METHOD AND STRUCTURE FOR MAGNETICALLY-DIRECTED, SELF- ASSEMBLY OF THREE-DIMENSIONAL STRUCTURES

CROSS-REFERENCE TO RELATED APPLICATIONS

[0001] The present application claims the benefit of U.S. Provisional Patent Application Serial No. 60/740,139, filed November 28, 2005, which is hereby incorporated by reference in its entirety, including all tables, drawings, and figures.

BACKGROUND OF THE INVENTION

[0002] This invention is directed to the self-assembly of three-dimensional structures, and more particularly, to the self-assembly of three-dimensional structures for integrated circuits utilizing shaped magnetic fields.

[0003] Over the past few decades, advances in the field of electronic and photonic devices for signal processing wireless communication, computing and the like have become more complex as they become more highly integrated. Most systems are an amalgam of heavily integrated microsystems. These systems are enabled by careful integration of subsystems of various physical size and function into compact modular devices. By way of example, the cellular phone tightly integrates CMOS electronics, RF components, photonics, micro-electro-mechanical systems (MEMS), passive discrete components, power supplies and circuit boards to enable wireless communication, data processing/storage and signal processing functionalities. The trend is for smaller, more compact, more integrated and more complex devices.

[0004] It is currently known in the art to create these devices by separately fabricating each of the components, through mutually incompatible processes, and then assembling into the sub-assemblies. For example, the semiconductor devices are batch fabricated using wafers having hundreds or thousands of individual components.

[0005] Reference is made to Fig. IA in which a schematic of a known assembly process is shown. Semiconductor devices are batch fabricated using wafers 10. After microfabrication, the wafers are diced and possibly packaged into subcomponents 12. In parallel with the semiconductor processing, passive electrical components such as capacitors, inductors, and resistors, as well as circuit boards 14 are fabricated. After individual

fabrication, all of the system subcomponents are assembled serially onto the circuit boards utilizing human manipulation, or more commonly robotic "pick and place" systems to form the completed electrical circuit board 15. This method has been satisfactory; however, it suffers from the disadvantage that the large number of components that can be batch manufactured overwhelm the throughput capabilities for the serial back-end packaging and assembly. Because of the serial nature, the throughput is limited by the number and speed of the robotic manipulators. Secondly, the shrinking physical size of the microelectronics, down to the micro- or nanoscale in many applications, requires precise manipulation. As components become smaller, the absolute alignment positioning tolerances scale equivalently; sometimes beyond the capabilities of robotic manipulation.

[0006] Lastly, for sub-millimeter parts, the adhesion forces between the part and the manipulator are significant compared to gravity, resulting in a sticking problem.

[0007] Self-assembly has been utilized to a small extent during the packaging and preparation process when forming components onto substrates. Self-assembly is the autonomous organization of components into patterns or structures without human intervention. It is the assembly of parts onto a fixed substrate or the assembly of homogeneous or heterogeneous mixtures of parts one to another. The process is inherently stochastic, relying on the random distribution, mixing and physical interactions between the parts.

[0008] The self-assembly processes are governed by two fundamental forces, one being played against the other. Namely, the two forces are the mixing forces causing the large-scale mechanical movement between the parts to be assembled and short-range bonding forces causing the parts to assemble to one another when in close proximity. To maintain a bond, the short range bonding force must be greater than the mixing force to provide a stable connection. The short range bonding force must also be sufficient to overcome other external forces that may act to separate the parts, such as gravity, surface tension, buoyancy, electrostatic forces or the like. Alignment of the parts in self-assembly is typically dependent on minimization of the total free energy.

[0009] It is also known in the art to effect mixing by fluid flow (typically for wet assembly) or by vibration energy (typically for dry assembly). Short-range bonding forces

have taken the forms of gravity, electrostatic forces, magnetic forces, and capillary forces (as seen in Fig. 2). It should be noted that the forces need merely be sufficient to temporarily hold two pieces of interest together even in the presence of the mixing forces. Once the parts have been assembled, permanent, mechanical and electrical connections can be made through the curing of polymer-based adhesives or reflowing of solder bumps.

[0010] More particularly, it is known to use electrostatic forces to more efficiently direct and hold one part to another. In one prior art embodiment, pattern electrodes have been formed on the surface of at least one of the parts to create electric field traps to capture and surface mount components and LEDs on silicon substrates. This method has been satisfactory; however, it requires the formation of an electrical circuit in a desired pattern on the substrate. This adds complexity to the structure of a substrate, is often substrate-limited, i.e., is too difficult to use between two free-floating bodies, and requires the input of energy.

