SYSTEMS AND METHODS FOR MANUFACTURING STACKED CIRCUITS AND TRANSMISSION LINES

FIELD OF THE INVENTION

The present invention relates to microwave and millimeter wave components circuits and systems, and more particularly, but not exclusively to stacked circuits and transmission lines and methods for the manufacture thereof.

BACKGROUND OF THE INVENTION

Passive and active RF components are integral to microwave and millimeter wave systems. Generally these components are designed based on the manufacturing methods and tolerances within the build process. Traditionally such processes include a computer numerically controlled (CNC) machine process or a die-cast process depending on the volume of the waveguide components to be made. These methods can suffer from multiple deficiencies such as the method of manufacturing being serial and not batch processed. For example, for CNC, the geometry of each machined part needs to be programmed into the machine for the build, and tolerances of the build depend on the tool cutters and temperature of the machine which can vary substantially. In addition, to achieve high resolution and accurate parts, the machine speed is often lowered and operated by a skilled machinist, increasing the overall cost. For die-cast processes, the resolution that can be achieved is often much coarser than the designs require, and unacceptable variation from die to die can reduce overall yield. Multiple part assemblies can also be complex and add to further errors in positional accuracy of pins, dowels, and features. The above drawbacks contribute to the high cost of passive and active microwave and millimeter wave components and modules, with recent years showing little improvement in the overall build process.

Planar circuits are alternative structures which can include microwave printed circuit boards with dielectrically loaded microstrip or coplanar structures. However, drawbacks for these circuits include insertion loss and lack of isolation between signal lines compromising signal integrity.

Another major drawback with both 3D machining and planar circuits is the lack of compactness or functional density. The machining of transmission lines such as waveguide channels are only performed in 2D surfaces in split waveguide formations. This limits the full 3D functionality where the active elements can only be placed in specific locations dictated by machining orientation. In planar circuits, a limitation of 3D multilayer parts includes poor thermal management due to high dielectric load between the interconnects and lack of inclusion of active elements such as integrated circuits in embedded architectures. Furthermore, planar multilayer circuits are heavy and can become a large burden for the overall system.

Millimeter-wave and THz waveguide structures made from cross-linked photoresist SU-8 are disclosed in Tian, Y, Shang, X & Lancaster, M J 2014, “Fabrication of multilayered SU8 structure for terahertz waveguide with ultralow transmission loss,” Journal of Micro/Nanolithography, MEMS, and MOEMS, vol 13, no. 1, 013002.,10.1117/1.JMM.13.1.013002, hereafter “Tian.” Such an approach is limited both in the process capability and the resulting structures. Metallizing a photoresist using methods such as electroless plating to create a sliced waveguide structure has many limitations.

A first limitation in the art is that a photoresist plastic such as SU-8 has very low thermal conductivity, so the electronic chips cannot dissipate the heat they generate through the plastic. A second is that a thin metal on plastic has a CTE mismatch preventing such structures from surviving the thermal cycles needed for consumer, industrial, and aerospace applications. A third is that such plastic structures and metallized plastic are not compatible with standard chip interconnect processes such as wirebonding. A fourth is that fusing metallized layers of plastic is difficult due to the inability for such structures to endure substantial mechanical compression without delamination, cracking, and peeling of the metal coatings on the plastic. A fifth limitation is the mechanical robustness of a stacked plastic part particularly when thin or small intricate features are required. A sixth limitation is the poor resolution offered after metallization of patterned plastic parts. In some cases, one might try electroplating rather than electroless plating on the plastic. As the parts are metal seeded and electroplated, current crowding effects unevenly electroplate the structure depending on the locations on the part exposed to the electroplating anode. This is even a larger problem for thicker electroplating in excess of 3 μm which would be required for mechanical strength of the parts. A seventh limitation is the accurate alignment of multi-stacking of parts due to the above (sixth limitation) over-plating of corners and edges. An eighth limitation is the overall number of stacks and their ordering and available features that can be created or used in a single monolithic plastic part. As each layer is added on top of a cured and exposed lower layer, it is chemically attacked throughout the fabrication process which will affect the interfaces between each layer causing delamination and poor adhesion. In addition, and more limiting in this eighth limitation, is that when attempting more than one layer of photoplastic in a monolithic construction, any added layer's photoexposure must fall inside the planar area of the previous layer's photoexposure so that the previous layer is not inadvertently photoexposed in an undesired region. A ninth limitation is the mechanical robustness of the metalized-plastic parts. For example, as the individual parts come together quite often mechanical screws are used to fixture parts and force the layers together for a no-gap connection. The metalized plastic parts cannot be tapped for a screw or pressed hard against each other for a firm contact. A tenth limitation is the lack of combined plastic (or non-conductor) and metal (or conductor) on the same integrated layer, which can be needed to isolate transmission lines from each other electrically and is an important attribute as the layers become more functionally capable. Thus, due to these limitations and more, there remains a need in the art for devices and methods that can achieve the above requirements while overcoming the limitations currently present in the art.

