Radio frequency signal low-loss double-sided metal-based circuit board

Abstract

The application discloses a kind of low-loss double-sided metal base circuit board of radio frequency signal, it is related to radio frequency circuit board technical field, two sides of continuous metal core are sequentially provided with dielectric layer and circuit layer containing radio frequency signal line, signal pad and ground conductor;Continuous metal core is provided with stepped hole, stepped hole includes main hole section and the end compensation hole section of at least one end and larger diameter located therein, and the center signal conductor insulated from continuous metal core and connected with the signal pad on both sides is arranged in the insulating filling body in stepped hole;Signal pad is connected with the radio frequency signal line on the same side through plane gradual change section;Each of two sides dielectric layer is provided with a plurality of ground blind holes around stepped hole and connecting ground conductor to continuous metal core.The main hole section and the end compensation hole section respectively adjust the impedance of core-penetrating main section and the capacitive boundary of hole mouth, ground blind hole shortens reflow path, plane gradual change section slows down transmission mode conversion, which collectively slows down the impedance mutation at core-penetrating transition and reduces additional loss.

Description

Technical Field

[0001] This invention relates to the field of radio frequency circuit board technology, and more specifically to a low-loss double-sided metal-based circuit board for radio frequency signals. Background Technology

[0002] Double-sided metal-based circuit boards typically use a continuous metal core that serves as the substrate for load carrying, heat dissipation, and grounding, with dielectric and circuit layers on both sides of the continuous metal core. To achieve electrical connection between the lines on both sides, isolation holes larger than signal holes are usually drilled in the continuous metal core. Insulating material is then filled into the isolation holes, and after the insulating material has cured, the signal holes are drilled and the hole walls are metallized.

[0003] The above structure can achieve DC insulation between the center signal conductor and the continuous metal core. However, when used for cross-plane transmission of radio frequency signals, the ordinary equal-diameter isolation hole, the circular signal pad, and the long-distance grounding point will jointly form an electromagnetic discontinuity: the center signal conductor generates a series parasitic inductance, the signal pad and the hole of the continuous metal core generate a parallel parasitic capacitance, and the surface return current may also detour to the long-distance grounding point.

[0004] Enlarging the isolation vias may reduce capacitive coupling, but it will also enlarge the return loop. Reducing the size of the signal pads may compress the via ring connection and pattern alignment margin. Adding grounding vias without coordinating the dimensions of the via openings, signal pads, and planar transition sections may also create new local electric field concentrations.

[0005] Therefore, even with continuous grounding of the continuous metal core, reliable insulation of the central signal conductor, and feasible processing, existing double-sided metal-based circuit boards still struggle to simultaneously constrain the parasitic inductance of the vertical signal path, the parasitic capacitance of the orifice and pad, and the double-sided return current routing. Impedance abrupt changes and additional losses are prone to occur at the through-core transition point. Summary of the Invention

[0006] (a) Technical problems to be solved This invention aims to solve the problem that when using equal-diameter isolation vias to achieve cross-plane transmission of radio frequency signals on existing double-sided metal-based circuit boards, the series parasitic inductance of the center signal conductor, the parallel parasitic capacitance of the pads and vias, and the surface return current routing together form electromagnetic discontinuities, making it difficult to achieve stable low-reflection through-core interconnects due to the mutual constraints of single compensation methods. Under the premise that the continuous metal core maintains continuous grounding and also has the functions of bearing and heat dissipation, and the center signal conductor is reliably insulated, the parasitic inductance, parasitic capacitance, and double-sided return current routing are constrained at the same time, so as to reduce the impedance change between the planar transmission mode and the through-core transmission mode and reduce additional losses.