[0011] One way of overcoming this shortcoming is to create direct forces between the component and the substrate such as magnetic forces. It is known in the art to use magnetic forces to self-assemble 50 μm nickel disks coated with immobilized biomaterials into an array of nickel disk pattern on the substrate. However, in the prior art, the magnetic forces were the result of the entire structure being a magnet where the structures were attracted to each other, but there is no control over selectivity, or interaction. Furthermore, the magnetic approach was also limited to substrate bonding, not to free floating bodies.

[0012] To overcome the shortcomings of these self-assembly structures and methods, capillary forces have also been used to drive self-assembly. At small scales, the capillary forces become dominant. The hydrophilicity of various regions of a surface is controlled to pattern liquid films on a substrate. When a hydrophilic contact pad on one side of a part comes in contact with a liquid droplet, the pad spontaneously wets, and capillary forces draw the part into alignment, thereby minimizing the fluidic interfacial surface energy. This technique has come into vogue to assemble small parts onto planar surfaces with submicron precision for micromirror arrays, inductors and micropumps by way of example.

[0013] These techniques and methods for manufacture have been satisfactory; however, they do not match the full functionality offered by robotic or human part manipulation such as orientational uniqueness, bonding selectivity, or inter-part bonding. As

a result, they limit the basic logical design rules available for a designer in the self-assembly process.

[0014] To provide orientational uniqueness, the method must restrict part bonding to a unique physical orientation between the two bodies to be bonded. As a result, process yield is improved by minimizing the number of misaligned, misfit or incorrectly bonded (e.g., upside down) bonds. Orientational uniqueness is necessary to ensure physically symmetric parts are bonded in the desired orientation to allow complex mating of interconnecting structures.

[0015] In the prior art, bonding forces would be sufficient to dominate the mixing forces even if incorrectly aligned, hi other words, certain incorrect orientations may result in a local minimization of energy, and the mixing energy is insufficient to move the part into the desired orientation to achieve the global energy minimum. Furthermore, for bonding approaches that have geometric symmetry, energy minimization may occur even in more than one physical orientation.

[0016] The gravity driven self-assembly process has attempted to overcome this issue. This is done by giving a specific shape to the receptor site hole. However, these approaches require parts with large scale asymmetrical physical geometries, adding cost and complexity to the batch manufacture process, hi the capillary driven self-assembly method, asymmetric bonding interfaces have been implemented, but the alignment precision decreased and the process yields dropped by nearly 70%. The drop in yield is attributed to local energy minima creating misfits and the reduced precision was attributed to a less sharp dip in the energy curve.

[0017] The self-assembly method must also lend itself to bonding selectivity, i.e., that the desired part bond with its intended mating part, but not with a third unintended part. Each of the gravity, electrostatic, electric field trap, and capillary methods can be adapted to provide bonding selectivity based on geometric shape, electromagnetic properties, surface and hydrophilicity between the parts.

[0018] By way of example, for capillary driven assembly methods, a method has been developed for activating or deactivating certain receptor sites for the self-assembly of

different components using sequential steps. Although allowing bonding selectivity, it requires substantial processing and precludes parallel assembly of heterogeneous mixtures. Furthermore, this sequential process lengthens the assembly process as a function of the number of different components.

[0019] For a gravity driven self-assembly, shape-matching techniques have been used to effect bonding selectivity for parallel self-assembly of a heterogeneous mixture of three different parts. It is comparable to the approach of a square block, which will not fit, into the proverbial circular hole. However, the number of mutually exclusive shapes may be limited and chip real estate may be wasted as a result of the need for size differentiation as one shape differentiator. Furthermore, the machining of arbitrarily shaped parts imposes additional processing complexity and cost. Therefore, the prior art provides no real solution for the bonding selectivity issue and the requirement in more complex applications for sequential selective bonding in a parallel process.

[0020] A self-assembly process should also enable inter-part bonding, namely that free floating parts bond to other free floating parts, rather than to fixed substrates. With respect to inter-part bonding, the electric field method becomes inapplicable with its requirement of the application of an electric charge to at least one of the bodies.