SUMMARY OF THE INVENTION

In one of its aspects the present invention may provide a stacked waveguide structure comprising a plurality of metal waveguide slices, which may be solid metal. Each waveguide slice may include at least one waveguide cavity disposed therein. Selected pairs of the waveguide slices may be disposed adjacent one another, with the waveguide cavity of each slice of a selected pair registered to one another so the waveguide cavities of the selected pair of slices communicate with one another to provide at least one waveguide within the stacked waveguide structure. The waveguide cavity of a selected slice may extend through the depth of the slice to provide openings on opposing surfaces of the slice or may extend partially into the depth of the slice. A selected slice may include two waveguide cavities oriented orthogonal to one another within the slice. Further, a selected pair of waveguide slices may each have a face disposed adjacent one another, with at least a portion of a waveguide disposed orthogonal to, or parallel to, the faces. The at least one waveguide may include a waveguide splitter and/or waveguide combiner. A plurality of waveguides may be provided in the stacked waveguide structure which do not communicate with one another.

In another of its aspects, the stacked waveguide structure may include an integrated circuit chip disposed in electromagnetic communication with a waveguide input and/or a waveguide output of the stacked waveguide structure. In addition, a waveguide transition may be provided between the integrated circuit chip and the waveguide input and output; the transition may include a waveguide cavity therein disposed in electromagnetic communication with the waveguide output. The transition may also include a probe disposed within the waveguide cavity of the transition, with the probe configured to convert electromagnetic energy disposed within the waveguide cavity of the transition into electrical energy within the probe. The probe may be disposed in electrical communication with the integrated circuit chip.

In yet a further of its aspects, the present invention may provide a method of creating a stacked waveguide structure, comprising depositing a plurality of layers on a substrate. The layers may include one or more of a metal material and a sacrificial mold material, thereby forming a plurality of solid metal waveguide slices each having at least one waveguide cavity disposed therein. The method may include the step of aligning and joining the plurality of waveguide slices to one another so the waveguide cavities of the slices communicate with one another to provide at least one waveguide within the stacked waveguide structure.

BRIEF DESCRIPTION OF THE DRAWINGS

The foregoing summary and the following detailed description of exemplary embodiments of the present invention may be further understood when read in conjunction with the appended drawings, in which:

FIG. 1A schematically illustrates an exemplary WR3 waveguide 16-way back-to-back power splitter/combiner in accordance with the present invention having a physically folded structure, with the structure of the waveguide air cavities depicted (i.e., an “air model”);

FIG. 1B schematically illustrates a wiring diagram representing the splitter/combiner of FIG. 1A;

FIG. 1C schematically illustrates the splitter/combiner of FIG. 1A with a portion cutaway to reveal 4-way splitter/combiner portions;

FIGS. 2A-2B schematically illustrate an exemplary design process for splitter portions of the waveguide power splitter/combiner of FIG. 1A, including both the electrical and subsequent mechanical design as a stacked architecture;

FIG. 3A schematically illustrates an isometric view of a physical realization of the power splitter/combiner of FIG. 1A designed using the process of FIGS. 2A-2B, and comprising 23 slices to form a single stack containing all electrical and mechanical features needed for the compact operation of the structure;

FIG. 3B schematically illustrates a cut-away view of the power splitter/combiner of FIG. 3A, with the waveguide air cavities shown by the shaded structures;

FIG. 3C schematically illustrates a transparent view of the power splitter/combiner of FIG. 3A;

FIGS. 4 and 4A schematically illustrate a single slice of the stacked waveguide architecture of the power splitter/combiner of FIG. 3A showing complex internal cavities that provide for high RF performance;

FIG. 5 schematically illustrates exemplary batch manufacturing process steps in accordance with the present invention for fabrication of a slice of a stacked waveguide architecture, such as a “multi-strata” slice shown in FIG. 4 or these same type of slices shown stacked together in FIG. 3A, for example;