[0007] (II) Technical Solution To achieve the above objectives, the present invention adopts the following technical solution: A low-loss double-sided metal-based circuit board for radio frequency signals includes a continuous metal core. A dielectric layer and a circuit layer are sequentially disposed on each side of the continuous metal core. Each circuit layer includes a radio frequency signal line, signal pads, and a ground conductor. The continuous metal core has a through-hole filled with an insulating filler. A central signal conductor, insulated from the continuous metal core, is disposed within the insulating filler. The two ends of the central signal conductor are connected to the signal pads on both sides. The through-hole is a stepped hole, including a main hole segment and an end compensation hole segment located at at least one end of the main hole segment. The end compensation hole segment extends inward from the surface of the continuous metal core to the main hole segment and has a diameter larger than that of the main hole segment. Each signal pad is connected to the RF signal line on the same side via a planar gradient section; Multiple grounding blind holes are provided in the dielectric layer on each side, surrounding the stepped hole. Each grounding blind hole connects the grounding conductor on that side to the continuous metal core.

[0008] Furthermore, there are two end compensation holes, located at both ends of the main hole section and coaxial with the main hole section, and the main hole section is located in the middle of the thickness of the continuous metal core; Multiple grounding blind holes on each side are located radially outside the end compensation hole section on that side, and terminate in the continuous metal area outside the end compensation hole section on that side where the continuous metal core is located. The diameter of the circle at the center of the grounding blind hole is larger than the diameter of the end compensation hole section.

[0009] Furthermore, the grounding conductors on each side form a C-shaped grounding ring around the central signal conductor. The C-shaped grounding ring has an opening at the location where the RF signal line enters the signal pad on that side, and multiple grounding blind vias on that side are arranged along the C-shaped grounding ring. The planar gradient section on this side connects to the signal pad through an opening. The line width of the planar gradient section, the gap between the planar gradient section and the adjacent ground conductor, and the local clearance profile of the ground conductor gradually change from the RF signal line to the signal pad.

[0010] Furthermore, the center signal conductor is a metallized via conductor coaxial with the stepped via, and the copper layer of the metallized via conductor is connected to the signal pads on both sides. The insulating filler fills the main hole segment and the end compensation hole segment and fits against the hole walls of the main hole segment and the end compensation hole segment. The insulating filler forms a circumferentially continuous insulating wall between the central signal conductor and the continuous metal core along the entire crossing length of the central signal conductor. The continuous metal core remains electrically conductive around the stepped aperture and is not divided into insulated islands.

[0011] Furthermore, the insulating filler includes a first dielectric region and a second adhesive dielectric region arranged coaxially. The first dielectric region surrounds the central signal conductor, and the second adhesive dielectric region is located between the first dielectric region and the hole walls of the main hole segment and the end compensation hole segment. The second adhesive dielectric region is a circumferentially continuous annular layer. The interface between the first dielectric region and the second adhesive dielectric region is continuous in the circumferential direction, and neither the first dielectric region nor the second adhesive dielectric region forms a through gap between the central signal conductor and the continuous metal core.

[0012] (III) Beneficial Effects The low-loss double-sided metal-based circuit board for radio frequency signals provided by this invention has the following beneficial effects: By setting a stepped hole consisting of a main hole segment and a larger diameter end compensation hole segment in the continuous metal core, the electric field spacing of the through-core main area and the capacitive boundary of the signal pad adjacent to the end face of the continuous metal core are adjusted by two hole diameters respectively, thus overcoming the inherent contradiction of enlarging the equal diameter isolation hole, which is equivalent to expanding the return loop.

[0013] Multiple grounding blind holes surrounding the stepped holes are set in the dielectric layers on both sides to form a double-ended grounding blind hole ring, so that the return current enters the continuous metal core through a parallel short path without detouring to the far-end grounding point, reducing the return inductance and limiting the leakage field. The continuous metal core also serves as a common RF reference body, carrier and heat diffusion substrate.

[0014] The signal pads are connected to the RF signal lines via a planar gradient section, so that the planar gradient section, the stepped via, and the double-ended grounded blind via loop act on the same local electromagnetic boundary: the end compensation via adjusts the capacitive boundary of the via opening, the outer grounded blind via converges the return loop, and the planar gradient section gradually converges the field distribution of the planar transmission mode to the via opening, which is beneficial for jointly constraining parasitic inductance, parasitic capacitance and return loop, and reducing reflection and additional loss at the through-core transition.