[0021] Inter-part bonding further requires the ability to bond in any arbitrary direction. Because of this, gravity-driven processes are inapplicable because gravity only acts in one direction. A collection of free floating parts would have no driving force to bond to one another unless oriented so that gravity is in the direction of at least one of the parts. As a result, short-range bonding forces must exist intrinsically between the parts, not from some externally applied source.

[0022] A secondary challenge for inter-part bonding is preventing agglomeration, where parts of a similar type inadvertently bond to each other rather than to the specified receptor site. As a result, capillary-driven assembly is inapplicable because agglomeration requires that there are no short range bonding forces between similar parts, while insuring there are bonding forces between dissimilar parts. If parts of a first type must have a wetted receptor, they will agglomerate to each other while the "dry" second type would not agglomerate.

[0023] For a capillary-driven self-assembly approach, it has been demonstrated that a two-step sequential self-assembly of encapsulated LED structures using shape- and solder-directed processes of free floating parts may overcome the inter-part bonding issue. The agglomeration problem is overcome by recessing the wetted receptor sites and cavities and limiting access to only the smaller parts that are intended to be bonded. Solder may also be used as the bonding liquid in a way that solid solder bumps are patterned at the wafer level and heated to melt to form wetted contacts on each of the individual components. When cooled, the solder also serves as the permanent electrical and mechanical contact. This process has been shown to have potential to be satisfactory, however, it suffers from the drawback that it requires sequential assembly steps and the limitation of a single electrical contact between the individual parts and the requirement for parts of dissimilar size to prevent agglomeration.

[0024J Accordingly, a method and structure for overcoming the shortcomings of the prior art and allowing self-assembly between two bodies while providing orientational uniqueness, bonding selectivity and inter-part bonding is desired.

BRIEF SUMMARY OF THE INVENTION

[0025J hi order to effect self-assembly, a first body is provided with a single magnet or plurality of magnets disposed on the body to form a spatially varying magnetic field in a first predetermined pattern. In an embodiment, the magnets are of a first polarity. A second body has a single magnet or plurality of magnets disposed on the body to form a spatially varying magnetic field in a second pattern, which is complimentary to the first pattern. In an embodiment, the second pattern is the mirror image of the first pattern.

[0026] The magnet or magnets of the second body may either be magnetically hard, having a magnetic polarity opposite to the polarity of the first magnet or magnets, or may be magnetically soft in which it becomes magnetized in the presence of the magnetic field from the first pattern. A magnetically hard magnet, or permanent magnet, has a permanent magnetic polarization. A magnetically soft magnet, or soft magnet, becomes magnetized, and exhibits a magnetic field, in the presence of another magnetic field. The size and arrangement of the first pattern and second pattern being sufficient so that when the first pattern is in close proximity to the second pattern, and substantially aligned, the attraction

force between the first magnetic field pattern and the second magnetic field pattern is sufficient to overcome a mixing force. The magnetic attraction force resulting from the interaction of the first magnetic field and the second magnetic field aligns the two bodies and bonds the two bodies. Full alignment of the two bodies with respect to each other results in an energy minimum, resulting in a stable physical orientation.

[0027] In one embodiment of the invention, a plurality of magnets is formed in at least one of the first body or second body. In another embodiment, a single magnet with a specific physical shape is formed in at least one of the first body or second body. In a third embodiment, one or more magnetic regions on the first body and/or second body are spatially magnetized to form the first or second shaped magnetic field pattern. In an embodiment, the orientation of the first magnetic field pattern relative to the second magnetic field pattern causes the first body to be uniquely oriented and bonded to the second body.

[0028] During use, a first body is formed with a single magnet or plurality of magnets disposed to produce a first magnetic field having a first predetermined pattern. A second body is formed with a single magnet or plurality of magnets to produce a second pattern complimentary to the first pattern. The first body and second body are mixed together. The first body being attracted to the second body when in sufficient proximity for an attraction between the first shaped magnetic field pattern to the second shaped magnetic field pattern to overcome a mixing force. The attraction between the first shaped magnetic field pattern and second shaped magnetic field pattern overcomes the mixing force when the first pattern substantially aligns with the second pattern. The full alignment of the first body and the second body results in an energy minimization and a stable bonding arrangement.

