EP0032332B1 - Radiofrequency power divider, and radiofrequency devices, especially solid-state devices, using the same - Google Patents

Description

The invention relates to a power distributor for radio waves. It relates in particular to the variants of such a distributor made in the solid state for application to microwave frequencies. It also relates to devices using one or more of these distributors.

Power divider and recombiner circuits are used whenever it is sought, for example, to add the powers of several independent amplifiers, in particular when a power higher than that supplied by existing amplifiers is required.

The circuits proposed so far generally do not meet a condition considered to be fundamental, which is that of allowing proper operation, even in the presence of identical mismatches on the loads of the different branches, or channels, of the divider, of the transistors for example, in high frequency distributed amplifiers; in the case of such amplifiers, an operation is always sought in a frequency range as wide as possible, the range in which such mismatches usually manifest themselves.

Among these mismatches, there are two kinds: some, as has just been said, resulting from the variation of the impedance of each elementary amplifier, consecutive to the displacement of the operating frequency, others resulting, at frequency given, disparities between amplifiers in the same batch.

In the context of the invention, there will be a major interest in the first of them, the most important by far with respect to the natural dispersion of manufacturing, in the current state of the art. We will therefore deal with the case of somehow symmetrical mismatches, where all the components loading the branches are supposed to be identical at the outset, then vary in the same way according to the frequency, the variations between elements of the batch being considered as of the second order compared to the previous ones.

Generally, to cope with the mismatch, we absorb the excess power, that reflected by the mismatched element, in a resistive member mounted between the branches of the distributor.

In the branches of a divider, like the one discussed above, these resistors were in the prior art mounted in parallel on each of the branches of the divider, as will be specified in the description which follows, and in some way so symmetrically with respect to these branches, so that the assembly, set for an initial state with zero absorption, only provided the desired absorption function in the event of an asymmetrical imbalance of the branches, that is to say -to say in particular of random mismatches of the branches, such as those arising from dissimilarities between amplifying elements.

However, such an arrangement was not adapted to the power absorption imposed by a symmetrical adjustment of the branches.

The article by Dr. W.C. TSAI (1978 IEEE MTT-S International Microwave Symposium Digest) has proposed a power distributor capable of solving this problem. However, this distributor constituted by Wilkinson type dividers operates on a narrow bandwidth.

The invention, as claimed, has for its object a power distributor capable of compensating for even symmetrical imbalances of its branches, in particular on the occasion of the impedance variations presented by the active elements which charge them when the frequency moves, this distributor being capable of operating over a much greater bandwidth than the distributors of the prior art. Such a distributor can a fortiori also compensate for the non-symmetrical imbalances that may arise.

• Figure 1 la forme générale, schématisée, d'un circuit diviseur de l'art antérieur ;
• Figure 2 le schéma d'un circuit diviseur de l'invention ;
• Figure 3 et 4 des compléments au schéma de la figure précédente ;
• Figure 5 une vue d'une réalisation en technologie microbandes d'un amplificateur utilisant le répartiteur de l'invention ;
• Figure 6 une variante de la figure précédente ;
• Figure 7 une vue d'une autre réalisation dans la même technologie.
• Figure 8 une vue d'un exemple d'utilisation du répartiteur de l'invention dans un amplificateur à plusieurs étages ;
• Figure 9 une vue schématique d'une autre forme de réalisation du répartiteur de l'invention.

The invention will be better understood by referring to the description which follows and to the attached figures which represent:

• Figure 1 the general form, schematically, of a divider circuit of the prior art;
• Figure 2 the diagram of a divider circuit of the invention;
• Figure 3 and 4 supplements to the diagram in the previous figure;
• Figure 5 a view of an embodiment in microstrip technology of an amplifier using the distributor of the invention;
• Figure 6 a variant of the previous figure;
• Figure 7 a view of another embodiment in the same technology.
• Figure 8 a view of an example of use of the distributor of the invention in a multi-stage amplifier;
• Figure 9 a schematic view of another embodiment of the distributor of the invention.