FIGS. 6A-6B schematically illustrate alternative examples of how a compact design of the present invention can be sliced in different directions for batch manufacturing, depending on tradeoffs of fabrication versus assembly into a stack;

FIG. 7 illustrates the calculated high frequency performance of a mathematical model of the power combiner of FIG. 1 showing low insertion loss due to accurate build structure and compactness;

FIG. 8 illustrates the calculated sensitivity analysis of the mathematical model of FIG. 7, but with slices placed out of alignment to show negligible performance variation of the combiner, indicating a high degree of manufacturing tolerance;

FIGS. 9A-9C schematically illustrate an exemplary waveguide to integrated circuit transition piece in accordance with the present invention, with FIG. 9A showing a top view, FIG. 9B showing a bottom view, and FIG. 9C showing a vertical waveguide to coplanar transition of FIGS. 9A and 9B, but with an IC in place;

FIGS. 10A-10B schematically illustrate top and bottom views, respectively, of an exemplary waveguide to coplanar transition in accordance with the present invention having a waveguide ridge and a horizontal probe;

FIGS. 11A-11B schematically illustrate top and bottom views, respectively, of an exemplary waveguide to coplanar transition designed to accommodate a quartz electric field probe;

FIGS. 12A-12B schematically illustrate top and bottom views, respectively, of an exemplary translation piece in accordance with the present invention built to accommodate various geometries of integrated circuits;

FIG. 12C schematically illustrates the transition piece of FIG. 9A stacked on top of the translation piece of FIG. 12A; and

FIG. 13 schematically illustrates the power splitter/combiner of FIG. 3A enclosed within a housing with a mounting plate disposed thereon for attachment of transition pieces thereto.

DETAILED DESCRIPTION OF THE INVENTION

In one of its aspects the present invention relates to multilayer transmission line devices and methods for design and manufacture thereof In certain aspects the present invention may provide methods for the design and manufacture of passive and active RF circuits, such as power amplifiers, oscillators, phase shifters, filters, time delay units, diplexers, etc. Such methods may also provide light weight and compact multilayer transmission lines, including waveguide, coax, microstrip, and grounded coplanar waveguide structures, for example. In another aspect, the present invention may facilitate manufacturing of the aforementioned structures with complex geometries without the need for substantial computer aided design (CAD) file manipulation.

Some exemplary device configurations in accordance with the present invention may include a complex 3D network design with few gaps between the interconnect transmission lines, such as waveguides or coaxial lines. In such an exemplary configuration, the transmission lines may be folded and ground layers can be shared between each folded line to generate high signal line density per unit volume, FIG. 1. The transmission lines of the present invention may be formed with an omni-directional propagation coax or rectangular coax or polarized propagation such as a rectangular waveguide. Both electric and magnetic channels can be used to compact the design even further by folding the path in multiple directions within a certain design volume.

Referring now to the figures, wherein like elements are numbered alike throughout, FIGS. 1A-1C schematically illustrate an air-model of an exemplary device in accordance with the present invention, a 16-way folded back-to-back rectangular waveguide splitter/combiner 100. (Since FIG. 1 is an air-model, “structural” elements, e.g., elements 130, 135, 140, 145, represent air or a void space in the final solid part.) The splitter/combiner 100 may include feed points for an input 110 and an output 120, and a splitter arm 130 operably connected to the input 110 to divide the input into 4 outputs, FIG. 1C. The location of the input and output 110, 120 may be selected with regard to the final configuration that is desired. The splitter arm 130 may include electric and magnetic field waveguide bends 104 to effect the folded structure and may comprise a generally planar structure disposed in the x-z plane, for example, to use the available volume. The waveguide bends 104 may be configured to keep the path length of the signal completely symmetric along the splitter arm 130 for efficient power splitting throughout the arm 130.

Four 4-way H-plane splitter modules 140 may be operably connected to the splitter arm 130 to further divide the signal into 16 portions. The H-plane splitter modules 140 may have a generally H shape and be disposed in planes that are perpendicular to the plane of the splitter arm 130, such as the y-z planes, for example. Each 4-way H-plane splitter module 140 may include four outputs 141, so that the four 4-way H-plane splitter modules 140 collectively have 4 times 4, or 16, total outputs 141. Thus, the splitter arm 130 and four H-plane splitter modules 140 may cooperate to provide the 16-way splitter of the splitter/combiner 100.

