to the invention has the advantage of allowing the
filter to be mechanized in a simple and very compact construction without slots or
critical dimensions to ensure a high RF power handling capability, increasing the
multipactor margin in space applications. In addition, the cost and the manufacturing
dimensional tolerance sensitivity are reduced. This construction allows using large
cavities to increase the Q in order to maintain low insertion losses at high
frequencies.
A further advantage of the invention is that fine adjustment of all elements,
i.e., cavities and coupling between them, is possible by means of tuning screws.
These are not part of the structure but elements to compensate for the mechanical
tolerances.
BRIEF DESCRIPTION OF THE DRAWINGS
The characteristics and advantages of the invention will become more clear
with a detailed description thereof, taken together with the attached drawings, in
which:
- Figure 1 shows the low pass prototype filter of 3th degree with 3 transmission
zeros at finite real frequencies according to the invention,
- Figure 2 shows the low pass prototype filter of 4th degree with 2 transmission
zeros at finite real frequencies according to the invention,
- Figure 3 shows the band pass prototype filter of 4th degree with 2
transmission zeros at finite real frequencies corresponding to low pass
prototype in figure 2,
- Figure 4 is a view of an exemplary rectangular waveguide realization of the
network shown in figure3,
- Figure 5 is a longitudinal view along line VV in figure 4,
- Figure 6 shows an example of insertion (A) and return losses (B) of a filter
with the characteristics described in figures 2 and 3.
DESCRIPTION OF THE INVENTION
In order to verify a band pass filter rejection specifications, the first task is to
generate the transfer function that in the more general case can be always expressed
as the ratio of two finite-degree polynomials with complex coefficients (from here
on, the degree of the transfer function is the degree of the numerator polynomial).
The evaluation of this mathematical response must fit with the out of band rejection
specifications and the return losses in the passband.
Once this task has been carried out, the next step is to synthesize a low-pass
prototype network, i.e., to obtain the values of the electrical components such as
capacitors, inductors, admittance/impedance inverters, frequency invariant
reactances/susceptances and transmission lines. The response of this electrical
network must be the same as that of the mathematical transfer function. For the case
of band pass filters a transformation is necessary to translate the low-pass response to
the considered frequency band. Equally important, since mechanical structure is the
rectangular waveguide, a one-to-one correspondence between the electrical circuit
elements and the physical resonant cavities and irises is necessary.
FIG. 1 shows as an example, the low pass prototype of a 3th degree network
with 3 transmission zeros in the real frequency axis, wherein ψ1, ψ2, ψ3, ψ4 and ψ5
represent electrical lengths corresponding to transmission lines, K is an admittance
inverter, 1 is a unity inverter, L1, L2, and L3 are inductors, and jX1, jX2, and jX3 are
frequency invariant reactances.
FIG. 2 shows another example, the low pass prototype of a 4th degree
network with 2 transmission zeros in the real frequency axis, wherein ψ1, ψ2, ψ3
and ψ4 represent electrical length corresponding to transmission lines, C1 and C2 are
capacitors, L1 and L2 are inductors, jX1 and jX2 are frequency invariant reactances
and jB1 and jB2 are frequency invariant susceptances.
FIG. 3 shows the equivalent rectangular wave guide H-plane structure to
implement the network corresponding to FIG. 2, wherein ψ2* and ψ3* represent
electrical length corresponding to transmission lines, K1, K2 and K3 are admittance
inverters, C1 and C4 are the equivalent circuit elements of the cavities with double
electrical behaviour, C2 and C3 are the equivalent circuit elements of the classical
resonant cavities. FIG. 4 is a view of an exemplary rectangular waveguide realization
of the network shown in FIG. 3 according to the invention. 1 and 2 are the input and
output of the structure respectively. C1 and C4 are cavities with double electrical
behaviour. C2 and C3 are resonant cavities. K1, K2 and K3 are inductive irises for
coupling adjacent cavities.
FIG. 5 is a perspective representation of a cross section along the line VV of
the rectangular waveguide of FIG. 4 wherein the position of the cavities can be
clearly observed.