All the dividing devices of which fig. 1 gives the diagram can be considered as deriving from the montage of Wilkinson, about which we will refer to the article by IRE Trans. MTT3, page 116, January 1960. The diagram represents the case of the power division starting from a high frequency generator R _G debiting on N branches, each charged by the same impedance of use to the second order near, R _s ; each line has m sections of impedances Z _{1 ′} Z ₂ ... Z _m , between which are mounted in parallel, as shown in the drawing, dissipative elements R _oi , R ₀₂ ... R _om . All these branches are voluntarily represented of identical constitution even if in reality they contain some differences of little importance between them. In the case of microwaves, the sections of impedances Z _{1 ′} Z ₂ ... Z _m are quite simply sections of coaxial lines, for example quarter wave, that is to say whose length is equal to a quarter of the average wavelength of the high frequency operating wave supplied by the generator R _G.

The larger the operating frequency band of the device, all other things being equal, the greater the number m.

This being the case, a brief analysis of the behavior of the device is carried out below for equal mismatches of the loads R _s .

The device being provided for given initial conditions, in accordance with the above, that is to say with N branches formed by m section of lines, all identical to each other, (same impedances Z ₁ ... Z _m same elements dissipatives R ₀₁ , R _0m and same charges R _s ), and adjusted so that the charges R _s are perfectly adapted, no power is reflected in the branches, in which the points located at the same distance from the source, such that L ₁ , M ₁ , N ₁ , are in phase on all the lines; no energy is then dissipated in the resistive elements placed in bridge between the lines. It is no longer the same if unequal adaptations occur in the branches, because then this phase concordance is no longer preserved and a current flows between the points such as L ₁ , M ₁ , N ₁ , L ₂ , M ₂ , N ₂ , etc.

On the other hand, in the event of symmetrical mismatches of the charges R _S , all the lines, identical, undergo the same disturbances and the points L, M ₁ Ni are found in phase, as well as M ₂ N ₂ P ₂ etc, as they were at the time of the initial setting. Under these conditions, the additional powers reflected by the mismatched charges cannot be absorbed by the resistors, at the terminals of which there appears no potential difference. There are various variants to the Wilkinson assembly described above on which we will not extend, all these variants, which fall under the same principle, having the same drawbacks.

We will also mention Lange, IEEE Trans MTT 17, December 69, page 1150, as prior art of these power distributors, as well as the patents of Marvin H. White and Gary H. Hoffmann granted in the United States under the respective numbers 3.593 .174 and 3.963.993.

To remedy this state of affairs, the invention provides an arrangement in which is ensured, permanently, that is to say whatever the state of the loads, a phase shift allowing the desired absorption. According to the invention, the distributor comprises several dividers arranged in cascade, each of them being connected by an output branch to the next; each of them further comprises, at its outlet, a certain number of branches on which the loads are mounted.

All these branches, as well the one used for the connection with the following divider, as those closed on the loads, will be called secondary branches, the first named being the main secondary branch. In addition, according to the invention, this branch is connected to the input branch in the next divider, called the primary branch, through a phase shifter if necessary; this divider also has, like the previous one, secondary branches, including a main one connected to the primary path of the following divider, possibly through a phase shifter, etc.

Figure 2 gives the general diagram of the distributor of the invention, limited to two dividers, D (i - 1) and D (i), diagram on which we find the main secondary pathways in B (i - 1) and B (i), the other secondary routes in C (i - 1) ... C _m (i - 1) and C ₁ (i) ... C _m (i), the primary routes A (i - 1) and A (i) and the phase shifters 0 (i - 1) and 0 (i), designated by the same letter as the phase shift that they introduce on the primary branch on which they are mounted. A network of resistive elements, represented as such and left without reference for clarity, is used, in addition, as in the prior art as described with reference to FIG. 1. These elements connect the secondary pathways to the secondary pathway main, star in the case of the example, any other arrangement being possible within the framework of the invention. Finally each of the dividers introduces itself a phase shift which will be called 0 _D with the index of rank i as for the rest of the elements above.

The loads are mounted, as we have said, at the ends of the secondary branches, according to the distribution deemed best in relation to the use for which the distributor is intended. In the example of FIG. 2, these loads are distributed in a number equal to the output of the dividers; but any other distribution could be envisaged where the index m would not have the same value for all the divisors.