The sixteen outputs 141 of the 4-way H-plane splitter modules 140 may be connected to active or passive components, e.g., an amplifier IC chip, external to the splitter/combiner 100, FIG. 1B. Signals from such external components may then be returned to the splitter/combiner 100 to be combined at output 120. In particular, the splitter/combiner 100 may include four H-plane combiner modules 145 each of which includes four inputs 146 to receive the outputs from the external components, FIG. 1A. The combiner modules 145 may have a generally planar H shape and be disposed parallel to the splitter modules 140. Each combiner module 145 may also include a single output operably connected to a combiner arm 135 to combine the four outputs from the combiner modules 145 into the splitter/combiner output 120, FIG. 1B. The combiner arm 135 may be disposed in a plane perpendicular to the planes of the combiner modules 145, such as the same plane as the splitter arm 130. Additional structures may be added to the splitter/combiner 100 to provide additional functionality, such as filters, bias feeds, signal amplification, and phase shifting based on the overall system requirements. In certain exemplary configurations, the splitter and combiner arms 130, 135 may be identical in size and shape, FIG. 1C, as may the H-plane modules 140, 145.

FIGS. 2A-2B illustrate a flowchart 200 of an exemplary design and manufacturing process in accordance with the present invention, where the splitter portion 101 (i.e., splitter arm 130 and H-plane splitter modules 140) of the 16-way folded back-to-back rectangular waveguide splitter/combiner 100 is shown, though the process could be used for other structures. This process may also be used for the combiner portion (i.e., combiner arm 135 and H-plane combiner modules 145).

The electrical and RF design may be performed prior to mechanical modeling, during which design performance may be optimized, such as a low insertion loss of each segment along with low reflected power from each port, i.e., low return loss. The exemplary design process may include the design of the 4-way splitter arm 130, step 202, FIG. 2A. The 4-way splitter arm 130 may provide a first branching. The 4-way splitter arm 130 may then be further broken down into a 2-way electric field waveguide splitter 131 and an electric and magnetic field waveguide bend 104, which may be separately optimized for RF performance, steps 204, 206. The 4-way H-plane splitter module 140 may also be broken down into a 2-way waveguide power splitter module 106 for performance optimization, steps 208, 210. Once the initial design of the 16-way splitter portion 101 is complete, and its RF performance such as insertion loss, isolation between ports and return loss optimized, the electrical and RF air-model may be converted into a solid, mechanical model, step 212.

As a first step the 16-way splitter portion 101 may be mirrored to provide the 16-way combiner (i.e., combiner arm 135 and H-plane combiner modules 145) to include the output network in a compact 3D volume, step 214. The overall design is folded to maximize use of the volume in which the waveguide splitter/combiner components are disposed. The mechanical model may be sliced across the volume into manufacturable slices that will fit together to form the final 3D volume 16-way splitter, step 224. The slice locations within the volume may be carefully engineered to match the fabrication process rules allowing for the high yield manufacturing, step 216. Once the slices locations are defined, each slice may then be modeled, step 218, with an associated layout. A mask set may then be generated where all the slices are placed on a single mask set and reviewed for accuracy and tolerance definition to the process design-rules, step 219. Based on the manufacturing tolerances, the slice thicknesses and their shape can be modified and re-simulated using 3D electromagnetic design tools to adhere to the fabrication process. This process may be iterated and trade-offs in performance versus manufacturing tolerance can be made for a high quality system build. Following the design optimization a sensitivity analysis, step 220, is performed to allow for performance variations due to the manufacturing and assembly tolerances to be minimized. The design for manufacturing cycle, step 224 may be an iterative process resulting in a final design of the slices where the final performance is insensitive to variations generated by the fabrication or assembly process. The design for manufacturing cycle may be completed by generating a test-plan, step 222, in which the step by step assembly and testing of the unit is described.

A particularly desirable manufacturing technology for use in fabricating the mechanical model is the metal-air-dielectric microfabrication PolyStrata® process. (U.S. Pat. Nos. 7,948,335, 7,405,638, 7,148,772, 7,012,489, 7,649,432, 7,656,256, 7,755,174, 7,898,356, 8,031,037, 8,698,577, 8,742,874, 8,542,079, 8,814,601 and/or U.S. Application Pub. Nos. 2011/0210807, 2010/0296252, 2011/0123783, 2011/0181376 and/or 2011/0181377 are incorporated herein by reference in their entirety, and hereinafter called the “incorporated PolyStrata® art.” As used herein, the term “PolyStrata” refers to the devices made by, or methods detailed in, any of the aforementioned and incorporated U.S. Patents and Published Applications.) Other technologies, such as computer control machining, laser forming, wire electrical discharge machining, and so forth may provide different approaches for fabricating some parts.

