Amplifiers – With semiconductor amplifying device – Including distributed parameter-type coupling
Reexamination Certificate
2001-06-05
2003-06-24
Choe, Henry (Department: 2817)
Amplifiers
With semiconductor amplifying device
Including distributed parameter-type coupling
C330S295000, C330S296000
Reexamination Certificate
active
06583672
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates to quasi-optic grid arrays, and, in particular, to techniques for controlling the distribution of dc bias across such arrays and improved arrays incorporating such techniques.
2. Description of Related Art
Broadband communications, radar and other imaging systems require the transmission of radio frequency (“RF”) signals in the microwave and millimeter wave bands. In order to efficiently achieve the levels of output transmission power needed for many applications at these high frequencies, a technique called “power combining” has been employed, whereby the output power of individual components are coupled, or combined, thereby creating a single power output that is greater than an individual component can supply. Conventionally, power combining has used resonant waveguide cavities or transmission-line feed networks. These approaches, however, have a number of shortcomings that become especially apparent at higher frequencies. First, conductor losses in the waveguide walls or transmission lines tend to increase with frequency, eventually limiting the combining efficiency. Second, these resonant waveguide cavities or transmission-line combiners become increasingly difficult to machine as the wavelength gets smaller. Third, in waveguide systems, each device often must be inserted and tuned manually. This is labor-intensive and only practical for a relatively small number of devices.
Several years ago, spatial power combining using “quasi-optics” was proposed as a potential solution to these problems. The theory was that an array of microwave or millimeter-wave solid state sources placed in a resonator could synchronize to the same frequency and phase, and their outputs would combine in free space, minimizing conductor losses. Furthermore, a planar array could be fabricated monolithically and at shorter wavelengths, thereby enabling potentially thousands of devices to be incorporated on a single wafer.
Since then, numerous quasi-optical devices have been developed, including detectors, multipliers, mixers, and phase shifters. These passive devices continue to be the subject of ongoing research. Over the past few years, however, active quasi-optical devices, namely oscillators and amplifiers, have evolved. One benefit of spatial power combining (over other methods) using quasi-optics is that the output power scales linearly with chip area. Thus, the field of active quasi-optics has attracted considerable attention in a short time, and the growth of the field has been explosive.
It is believed that the first quasi-optical grid array amplifier was a grid developed by M. Kim et al at the California Institute of Technology. This grid used 25 MESFET differential pairs, demonstrating a gain of 11 dB at 3 GHz. As shown in
FIG. 1
, a typical grid amplifier
10
is an array of closely-spaced differential pairs of transistors
14
on an active grid
12
sandwiched between an input and output polarizer,
18
,
24
. An input signal
16
passes through the horizontally polarized input polarizer
18
and creates an input beam incident from the left that excites rf currents on the horizontally polarized input antennas
20
of the grid
12
. These currents drive the inputs of the transistor pair
14
in the differential mode. The output currents are redirected along the grid's vertically polarized antennas
22
, producing a vertically polarized output beam
30
via an output polarizer
24
to the right.
The cross-polarized input and output affords two important advantages. First, it provides good input-output isolation, reducing the potential for spurious feedback oscillations. Second, the amplifier's input and output circuits can be independently tuned using metal-strip polarizers, which also confine the beam to the forward direction. Numerous grid amplifiers have since been developed and have proven thus far to have great promise for both military and commercial RF applications and particularly for high frequency, broadband systems that require significant output power levels (e.g. >5 watts) in a small, preferably monolithic, package. Moreover, a resonator can be used to provide feedback to couple the active devices to form a high power oscillator.
One challenge in the design of active grid arrays relates to DC biasing of the differential pair cells in the array. Unfortunately, since any practical design requires the whole array to be biased with a single bias bus, one specific problem that affects the operation of active grid arrays is the uneven distribution of bias voltage across a given array of cells or a row of cells in an array. Due to the finite conductivity and limited thickness of metal lines used in actual circuits, grid arrays that require appreciable bias current to be distributed among many active devices (e.g., transistors) from a single source at the edge of the grid may suffer from unequal distribution of DC power. Specifically, the cells in the middle of the array tend to get less bias current that those closer to the dc source. This DC power distribution problem limits the effectiveness of large, high-power grid array components (e.g., grid amplifiers and grid oscillators).
It would thus be desirable to have a grid array in which the power distribution across the entire array could be controlled, by either providing an equal distribution of DC bias to each cell in the array or by providing a different, but desired, power distribution at the control of the array designer. The present invention fulfills this need.
SUMMARY OF THE INVENTION
The present invention, which addresses this need, resides in a method of controlling the distribution of a bias in a periodic array of electronic devices, such as a quasi-optic grid array. The method includes the steps of first sensing the voltage drop on the bias line resulting from the intrinsic resistance in the bias line at either all or selected electronic devices in the array, and then applying a prescribed bias voltage via an impedance network at the control input of either selected or all devices in the array based upon the sensed voltage drop. In this way, the bias voltages to each control input of either the selected or all electronic devices in the array can by tailored and thus will not suffer from an undesirable, non-ideal, and/or uneven bias distribution across the array.
More particularly, a method of controlling the distribution of bias in a periodic array of active cells that is biased with a DC bias supply, a bias supply line and a bias-return conductor line, is disclosed. The method includes sensing the voltage difference between consecutive cells on the bias-return conductor line resulting from the internal conductor resistance in the bias-return conductor line; and based on the sensed voltage difference, applying to each cell a prescribed control bias voltage. The prescribed control voltage is applied via an impedance network at each cell along the array that allows the control bias voltage to track the voltage difference on the return line.
The impedance network of the present invention includes a voltage divider circuit disposed between the bias-return current line and a reference voltage conductor return line, and a tracking resistance inserted in the reference-voltage conductor return line and is connected to the voltage divider circuit.
In one preferred embodiment, the impedance network maintains a relatively constant control bias voltage distribution to each cell in the array relative to the sensed voltage on the bias return line at each cell. In this way, the input control biasing voltage (V
gs
in the case of FET's) is the same across the entire grid array, regardless of the size of the array. In one specific embodiment, the active cells are differential pair amplifiers and the periodic array is a grid amplifier.
An improved periodic active grid array is also disclosed by the present invention. The array includes (a) a plurality of active cells that are combined at their outputs in a periodic arrangement via bias-supply
Deckman Blythe C.
Rutledge David B.
California Institute of Technology
Choe Henry
LandOfFree
Method for controlling bias in an active grid array does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method for controlling bias in an active grid array, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for controlling bias in an active grid array will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3092902