[0015] The insulating filler forms a circumferentially continuous insulating wall around the center signal conductor and can be radially divided into a first dielectric region and a second adhesive dielectric region. This allows the electric field transmission, hole wall wetting, adhesion, and stress buffering to be handled by different regions, which is beneficial for balancing interface reliability and reducing additional losses caused by dielectric loss. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the overall structure of the double-sided metal-based circuit board and the radio frequency signal return path of the present invention; Figure 2 This is a schematic diagram showing the axial cross-section and dimensions of the stepped through-core structure of the present invention; Figure 3 This is a schematic diagram showing the layout and partial enlargement of the upper and lower surfaces of the through-core structure of the present invention; Figure 4 This is a flowchart illustrating the design, fabrication, and testing process of the stepped through-core structure of this invention. Figure 5 This is a schematic diagram of the structure of a parallel embodiment of the present invention; Figure 6 This is a flowchart of the anomaly handling process in the manufacturing and testing of this invention; Figure 7 This is a schematic diagram illustrating the recording relationship of design, manufacturing, and testing data for this invention; Figure 8 This is a schematic diagram of the test ticket and measurement topology of the present invention. Detailed Implementation

[0017] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. The described embodiments are only some embodiments of the present invention, and equivalent substitutions or modifications made by those skilled in the art without departing from the technical concept of the present invention should fall within the protection scope of the present invention.

[0018] Please see Figures 1 to 8 This invention provides a low-loss double-sided metal-based circuit board for radio frequency signals.

[0019] In double-sided metal-based circuit boards, to achieve electrical connection between traces on both sides of a continuous metal core, isolation vias larger than signal vias are typically created on the continuous metal core. These vias are filled with insulating resin, and after the resin cures, the signal vias are drilled and the via walls are metallized. This method solves the DC insulation problem between the center signal conductor and the continuous metal core. However, when used for cross-plane transmission of radio frequency signals, ordinary equal-diameter isolation vias, circular signal pads, and distant grounding points can collectively create significant electromagnetic discontinuities: the center signal conductor generates series parasitic inductance, the signal pads and via openings of the continuous metal core generate parallel parasitic capacitance, and surface return current may even bypass to the distant grounding point. Enlarging only the isolation vias, while reducing capacitive coupling to some extent, will enlarge the return loop; shrinking only the signal pads may compress manufacturing margins; and adding only grounding vias without coordinating the dimensions of the via openings, signal pads, and planar transition sections may also create new local electric field concentrations.

[0020] Therefore, simply combining known methods such as low-loss materials, signal hole insulation, grounding blind hole conduction, and signal pad compensation cannot automatically achieve wideband and repeatable low-reflection through-core interconnects.

[0021] The technical problem addressed by this embodiment is: under the conditions that the metal core is continuously grounded and also has the functions of bearing and heat dissipation, and the central signal conductor is reliably insulated and can be fabricated, how to simultaneously constrain the parasitic inductance of the vertical signal path, the parasitic capacitance of the orifice and pad, and the double-sided return current bypass, so as to mitigate the impedance change between the planar transmission mode and the through-core transmission mode, and reduce the additional losses introduced by dielectric loss, conductor roughness and return current diffusion.

[0022] Therefore, this embodiment does not treat a specific aperture or a particular low-loss material as an isolated solution. Instead, it uses a continuous metal core 1 as a common RF reference body. A stepped aperture with a main aperture segment 11 and an end compensation aperture segment 12 is provided at the core-penetrating position. A low-loss insulating filler 6 is formed within the stepped aperture, and a center signal conductor 7, insulated from the continuous metal core 1, is formed within the low-loss insulating filler 6. The upper and lower ends of the center signal conductor 7 are connected to the RF signal lines on both sides through signal pads and planar gradient segments, respectively. The grounding conductors on both sides are connected to the continuous metal core 1 nearby through grounding blind holes 8 provided around the core-penetrating position. The design phase uses the same target impedance. Z tar The process involves constraining stepped vias, center signal conductor 7, signal pads, double-ended grounding blind via rings, and planar gradient sections. During the manufacturing stage, the actual dielectric layer thickness, via copper layer thickness, and center deviation are recorded. During the testing stage, the structural measurements and RF response are used to correct the geometric parameters of subsequent batches, thus forming a closed-loop technology chain in which structure, electromagnetic design, and manufacturing testing correspond to each other.