In operation, the device is first adjusted, including the phase shifters 0 (i), so that, for practically zero reflections in the secondary branches, the absorption is negligible in the resistive elements. When the charges on which the secondary branches C are closed are mismatched, and substantially all of the same amount, the phase differences introduced between the outputs of the successive dividers cause the waves reflected in the main secondary channel B (i) and in the different secondary pathways C (i) are not in phase; a potential difference then appears between point a on the one hand and points b, c, d on the other hand, and an absorption of the power reflected in the resistive elements is made possible. Identical variations in load in the branches of the distributor are thus possible without interfering with operation, contrary to what takes place in devices of the same kind of the prior art.

The structure is completed (i = n) by connecting the main secondary route to a load. Two assemblies are possible: either we introduce a phase shifter Ø ≠ 0, preferably as in the above Ø = (2K + 1) π / 2, between A and the main secondary channel B of the last divider, the other secondary branches all being related to other charges; or all the secondary branches of the last divider, including the main secondary path, are directly connected to loads: see Figures 3 and 4 where all the other secondary routes are indistinctly designated by C.

The distributor of the invention appears as a progressive structure composed of a succession of phase dividers offset from one another.

We have spoken of phase in the above: it must of course be specified that, so that in the absence of reflections in the lines, no current flows in the resistive elements, there must also be an amplitude relationship between the voltages appearing at insertion points a, b, c, and d; this relation is P _o R _o = P _(i) R _(i) (1) if R _(i) and R _o denote the characteristic impedances of the main secondary line and that of all the other secondary lines, and P _(i) and P _o the high frequency powers in these lines.

We now come back to the phase shift. The total phase shift between a secondary branch C (i) and a secondary branch C (i - 1) is 0 (1) + 0 _D (i) with the previous notations.

The phase shift will obviously be effective only if this sum differs from kπ, which is always easily achieved. We can take in particular Ø (i) + Ø _D (i) = (2 k + 1) π / 2.

The reflected power is, in this case, absorbed in the resistive elements, and one can make so that a collective mismatch of the charges has no influence on the reflection coefficient at the input, on the left of the figure 2, which is that of using the distributor as a divider, for a source whose power arrives in the direction of the arrow.

This characteristic is particularly advantageous in the case where the payloads are transistors used for amplification over a wide bandwidth. We know, in fact, that the gain of a transistor decreases by about 6 dB per octave depending on the frequency. At the lowest frequencies in the range, there is therefore a gain gain and the input power should be reduced accordingly. This is achieved either by introducing selective losses into the input circuit, or by voluntarily mismatching at the bottom of the band. The latter solution is the simplest, but it is necessary to absorb the power thus reflected. The structure according to the invention makes it possible to obtain this result in a particularly simple manner even in the case of a large number of transistors and over a wide bandwidth. Individual mismatches are also absorbed in the same way as in a phase divider.

Reference has hitherto been made mainly to the structure of the invention in its use as a power divider. Of course, the same structure can be used as a power recombiner provided that the different power sources have suitable relative phases. This condition can be ensured, in the case of an amplifier, thanks to a distribution of the input powers by a divider of the same type. Two distributors of the invention will then be used, one acting as a divider and the other as a recombiner. FIG. 5 shows an exemplary embodiment thereof in the micro-band technique. The view shown is a plan view of the useful part of the insulating substrate 10 on which the transmission lines are made by linear conductors applied to this substrate, coated with metal on its opposite face. It is an amplifier device for microwave with four active active elements consisting of four transistors T ₁ , T ₂ , T ₃ and T ₄ .

These separately manufactured elements are, in the example, attached to the substrate.

The example in the figure implements two distributors of the invention as described above, designated as a whole by 1 and II; the first of them operates as a power divider to the four transistors T ₁ to T ₄ from a source not shown, located at the input of part 1 (arrow), and the second as a recombiner of the powers of output of these four transistors. Each of them has three phase dividers, such as D (i) of the preceding diagrams, with two secondary branches each, including the main secondary branch (m (i) of the preceding diagrams is therefore here equal to 1), and a terminal line.

For the set I, these dividers carry marks D as in the above and consist of lines (1,7), (2,6) and (3,5) as shown in the drawing; the terminal line 4 is charged by the last transistor; the drawing shows the location of the resistive elements without mark. Similar benchmarks are assigned to Set II, with the changes necessary to avoid confusion. The recombined power comes out in the direction of the arrow.