FIGS. 3A-3C schematically illustrate an exemplary physical implementation of the air-model design of FIGS. 1A-1C to provide a waveguide block 300 with a 16-way combiner network interwoven with a 16-way divider network. In particular, FIG. 3B illustrates how the input 110, output 120, 4-way H-plane splitter and combiner modules 140, 145, splitter and combiner arms 130, 135 of the air-model of FIGS. 1A-1C may be disposed within the waveguide block 300.

The waveguide block 300 may comprise multiple slices 301 with each slice 301 fabricated independently based on the manufacturing process chosen. Each slice 301 may contain a respective portion of the air-model structures (e.g., input port 110, output port 120, 4-way H-plane splitter and combiner modules 140, 145, splitter and combiner arms 130, 135) of the air-model of FIGS. 1A-1C. Once the slices 301 have been fabricated, the slices 301 may be aligned and assembled to each other with a high degree of tolerance to provide a single rectangular volume. To aid in the alignment, registration features, such as dowel holes 310 may be provided in each slice 301 into which precision dowels may be inserted, which may optionally align the slices 301 to a secondary structure such as a flange, heat sink or integrated circuits. Alignment features may be internal or external, such as external grooves or notches 312, 314. Optional cavities 302 which do not communicate with the waveguide structures (e.g., structures 130, 135, 140, 145) may be provided to reduce the weight of the waveguide block 300 and provide air waveguide(s) through which the alignment of the slices 301 may be measured. For instance, a particular cavity 302 may be present at the same location in each and every slice 301 to provide, upon assembly, a waveguide that extends along the entire length of the waveguide block 300. After assembly of the waveguide block 300, energy can be launched through the waveguide formed by the assembled cavities 302 and the output power measured to determine insertion loss. Low insertion loss would indicate proper alignment of the slices 301, and conversely, a high or unacceptable insertion loss would indicate that the slices 301 are not well aligned.

FIGS. 4, 4A schematically illustrate a representative slice from the stack of slices illustrated in FIGS. 3A3C, which in this particular case is the fifth such slice 400 from the front of waveguide block 300, FIG. 3C, where the reference numerals have the same meaning as variously discussed in connection with FIGS. 1A3C. The selected slice 400 illustrates how particular portions of the rectangular waveguide splitter/combiner 100 air structures may be divided into, and disposed within, a particular slice 400. For instance, at the location of the fifth slice 400, portions of the splitter arms 130, 131 may be present, as well as portions of the 4-way H-plane combiner module 145 and combiner arm 135. The various structures contained within a slice 301, 400 may optionally extend through the depth of the slice, such as portion of the splitter arm 130, for example. Other structures may extend only a portion of the depth into the slice 301, 400, such as the 4-way H-plane combiner module 145, for example. Since the 4-way H-plane combiner module 145 does not extend the full depth, the remaining portions of the combiner module 145 may be present in an adjacent slice 301, so that when slice 400 is combined therewith, the waveguide associated with the combiner module 145 is complete. In addition, further structures may be provided to enhance performance of the device, such as septums 416, 424 which may provide wave impedance matching for low resistive power combining (or splitting, for septums present in the 4-way H-plane splitters 140).

The slices 301 may be fabricated with features such as those shown in FIGS. 4, 4A using PolyStrata® copper microfabrication multi-layered process, where each layer deposited in the process can be overlapped or exposed to build the complex layout of the design required for high electrical and mechanical performance. Specifically, layer upon layer of high resolution copper strata may be electroplated through a sacrificial plastic mold or template followed by a planarization step. Each deposited layer may then be aligned to the other as the slice 301 is fabricated and can be patterned with any shape.

The PolyStrata process may be particularly well-suited, because alignment of each deposited copper layer to another within the slice 301 can be achieved with much higher precision than required for the waveguide block 300. This allows for complex features to be built in each slice 301, and for 3D volumetric complexity to increase with each slice 301 that is stacked and bonded together. In this multilayered approach micromachined, RF cavities may be built interior to, or enclosed within, a slice 301, even though the cavity may not be accessible from either face of the slice 301. In addition, the PolyStrata® manufacturing technology allows for various metals to be incorporated into the slice 301 such as copper, silver, nickel, or gold and others depending on the requirements. Passivation layers may also be added to each slice 301 either on the surfaces only or the entire structure. The passivation may be dielectric or conductive, such as metals, for mechanical and electrical improvements to the structure. Some metals may be added to the surface to increase surface-to-surface bondability, such as adding a gold surface coating onto a copper fabricated structure. Bonding of copper surfaces to copper surfaces has been demonstrated under pressure. This can be accomplished at elevated temperatures, as well as room temperature when the surfaces are clean. Publications on surface activated bonding, the use of ultra-thin and mono-layer coatings to prevent oxidation exist in the literature. It should be clear various techniques can be used to join the independent slices 301 without causing substantial deformation to their mechanical dimensions.