BRIEF DESCRIPTION OF THE DRAWINGS

[0029] For a fuller understanding of the invention, reference is had to the following description taken in connection with the accompanying drawings in which:

[0030] Fig. IA is a schematic diagram of a method for serial robotic self-assembly of sub-components in accordance with the prior art;

[0031] Fig. IB is a schematic diagram of a method for parallel self-assembly in accordance with the prior art;

[0032] Figs. 2A-2E are schematic diagrams of prior art self-assembly methods for components on a substrate;

[0033] Fig. 3 is an exploded perspective view of a self-assembled circuit constructed in accordance with the invention;

[0034] Fig. 4 is a graphical representation of the magnetic forces and potential energy between two uniformly magnetized permanent magnets for demonstrating an aspect of the invention;

[0035] Fig. 5 is a schematic and graphical representation of the magnetic forces at work in accordance with the invention;

[0036] Figs. 6A-6E show a schematic representation of the fabrication of wafers in accordance with the invention.

DETAILED DESCRIPTION OF THE INVENTION

[0037] Reference is now made a specific embodiment, shown in Fig. 3, in which an assembly, generally indicated as 30, includes a first member 32, a second member 34 and a third member 36. First member 32 is a printed circuit board; the circuit having a generic flip chip circuit and capacitor by way of example mounted thereon. Circuit board 32 is provided with circuit structure including wiring or trace leads 38 conductively coupled to solder bumps 40 to allow electrical connection with other members 34, 36. Circuit board 32 (interchangeably referred to as first member 32) includes a first plurality of magnets 42 arranged in a pattern 44 to provide a shaped magnetic field. Circuit board 32 also includes a second plurality of magnets 46 arranged in a second pattern 48.

[0038] Second member 34, in this embodiment a flip chip, is formed with a plurality of magnets 50 thereon and solder bumps 52. Magnets 50 are provided in a predetermined pattern 54, which is the mirror image of pattern 44 to provide a mirror shaped magnetic field.

[0039] Similarly, member 36, a capacitor in this example, includes solder bumps 56 and a plurality of magnets 58 in a pattern 60 which is the mirror image of pattern 48.

[0040] Magnetic forces arise between the patterns of magnets as a result of the interaction of the magnetic fields between any two magnets. The magnetic fields are related to the material. While all materials have magnetic properties, what are commonly referred to as magnets are typically ferromagnetic in nature and can be classified as either hard or soft magnets. Hard magnetic materials possess a strong permanent magnetization, while soft magnetic materials incur a strong magnetization only upon the application of an external magnetic field. The use of different magnetic materials possessing different magnetic properties and strengths provides flexibility in the design.

[0041] By utilizing complimentary magnetic field patterns, a magnetic "lock and key" matching mechanism is provided. The size, shape, material type, and number of the individual magnets can be used to control the total magnetic patterns of the magnets, resulting in complex magnetic force field interactions, permitting a degree of pattern- matching between the components.

[0042] If sufficiently asymmetric and diverse patterns are generated, it enables orientational uniqueness, bonding selectivity and inter-part bonding. These capabilities are achieved simultaneously in one massively parallel self-assembly step. In an embodiment, agglomeration is prevented because the individual magnets for any given component type are magnetized in the same direction {e.g., all having north or south polarities for the hard magnets). Jn additional embodiments, more complicated magnetic field patterns are utilized with agglomeration prevented as the magnetic attraction between parts of the same type are lower than the mixing force used. Thus, parts of the same type will repel each other, increasing the overall speed of the matching process while preventing misalignment of two bodies.

[0043] Reference is now made to Figs. 4 and 5 for illustrating how the inherent properties of the individual magnets in the array will aid the self-sorting process. As shown in Fig. 4, two uniformly magnetized magnets are separated axially by a gap g and a lateral displacement, d . Under quasi-static conditions, the magnetic force, F , between the two magnets can be computed using several approaches: Maxwell's stress tensor, energetic principles (method of virtual work), or equivalent source models. For all but the simplest

geometries, finite-element methods axe typical employed and for finite-element modeling, the method of virtual work has been shown to be the most precise.

[0044] In this approach, the magnetic force is calculated by the derivative of the magnetic energy, U , with respect to a virtual displacement, ds:

The energy is related to the magnetic field distribution B by,

where V is the volume, μ ₀ is the magnetic permeability of free space. From Eq. (1), it is clear the force acts in a direction so as to reduce the energy stored in the magnetic field. As shown in Fig. 4, the force acts to close the magnetic gap and can be decomposed into an axial pull-in force, F _g , and a radial alignment force, F _d , both functions of the gap, g , and displacement, d .