The phase shift Ø (i) + Ø _D (i), is chosen in the example equal to π / 2 by the use of lines 1, 2, 3, 4 of lengths equal to a quarter wave for the central frequency ; here we have 0 (i) = 0 because for each divider point B (i - 1) of the diagram in Figure 2 is confused with point A (i): points u, v, w in the figure; the characteristic impedances of these lines are all 50 ohms; those of lines 5, 6 and 7 are, on the contrary, unequal to one another and are chosen respectively from 50, 25 and 16.6 ohms to satisfy condition (1).

It often happens, however, that the impedances of the active elements, such as T ₁ T ₂ T ₃ T ₄ , are very low (a few ohms) and that the adaptation to 50 ohms is difficult. Figure 6 gives a variant of the arrangement of the previous figure particularly suitable for this case, in which, in each divider, an impedance transformer is provided on the secondary branch, as shown in the drawing, where these transformers carry the marks t ₁ t ₂ t ₃ ; each of them has a second line 01, 02, 03 coupled to the secondary lines 1, 2 and 3.

FIG. 7 shows, also in micro-band technique, another example of application of the distributor of the invention, to the addition of the powers of several diodes with negative resistance. tive; the variant of the figure uses impedance transformers as described with reference to the previous figure.

The output power is collected in the direction of the arrow on a frequency fixed by a “locking” signal using a circulator (curved arrow), frequency of the order of 10 to 20 gigahertz for example.

In the figure the diodes bear the marks d ₁ d ₂ d ₃ d ₄ ; those of the other elements, without designation, are easily deduced from the preceding figures.

We have described above embodiments applied to four active elements; it goes without saying that .these examples are not limiting, the number of these elements being able, within the framework of the invention, to exceed this value or to be less.

Another application of the distributor of the invention consists in the repeated use of dividers, recombiners, D ₁ R ₁ D ₂ ... R ₃ connected in series. FIG. 8 shows a particularly simple example, where each of the dividers and recombiners has only two branches, and where each recombiner and the divider which follows it are directly linked together. The example shows a three-stage amplifier with two transistors each; the gain in decibels is multiplied by the number of stages between input and output (arrows). An advantage of the structure shown is to minimize the necessary impedance transformations.

In the examples of embodiment in the solid state of the distributor of the invention given above, lines have been shown in the form of strips deposited on a substrate. It will be noted that these lines could, in the context of the invention, be replaced by cells with localized elements - self - inductors and capacitors - bandpass, low pass or high pass type, as shown in the diagram of the Figure 9, where these elements have been left unmarked; on the other hand, we have reproduced on this figure the reference points necessary to enable the correspondence to be drawn between this and the diagram of FIG. 2. Each of the cells is one of the divisors of this figure. The part of the figure to the left of point A _o is an impedance adapter, and that to the right of point B _{2 is} a terminal line. All secondary lines are designated by the same letter C.

In addition, it has been recognized that in a hybrid form the active elements are diodes or transistors attached to the substrate; within the limits of the invention, these elements could be integrated into the substrate.

The invention also includes, on the contrary, the case of embodiments by conventional means, other than those of the solid state, both with regard to the active elements and the transmission elements.

Claims (8)

1. A high frequency power distributing assembly for distributing power from one channel to several other channels and for recombining in a common channel the power supplied by several channels, in a given operation frequency band, comprising more than two dividers ((D(i - 1), D(i)) arranged in series and each comprising one input or main branch (A(i - 1), A(i)) and at least two outputs or secondary branches (B(i - 1), ), C₁(i - 1), C₂(i - 1), C_m(i - 1); B(i), C₁(i), C₂(i), C_m(i)), with one secondary branch called main secondary branch constituting the input or main branch of a next following divider, charges being connected or powers being applied to the terminals of the secondary branches which are not main secondary branches, in such a way that the characteristic impedance R_(i) of the main secondary channel and the characteristic impedance R_o of the other secondary channels are connected via the relation

in which P_(i) and P_o represent the high frequency power of these lines, and that Ø_D(i) ≠ 2 k π, in which Ø_D(i) constitutes the phase shift introduced by the divider of the rank i, characterized in that all the secondary branches of a divider other than the main secondary branch (C₁(i - 1), C₂(i - 1), C_m(i - 1)) and (C₁(i), C_m(i)) are connected via a dissipating element other than zero to a given common point of the main output in such a way that the simultaneous reflections on the outputs are absorbed in said dissipating element.