FIG. 5 illustrates an exemplary batch manufacturing process of the slices 301 in accordance with the present invention, in which all the slices 301 needed for the waveguide block 300 can be included in a single fabrication mask set, allowing the slices 301 to all be built at the same time on separate areas of wafers or panels. The batch processing allows for higher resolution and alignment part-to-part, since all parts see the same fabrication process. The manufacturing cost may also be much lower than other processes that generally build parts in series like computer numerically controlled machines.

Turning to FIG. 5 in more detail, a photopolymer 502 can be spun on a carrier wafer 501 and patterned by an appropriate ultraviolet or other wavelength exposure, Step 1. The photopolymer 502 may be developed to define a template 503, Step 2, which may be electroplated to fill the template 503 with a metal 504, Step 3. Once the electroplated metal has reached a level above the height of the template 503, the metal may then be planarized and ready for the repeat of the lithography process of Steps 1-3. For example, a second layer of the photopolymer 505 may be deposited, Step 4, exposed and developed to provide an additional template 506, Step 5, which may then be filled with a metal 507 and planarized at Step 6 in a manner similar to that at Step 3. This process may be repeated as many times as needed to provide additional layers, such as a layer comprising a photopolymer 508 and a metal 509 as illustrated in Step 7. Once the layers needed have been processed, the photopolymer of the templates 503, 506, 508 may be removed to expose the air filled electroplated part 510, Step 8. The part 510, which may be a slice 301, may be released from the carrier wafer 501 following a selective etch process allowing stacking of multiple parts 510 together to form higher functionality circuits and systems. Thin seed layers that may be used whenever electroplating is to be done on any previous region that is non-conductive. Thus before each photopolymer template 503, 506 is formed, one may typically provide a thin seed layer that can be removed when the template material is removed or when each layer is planarized. Details of the PolyStrata® process are discussed in the other patents referenced herein.

Additional aspects of the fabrication process of the present invention include the high resolution alignment of the slices 301 to each other, such as through dowel holes 310, and control over the surface roughness of each slice 301, both of which are achievable via the PolyStrata® process. In this fabrication process the surface of each copper layer may be ultra-flat and smooth through a chemical mechanical polishing allowing for a high level of contact between slices 301 as they come together. This is important, because any gap between the slices 301 can reduce the performance required through high frequency leakage paths created in between the slices 301. Optionally, after assembly the slices 301 may be electroplated to metalize the rectangular waveguide block 300 and seal the inside channels (e.g., 130, 135, etc.) of the waveguide block 300. Depending on the system needs, the outside or exposed interior surfaces of the waveguide block 300 may also be electroplated, immersion plated, or passivated using an insulating material for environmental proofing considerations.

Other aspects of the fabrication process may include permanent attachment of the slices 301 together during assembly. Possible approaches may include metal-to-metal compression bonding which can be assisted through high heat and/or ultrasonic power, epoxy attach, or eutectic bonding, for example. The slices 301 may be permanently or temporarily attached to each other or other machined parts using various combined techniques allowing sections to be removed or replaced as necessary.

Once the air-model of the combiner/splitter 100 is created, the model may be sliced in a variety of different orientations. For example, as illustrated in FIGS. 6A-6B, respectively, the slice orientation may be perpendicular to the longitudinal axis of the combiner/splitter 100 or may be parallel to the longitudinal axis. Slicing parallel to the longitudinal axis, FIG. 6B, requires fewer slices but a larger area for each slice, while slicing perpendicular to the longitudinal axis, FIG. 6A, requires more slices but each are smaller in cross-sectional area. The slices in FIG. 6A may be 1 mm thick, for example. The tradeoff between number and slices 301 and size of each slice 301 may be based on the complexity of the circuit or device and the fabrication process. In general, yield and uniformity of the manufacturing process will determine the best option.