[0023] Please refer to Figure 1 As shown, the circuit board consists of a continuous metal core 1, an upper low-loss dielectric layer 2 and an upper circuit layer 4 located above the continuous metal core 1, and a lower low-loss dielectric layer 3 and a lower circuit layer 5 located below the continuous metal core 1. The upper RF signal line 41 in the upper circuit layer 4 connects to the lower RF signal line 51 via an upper transition section 43, an upper signal pad 44, a center signal conductor 7, a lower signal pad 54, and the lower transition section 53. Correspondingly, the return current enters the continuous metal core 1 from the upper ground conductor 42 through the upper grounding via 8, and then reaches the lower ground conductor 52 via the lower grounding via 8. The continuous metal core 1 is not divided into mutually insulated ground islands around the core penetration location, allowing the signal path, insulating dielectric, and return current path to work collaboratively within the same local three-dimensional region.

[0024] like Figure 1 and Figure 2 As shown, the double-sided metal-based circuit board referred to in this document refers to a board with low-loss dielectric layers and patterned conductor layers on both sides of the continuous metal core 1, allowing radio frequency signals to reach the radio frequency signal lines on the other side through local vertical interconnects; the upper and lower sides only indicate the relative positions in the illustration and do not limit the actual mounting direction of the board. The continuous metal core 1 referred to in this document means that, except for signal through holes and necessary process holes, the continuous metal core 1 maintains continuous conductivity and is connected to the radio frequency reference ground around the through-hole area; this term does not require that the continuous metal core 1 is completely without holes, nor does it include floating metal islands that are insulated from each other around the through-hole location and cannot form a common return path.

[0025] Combination Figure 2It is understood that the stepped hole includes a main hole segment 11 penetrating the continuous metal core 1, and end compensation hole segments 12 located at at least one end, preferably at both ends, of the main hole segment 11. The effective diameter of the main hole segment 11 is... D 1. The effective diameter of the end compensation hole section 12 is D 2. The depth of the end compensation hole section 12 along the stepped hole axis, from the end face of the continuous metal core 1 to the interface between the main hole section 11 and the end compensation hole section 12, is: H c The effective diameter is based on the actual measured size after the stepped hole has been machined, deburred, and cleaned. In this main embodiment, D 2 greater than D 1. The main hole section 11 mainly defines the electric field spacing of the through-core main body area, and the end compensation hole section 12 mainly changes the capacitive boundary of the upper signal pad 44 and the lower signal pad 54 adjacent to the end face of the continuous metal core 1; the two have different adjustment functions, and the stepped hole should not be understood as a regular countersunk hole used only for deburring.

[0026] refer to Figure 2 The low-loss insulating filler 6, as referred to in this paper, is an electrically insulating medium that is filled into the stepped holes and cured. It forms a continuous insulating wall between the central signal conductor 7 and the continuous metal core 1, while simultaneously supporting subsequent signal hole drilling and hole wall metallization. The term "low loss" in this paper refers to the dielectric polarization loss relative to the target frequency band. In engineering practice, the relative permittivity at the corresponding frequency point within the target frequency band should be considered. ε r tan tangent of dielectric loss angle δ The selection should be based on a combination of factors, including curing shrinkage rate, hygroscopicity, coefficient of thermal expansion, and interfacial bonding ability, and should use traceable test values ​​or verified technical data of the batch of materials used as modeling inputs, rather than solely on the product name. The center signal conductor 7 is a metallized through-hole conductor located within the low-loss insulating filler 6, preferably coaxial with the stepped hole. Its copper layer 71 is conductive to the upper signal pad 44 and the lower signal pad 54, and is insulated from the continuous metal core 1 over its entire length.

[0027] like Figure 3 As shown, a double-ended grounding blind via ring refers to multiple grounding blind vias 8 arranged around the core axis at the upper and lower ends of the central signal conductor 7. Each grounding blind via 8 passes through the low-loss dielectric layer on the corresponding side from the upper grounding conductor 42 or the lower grounding conductor 52, and terminates at and is connected to the continuous metal core 1; multiple grounding blind vias 8 can be arranged along a C-shaped grounding ring 81. The C-shaped grounding ring 81 has an opening at the RF signal line entry point, allowing the RF signal line to approach the corresponding signal pad, while forming a shorter parallel return path in other directions.