Note that this structure is composed by two different types of resonant
cavities; C2 and C3 are two inductive coupled (shunt reactive iris K2) transmission
cavities. C1 and C4 are a new structure having a double controlled electrical
behaviour; each one operates as a resonant cavity in transmission at central
frequency of the passband and simultaneously introduces a controlled transmission
zero out of the band. C1 and C4 are inductive coupled to C2 and C3 by means of
shunt reactive irises K1 and K3 respectively. Therefore, it is possible to guarantee the
required return losses and at the same time to introduce a transmission zero in the
desired position. If the rectangular waveguide elements are correctly dimensioned the
electrical response of the structure will be very similar to that predicted by the
mathematical filtering function. However, in the synthesis process an attention will
be paid to the circuital values, in order to obtain the most adequate results because of
the mechanical constraints.
The dimensions of the cavities with double behaviour can therefore be
obtained by way electromagnetic simulation in conjunction with optimisation of their
structure. This simulation can provide different dimensions depending on the
requirements of the design in each particular case. In the example of figures 4 and 5
these double behaviour cavities C1 and C4 are shown to represent an extension to a
side of the filter structure, clearly being different from the rest of the cavities, e.g. C2
and C3.
Based on design parameters, the cavities with double behaviour may adopt a
variety of structures, for example instead of being extended to a side of the
waveguide, i.e. having a width larger than the general width of the waveguide
structure as shown in figures 4 and 5, they could have a width being smaller than the
general width of the rectangular waveguide (not shown).
The use of minimum phase networks like the one shown above, is more
desirable because element-value sensitivity is less and network complexity is
reduced. If cross-couplings are employed, the designer does not have specific control
over the positions of the zeros because there is not a one to one correspondence
between zeros and cross-couplings. For this reason such kind of structures are very
sensitive and difficult to adjust. However, by implementing the extracted-pole
technique every transmission zero, is controlled independently. This is a very
important asset from the engineering point of view in order to minimize the
sensitivity of the network for mass production.
The synthesis technique is based on a systematic process to extract the
(attenuation) poles. In order to deal with asymmetric electrical responses, where the
transmission zeros on the imaginary axis of the complex plane are asymmetrically
disposed, it is necessary to extract them individually.
Filters exhibiting their maximum number of finite transmission zeros at real
frequencies (N zeros corresponding to N degree) make possible to design transfer
functions with very high selectivity. Since an N-degree filter with N transmission
zeros has finite insertion losses at infinite frequency (the transfer function is a
rational expression with polynomials in the numerator and denominator of the same
N degree), a change in the impedance level must be introduced in the network to
assure this behaviour. This impedance level change has its circuital representation as
an ideal transformer of relation 1:K, as shown in FIG.1.
The synthesis process is composed by two different steps. The initial part of
the synthesis procedure is carried out in terms of the transfer function of the filter and
includes several extraction cycles (as many as finite real transmission zeros) in order
to extract the finite poles, each cycle comprising the steps of determining the phase
lengths of the unity inverters, the residue of every pole (shunt series resonator), the
capacitors and the invariant shunt reactances to cope with the asynchronously tuned
network.
Once the element values of the extracted pole prototype network have been
obtained the synthesis procedure improves further to transform the prototype
network into the equivalent rectangular waveguide structure arrangement according
to the invention. The transformation converts the phase lengths of the transmission
lines into an admittance inverter plus a new phase length, as shown in FIG.3.
Figure 6 illustrates a simulation of the electromagnetic response of the
waveguide structure wherein curve A represents insertion loss and curve B represents
return loss. In this figure the effect of a cavity with double behaviour in introducing
controlled transmission zero out of the band can be seen in the deep minimal
insertion loss shown in curve A.
Filters obtained according to the invention can be connected in arrangements so as to
provide a multiplexing or demultiplexing network. Examples of such connections are
by connecting a plurality of filters by rectangular waveguides sections and tee's
(connection devices in the form of T, known in the art) the height of each is equal to
that of the rectangular waveguide of the filters, i.e. in H-plane configuration. An
alternative connection is obtained by means of transmission line sections and tee's,
and in particular said transmission line sections can be coaxial.