In some configurations, the arrangement or ordering of the slices 301 through the overall system stack can be changed to create differing sub-components or a different system altogether. The arrangement or ordering of the slices 301 can also be used to validate the performance of the system components independently before full assembly and characterization. Furthermore in some instances the slices 301 can be rotated or flipped to create other structures, for example, filters with multiple poles that can be reconfigured based on system need. An advantage of using these separate multi-layer slices 301 as building-block pieces and aligning and stacking them using means such as dowels, is that the assembly can be tested for performance and re-built or adjusted as needed before the parts are more permanently committed to an arrangement. The slices 301 can be configured or reconfigured from an inventory of such “building blocks” to rapidly create custom system configurations. This is a particular advantage over the alternative of milling where extremely high precision and suitably low surface roughness CNC milling of bulk metals is a slow and serial production processes requiring machines that are often hundreds of thousands of dollars, for example some of those made by Kern Microtechnik Gmgh in Germany. Prototyping time and cost can be greatly reduced in the approach of the present invention, such that custom hardware can be made from an “off the shelf” inventory of suitable “slices.”

The quality performance metrics for the design of FIGS. 1A-3C are shown in FIG. 7 where the scattering parameters of a 3D electromagnetic simulation is plotted versus frequency showing low insertion loss 36, isolation 37, and high return loss 38, indicating a high degree of impedance match across the frequency of 225-240 GHz. The design can be expanded to scale in frequency while maintaining a similar high degree of performance. A main reason for the low insertion loss 36 and isolation 37 between the ports is not only a high degree of precision in manufacturing, but also the actual compact size of the waveguide block 300. This is in direct contrast to the size and accuracy of existing state of the art manufacturing techniques or combiners and other circuits and systems where parts are machined using computer numerically controlled tools.

FIG. 8 represents the similar metrics for waveguide block 300, such as insertion loss, but with a sensitivity analysis performed on the structure showcasing shifts in the position of the slices 301 as a function of relative position between each slice. Each slice 301 was moved about 10 and 20 μm top and bottom, left and right, and the maximum deviation from the original performance is shown in FIG. 8. The lower two curves (return loss and isolation) show little variation from the data shown in FIGS. 7, since the part is well impedance matched (return loss variations) and the isolation between ports is unaffected. The data also show that up to 20 μm of movement between the slices 301, or a similar manufacturing design change variation in assembly, does not change the insertion loss 39, 40 substantially at a frequency range of 225-240 GHz. The design sensitivity is both a function of the manufacturing tolerance and assembly of the waveguide block 300. For this exemplary design, a manufacturing tolerance of 2 μm or better throughout the slice fabrication of slices 301 can be achieved using the PolyStrata® process. The assembly tolerance is also ensured through use of alignment features embedded in each fabricated slice, such as the dowel holes 310 or positive and negative 3D form fitting features. Together the manufacturing and assembly tolerances can ensure that the design performance is maintained within the sensitivity analysis boundaries.

In another of its aspects, the present invention may provide transitions 900, 1000, 1100 for connection between the rectangular waveguide block 300 and passive or active electronic/waveguide components, such as, power amplifiers, transistor circuits, or integrated circuit chips 990, for example, FIGS. 9A-11C. Considering the exemplary transition 900, FIG. 9A schematically illustrates the transition 900 oriented faced-up, where as FIG. 9B shows the transition 900 faced-down. An IC pedestal 942 may be provided on the transition 900 for mounting an integrated circuit chip 990 thereto, as seen in partial cross-section in FIG. 9C. The transition 900 may also include coplanar waveguide probes, such as two coplanar waveguide probes 950, disposed in waveguide cavities 952 of the transition 900. The waveguide probes 950 may be located adjacent the IC pedestal 942, so that with the integrated circuit chip 990 in place the waveguide probes 950 may be electrically connected to the integrated circuit chip 990, such as by a wirebond 951 to an IC bonding pad 955, FIG. 9C. The waveguide cavities 952 may be configured to communicate with the outputs and inputs 141, 146 of the waveguide block 300. Thus, power traveling through the waveguides of the outputs 141 and inputs 146 of the waveguide block 300 may be directed into the cavities 952 of the transition 900, wherein the power is converted from a waveguide mode into an electrical mode in a conductor by operation of a ridge waveguide structure 953, FIG. 9C. The ridges (also called fins) of the ridge waveguide structure 953 may configured to impedance match the waveguide to the coplanar transmission line 954 for connection to the integrated circuit chip 990. The transition 900 may be configured to provide a shortest distance between the rectangular waveguide block 300 and the integrated circuit chip 990, FIG. 9C, and hence reduce the insertion loss and minimize any variations in manufacturing build.