[0028] The hole center arrangement circle referred to in this article is the circle containing the center of each grounding blind hole 8, with the axis of the stepped hole as the center. The hole center arrangement circles of the upper and lower grounding blind holes 8 are set coaxially, and the angular positions of the grounding blind holes 8 on both sides can be staggered.

[0029] Combination Figures 1 to 3 As can be seen, the planar gradient section referred to in this paper is the region where the upper gradient section 43 or the lower gradient section 53 gradually changes the signal line width, the gap between adjacent ground conductors, or the local clearance profile from the conventional planar RF signal line to the corresponding signal pad; its function is to gradually converge the field distribution of the planar transmission mode to the through-core area, rather than merely meeting the wiring shape requirements. The quasi-coaxial vertical interconnect structure referred to in this paper is an approximately coaxial field constraint region formed by the central signal conductor 7 acting as the inner conductor, the low-loss insulating filler 6 acting as the main electric field medium, and the continuous metal core 1 and the double-ended grounded blind hole ring jointly providing the peripheral reference ground; this term is used to describe the physical mechanism and is not equivalent to a standard coaxial connector with a complete cylindrical outer conductor.

[0030] refer to Figure 2 and Figure 4 The impedance coordinated control referred to in this paper is based on the target impedance. Z tar To ensure common constraints, joint determination D 1. D 2. H c Effective outer diameter of center signal conductor 7 d Equivalent diameter of signal pads D p The number of grounding blind holes 8, the radius of the circle at the center of the holes, and the length of the transition section. L t The process involves backfilling the design model with actual material parameters, dielectric layer thickness, via copper layer thickness, and center deviation data. Local equivalent impedance is the voltage and current relationship that varies with position when a radio frequency wave propagates through the planar gradient section, signal pads, center signal conductor 7, and peripheral grounding structure. It is affected by both distributed inductance and distributed capacitance and cannot be replaced by a single lumped element value.

[0031] The subsequent batch feedback correction mentioned in this article occurs between design, prototyping and batch manufacturing. It is to convert the test results into geometric parameters or process inputs for the next controlled version, and does not refer to the active electronic tuning of the radio frequency signal during the service of the circuit board.

[0032] To avoid ambiguity in parameter notation across different paragraphs, this paper adopts the following unified definition for geometric parameters: D 1 represents the effective diameter of the main hole segment 11 after machining, deburring, and cleaning. D 2 represents the effective diameter of the end compensation hole section 12 after machining, deburring, and cleaning. Hc The depth of the end compensation hole section 12 along the stepped hole axis direction, from the end face of the continuous metal core 1 to the interface between the main hole section 11 and the end compensation hole section 12; d The effective outer diameter of the center signal conductor 7 after the hole wall metallization is completed; D p The equivalent diameter of the signal pad; L t The effective length of the planar gradient segment along the centerline of the radio frequency signal propagation; e The center deviation of the axis of the center signal conductor 7 relative to the axis of the stepped hole; the diameter of the circle at the center of the hole is the diameter of the circle containing the centers of each grounding blind hole 8 with the axis of the stepped hole as the center.

[0033] This document makes the following unified provisions for electrical parameters: ε r The relative permittivity of the low-loss insulating filler 6 at the corresponding frequency point within the target frequency band; tan δ For low-loss insulating filler 6, the dielectric loss tangent is at the same frequency. Z tar The target impedance used in the design; Z v This is the estimated initial characteristic impedance of the transmission section of the through-core body; ε r,eff The effective relative permittivity is the sum of the cured low-loss insulating filler 6 and adjacent layers. Its value is calculated using a three-dimensional electromagnetic field model based on the corresponding material batch, layer thickness, and geometry, or calibrated using test tubes with the same material and layer. The relative permittivity, dielectric loss tangent, and effective relative permittivity are all dimensionless positive values ​​and must correspond to the specific material batch and test frequency. The units for the target impedance and initial characteristic impedance estimates are all ohms.