[0045] Magnetic pattern-matching is possible by combining the interactions of multiple magnets, as depicted in Fig. 5. Here, a plot of potential energy of the system versus lateral displacement, d , exhibits several local minima, as individual magnets align, but only one global minimum, where all the magnets align. Physically, as the arrays shift with respect to each other, overlap of individual magnets may cause small bonding forces. However, the peak bonding force for the system is not reached until all magnets are acting in parallel. This same effect would apply for rotational misalignment, denoted θ , and therefore could be used to realize orientational uniqueness.

[0046] This concept can also be extended to achieve bonding selectivity by considering the potential energy between different magnetic patterns. As dissimilar patterns overlap, individual magnets may cause small bonding forces, the parts will continue to mix until a part with the correct pattern fills the vacancy, thereby minimizing the total free energy. Again, by setting the mixing force higher than any attraction other than that provided by full alignment, the correct orientation will be self-selecting. In an embodiment, if magnets within the pattern are of different polarity, the patterns could cause repulsion between two bodies unless orientation between the patterns was correct so that all magnets attracted all opposed magnets.

[0047] Referring back to Fig. 3, during use, a first body such as circuit board 32, is provided with at least one predetermined pattern 44 of a plurality of magnets 42. A desired matching body, such as flip chip 34, is provided with a pattern 54 of a plurality of magnets 50. Magnets 50 are the opposite in polarity of magnets 42, although all of magnets 50 need not be the same polarity. Magnets 50 and/or magnets 42 are hard magnets. However, either one of magnets 50 or 42 may be hard magnets, while the other is a soft magnet. Member 32 and member 34 are mixed with each other. Because of the complementary magnetic patterns, when the magnetic attraction between member 34 and member 32 is sufficient to overcome the mixing forces, member 34, the flip chip, will be affixed to member 32 on the circuit board in the desired orientation so that solder bumps 52 are aligned with solder bumps 40. Because of the need for the magnetic attraction to overcome the mixing force, all non-aligned patterns will be dislodged or repulsed. For this reason, pattern 48 would not sufficiently attract pattern 54 because the overlay of magnets 50 with magnets 46 would not provide sufficient one-to-one correlation to overcome the mixing forces.

[0048] Similarly, second member 36 thrown into the same mixing process would, in parallel, become attracted to circuit board 32 because of the complementary magnetic patterns mirror image magnet patterns 48, 60. In this way, capacitor 36 and flip chip 34 are self-selecting for their appropriate position relative to circuit board 32. The same would be true if circuit board 32 were in fact broken into two pieces so that one half of circuit board 32 would be self-selecting with flip chip 34 while the other half would be self-selecting with capacitor 36 to prevent misfit or disorientation.

[0049] The above examples were described in connection with patterns of a plurality of magnets. However, a patterned magnetic field is needed. Therefore, it is within the scope of the invention to provide a patterned magnetic field from a single magnet having a unique physical shape attracting a second single magnet of mirrored unique physical shape. A third manner to obtain the patterned magnetic field is to provide a single physical magnet with varying magnetic properties (magnetization pattern) within itself in the desired pattern. In various embodiments, various combinations of permanent magnets, soft magnets, magnet arrays, magnet shapes, magnet magnetization profiles, and magnet patterns can be used to achieve a desired magnetic field pattern.

[0050] It is preferable for the magnetic materials to be integrated into the component structures. The magnetic structures are relatively small in volume as compared to the bodies in which they reside and therefore large amounts are not required reducing the relative cost. The magnetic processing, as will be described below, can be performed in batch operations. For wafer-level processing, micromagnetic components can be integrated at pattern by electrodeposition as seen in Figs. 6A through 6E.

[0051] In the electrodeposition approach, the micro-magnets are patterned generally through photolithography. A substrate 100 is provided at step 6 A with a deposited seed layer 102. This conductive seed layer 102 is a thin layer of 0.1 μm. In step 6B, small geometric shapes, having a height in the micrometer to millimeter range, are patterned into a photoresist layer 104, serving as the electroplating mold. In a step 6C, magnetic materials 106 are electroplated into the mold formed by photoresist layer 104. In a step 6D, the mold and seed layer 102 are removed by wet etching as known in the art. In step 6E, the wafer is diced as known in the art.

[0052] This process is simple, low cost and CMOS compatible and is efficient for integrating a variety of magnetic materials with sufficient thickness (1 μm to 100 μm +) to provide useful magnetic fields. Many different thick film electroplated permanent magnet materials have been developed for MEMS applications possessing a wide range of magnetic strengths, which would lend themselves to the process.