Connection of the transition 900 to the waveguide block 300 may be effected by a housing 1300 and mounting plates 1310, FIG. 13. The housing 1300 may retain the waveguide block 300 and provide surfaces to which the mounting plate 1310 may be attached. Each mounting plate 1310 may include eight locations at which a transition 900 may be attached. The transitions 900 may be aligned to the mounting plates 1310 via dowel holes 943 in the transitions 900 that are registered to dowel holes in the mounting plate 1310, and may be secured to the mounting plates 1310 via screws 1305 inserted through screw holes 946, FIGS. 9A, 13. Use of the mounting plates 1310 assists in ensuring that the waveguide cavities 952 communicate with the outputs 141 of the 16-way splitter modules 140 and inputs 146 of the 16-way combiner modules 145 of the waveguide block 300.

FIGS. 10A-10B schematically illustrate an alternative configuration of a transition 1000 in accordance with the present invention, having a horizontal coplanar waveguide 1053 instead of the vertical transition in FIG. 9. Like the transition 900, the transition 1000 may include dowel holes 1043, screw holes 1046, waveguide cavities 1052, and an IC pedestal 1042. In addition, FIGS. 11A-11B schematically illustrate another alternative configuration of a transition 1100 in accordance with the present invention which may also include dowel holes 1143, screw holes 1146, waveguide cavities 1152, and an IC pedestal 1142. Additionally, notches 1153 are provided in which a quartz electric field probe 1150 may be positioned to enable high degree of alignment between the probe, the integrated circuit, and waveguide cavities 1152. The probe 1150 may be attached directly to the integrated circuit bonding pad 1142. Guides 1151 on the transition part 1100 may be micro-machined for alignment of the quartz probe 1150 to an integrated circuit on the bonding pad 1142.

Each of the transitions 1000, 1100 may be mounted on the mounting plates 1310 in a similar manner to the transition 900. The transitions 900, 1000, 1100 can not only be used for a low loss interface between a rectangular waveguide, e.g., inputs and outputs 141, 146, and integrated circuit chip 990, but may also serve as a performance enhancer when the transition 900, 1000, 1100 and the integrated circuit 900 are connected together in a system. In one such exemplary configuration of the transition, e.g., transition 900, the impedance matching can be optimized to be inductive at the design frequency inclusive of the wirebond 951 for connection to the integrated circuit chip 990, and the respective IC bonding pads 955. The wirebond 951 and IC bonding pads 955 could be capacitive, so the wirebond inductance and the chip bonding pad capacitance may resonate together and create a low loss signal path through the chip 990. A benefit of larger capacitance, which larger bonding pads 955 of a chip 990 create, is the ease of attachment using wirebonds 951 leading to higher reliability and yield of the overall manufacturing process.

In yet another aspect of the present invention, FIGS. 12A-12C schematically illustrate a translation 1200 which allows waveguide cavities 1052 to be translated to another location to accommodate any size integrated circuit. The translation 1200 can be placed under a transition piece, such as transition 900, FIG. 12C, and can also be used for further impedance tuning if necessary in addition to that in the transition 900. Like the transition 900, the translation 1200 may include dowel holes 1243, screw holes 1246, waveguide cavities 1252, 1253. The cavities 1252, 1253 show a staggered waveguide taper where every layer in the slice is used to slightly move the waveguide cavity 1252, 1253 in one direction resembling a staircase. This allows for high performance transition of the cavity 1252, 1253 from one location to another and allow the accommodation of various size ICs. In addition, the translation 1200 may included visual alignment features 1210. The transitions 900, 1000, 1100 and translation 1200 may be made by the process illustrated in FIG. 5.

These and other advantages of the present invention will be apparent to those skilled in the art from the foregoing specification. Accordingly, it will be recognized by those skilled in the art that changes or modifications may be made to the above-described embodiments without departing from the broad inventive concepts of the invention. It should therefore be understood that this invention is not limited to the particular embodiments described herein, but is intended to include all changes and modifications that are within the scope and spirit of the invention as set forth in the claims.

SYSTEMS AND METHODS FOR MANUFACTURING STACKED CIRCUITS AND TRANSMISSION LINES

Information

Publication Number

Date Filed

Date Published

Inventors

Original Assignees

CPC

International Classifications

Abstract

Description

Claims

GOVERNMENT LICENSE RIGHTS

PCT Information