2. A distributing assembly according to claim 1, characterized in that a phase shifter (Ø(i - 1), Ø(i)) is inserted into the main branch (A(i - 1), A(i)) of at least one of the dividers (D(i - 1), D'(i)), the phase shift induced by said phase shifter being such that Ø(i) + Ø_D(i) ≠ 2 k π.

3. A distributing assembly according to claim 1, preferably intended for microwaves, characterized in that, said assembly being realized according to the microstrip technology using linear conductors supported by one face of an insulating substrate (10), the opposite face of which is metal coated, the dividers (D'(i=3), D'(i=2), D'(i=1) ; D(i=1), D(i=2), D(i=3)) are made of aligned U-shaped conductors, the central part of an U being situated at the level of the end of the branches belonging to the preceding U, whereas its lower branch (31, 21, 71) or its upper branch (1, 2, 3) is situated in alignment with the upper branch (81, 11, 61) or the lower branch (8, 7, 6) respectively of the latter and in contact therewith, these branches (31, 21, 71, 1, 2, 3) extending all over substantially a quarter wavelength corresponding to the central operational frequency, and that the upper branch (51) or the lower branch (5) of the first divider (D'i=3) and (Di=3) is extended by a line (41 or 4), the length of which is substantially equal to a quarter of said wavelength.

4. A distributing assembly according to claim 3, characterized in that in each divider an impedance transformer (t₁, t₂, t₃) is provided in the secondary branches except the main secondary branch.

5. A distributing assembly according to claim 3, characterized in that supplementary linear conductors ((01), (02), (03), (04)) are disposed parallelly to the upper branches (1, 2, 3) of the U and of the terminal line and constitute impedance transformers (t₁, t₂, t₃, t₄) therewith, and that the charges (T₁, T₂, T₃, T₄) are connected to these supplementary conductors at the end of the branches of the U and at the end of the terminal line.

6. The use of a distributing assembly according to any one of claims 1 to 5 in a microwave device for recombining the powers of n generators (T₁, T₂, T₃, T₄), characterized in that the assembly comprises n - 1 dividers (D'(i=1), D'(i=2), D'(i=3)) with two secondary branches ((11, 71) ; (21, 61) ; (31, 51)), the first divider (D'(i=3) being connected via one (51) of its secondary branches to one of the generators (T₁) via a phase shifter (41 ).

7. The use of a distributing assembly according to any one of claims 1 to 5 in a microwave device for the division of a power into n partial powers, characterized in that the assembly comprises n - 1 dividers (D(i=1), D(i=2), D(i=3)) with two secondary branches ((1, 7) ; (2, 6) ; (3, 5)), one (5) of the secondary branches of the final divider (D₃) being connected to one (T₄) of the charges (T₁, T₂, T₃, T₄) via a phase shifter (4).

8. The use of a distributing assembly according to any one of claims 1 to 5 in a microwave device for dividing the power of a source into partial powers supplying the inputs of n amplifiers (T₁, T₂, T₃, T₄) and for recombining the output powers of these n amplifiers (T₁, T₂, T₃, T₄), characterized in that it comprises a first distributing assembly (I) comprising n - 1 dividers (D(i=1), D(i=2), D(i=3)) with two secondary branches ((1, 7) ; (2, 6) ; (3, 5)), one (5) of the secondary branches of the final divider (D₃) being connected to a charge (T₄) via a phase shifter (4), the charges being constituted by the input of the amplifiers (T₁, T₂, T₃, T₄), and a second distributing assembly (II) comprising n - 1 dividers (D'(i=3), D'(i=₂), D'(i=1)) with two secondary branches ((31, 51) ; (21, 61) ; (11, 71)), one (51) of the secondary branches of the first divider (D'(i=3)) being connected to a generator (T₁) through a phase shifter (41), the generators being constituted by the outputs of the amplifiers (T₁, T₂, T₃, T₄).

Images