[0053] Other methods exist in the art for patterning small magnetic patterns onto micro-fabricated parts, such as sputtering. For non-micro-fabricated parts, screen printing particle could be used. Also, new liquid metal jet printing methods, similar to inkjet printing, may be used to deposit small molten droplets of metallic alloys in pattern arrays. Lastly, magnetic materials may be diffused in a pattern into a substrate or body.

[0054] As a result of providing a predetermined magnetic pattern and a complementary magnetic pattern thereof on two respective bodies, it is possible to self- assemble small free floating bodies on a large parallel basis. By providing specific magnet patterns, the bodies will be self-selecting and self-orienting to each other so that only the correct bodies in the correct orientations will bond in a mixing process, whether wet or dry.

[0055] In a specific embodiment, the first magnetic field on the first body can be n- fold symmetric and the second magnetic field on the second body can be n-fold symmetric such that there are n orientations that the second body can be fully aligned and bonded with the first body. In other embodiments, the complementary first and second magnetic fields are asymmetric such that only one orientation of the first and second bodies results in full alignment and bonding.

[0056] In a specific embodiment, a structure can have additional bodies that can align and bond to the second body, or any other body of the structure, due to magnetic fields on the additional bodies and the bodies to which the additional bodies align and bond. As an example, a robot structure can have a torso that aligns and bonds with a right arm, a left arm, a right leg, a left leg, and a head. Each of the parts, or bodies, can selectively bond to the torso such that each part bonds only to the appropriate location and not to other locations. In a further embodiment, a right hand, a left hand, a right foot, and a left foot, can align and bond with the right arm, the left arm, the right leg, and the left leg, respectively. The alignment and the bonding can be done in stages or simultaneously. In this way, a plurality of torso and a corresponding plurality of associated parts could be mixed such that the robots are self-assembled. In an analogous way, other structures, such as MEMS structures, could be assembled. Such structures could include a variety of parts to be assembled, such as gears, blocks, levers, and other structures that can be held together via magnets incorporated with the parts.

[0057] hi an embodiment, the first body and second body that will be in contact after alignment and bonding can incorporate other attachment and/or connectivity means, such as glues for mechanical adhesion and solder bumps for electrical connectivity or bonding. In an embodiment, solder bumps can be used such that heating after alignment and bonding can cause a portion of the first body to be soldered to a portion of the second body. In another embodiment, glue can be used such that heating after alignment and bonding can cause a portion of the first body to be glued to a portion of the second body.

[0058] In a specific embodiment, shielding material can be incorporated with the first and/or second body to shield the magnetic fields of the first and/or second body from portions of the aligned and bonded structure or other objects proximate to the structure. Such shielding materials are well known in the art. As an example, soft magnetic materials can be

used as shielding. A layer of soft magnetic material can be placed on one or more bodies and then the magnets for the body grown on the soft magnetic material layer such that the portion of the body below the layer is shielded from the magnetic fields of the magnets grown on the layer.

[0059] In a specific embodiment, demagnetizable magnets and/or magnetic materials can be used such that after alignment and bonding, and optional implementation of additional attachment means, the demagnetizable magnets and/or magnetic material can be demagnetized, hi an example, magnets and/or magnetic material that can be demagnetized by heating can be used. Such materials are well known in the art. hi a specific embodiment, magnets and/or magnetic materials that can be demagnetized by heating to demagnetization temperatures under 500 ⁰ C can be used.

[0060] Preferably, the magnetic attraction force between two bodies are at least as great as the weight force of the body free to move, and at least as great as the weight force of the lightest body if both bodies are free to move. Preferably, the magnetic attraction force is two to ten as great as the weight force of the body free to move, and at least two to ten as great as the weight force of the lightest body if both bodies are free to move.

[0061] Thus, while there have been shown, described and pointed out novel features of the present invention as applied to preferred embodiments thereof, it will be understood that various omissions and substitutions and changes in the form and detail are contemplated so that the disclosed invention may be made by those skilled in the art without departing from the spirit and scope of the invention. It is the intention therefore to be limited only as indicated by the scope of the claims appended hereto. It is also to be understood that the following claims are intended to cover all of the generic and specific features of the invention herein described and all statements of the scope of the invention, which as a matter of language might be said to fall therebetween.