[0034] like Figure 1 As shown, this embodiment is used in the 1 GHz to 12 GHz frequency band with a target impedance. Z tarThe example is a 50 Ω copper-based metal core board. The continuous metal core 1 has a thickness of 0.8 mm; a 0.203 mm thick upper low-loss dielectric layer 2 and a 0.203 mm thick lower low-loss dielectric layer 3 are respectively formed on its top and bottom, and 18 μm thick copper foils are respectively used to form an upper circuit layer 4 and a lower circuit layer 5. These parameters are used to illustrate a specific structure and do not represent a limitation on the RF test results of the finished product. The continuous metal core 1 simultaneously serves as a common RF reference, the outer conductor of the quasi-coaxial vertical interconnect structure, and the heat-diffusing substrate, thus avoiding the need for additional independent and heavy grounding components within the board. In an alternative approach, the thickness of the low-loss dielectric layer and the copper thickness can be reselected based on the target power, linewidth, and processing capabilities; however, if the upper and lower layers are asymmetrical, the designer should create separate models for the upper and lower ends and should not use the same end compensation amount from a symmetrical structure.

[0035] Combination Figure 2 It can be seen that the continuous metal core 1 forms a coaxial stepped hole at a predetermined core-piercing position, the main hole segment 11 is located in the middle of the thickness of the continuous metal core 1, and the two end compensation hole segments 12 extend inward from both sides of the continuous metal core 1 and communicate with the main hole segment 11. In a specific embodiment, D 1 is 1.10 mm. D 2 is 1.30 mm, each end H c The thickness ranges from 0.10 mm to 0.15 mm. D 1. Define the main insulation distance from the center signal conductor 7 to the hole wall of the continuous metal core 1. D 2 and H c Together, they reduce the excessively strong capacitive coupling caused by the upper signal pad 44 and the lower signal pad 54 being close to the end face of the continuous metal core 1. If only increasing D 2. Without checking the termination position of the grounding blind hole 8, the end clearance may encroach upon the machinable area of ​​the grounding blind hole 8; if H c If the depth is too deep, the main quasi-coaxial transmission section will be shortened and may result in overcompensation.

[0036] therefore, D 2 and H c The remaining load-bearing cross section of the continuous metal core 1, the minimum insulation wall thickness, and the termination position of the grounding blind hole 8 should be determined together with the signal pad dimensions.

[0037] In embodiments where the asymmetry of the stacked structure or the unilateral electromagnetic discontinuity is more pronounced, the end compensation hole segment 12 may also be provided only at one end of the main hole segment 11. In this case, the side where the end compensation hole segment 12 is provided is... D 2 and H cAdjust the capacitive boundary of the orifice, and adjust the end impedance on the other side through the corresponding signal pads, planar gradient section, and grounding blind hole 8; the parameter solving module 102 should model the two ends separately, without using the same compensation parameter for the symmetrical structure. This implementation does not change the basic relationship of the insulation between the center signal conductor 7 and the continuous metal core 1 and the proximity conduction of the double-ended grounding blind hole 8.

[0038] like Figure 2 As shown, a low-loss insulating filler 6 is completely formed within the stepped hole. The low-loss insulating filler 6 is fitted to the hole walls of the main hole section 11 and the end compensation hole section 12, and its central region is used for subsequent drilling of the center signal hole. Exemplarily, the relative permittivity of the filling medium at the corresponding frequency point within the target frequency band... ε r The value can be selected from 3.2 to 3.6, and the dielectric loss tangent is tan tan δ Material systems with lower relative permittivity and batch traceability data should be prioritized; the relative permittivity range is used for the initial electromagnetic dimension estimation, and the actual value after curing should still correspond to the corresponding material batch and testing frequency. Lower relative permittivity ε r It helps to reduce the distributed capacitance under given geometric conditions and achieve a lower dielectric loss tangent (tan ... δ It is beneficial to reduce the dielectric loss in the filled area; however, one should not pursue lower losses at all costs. ε r and tan δ Ignoring curing shrinkage and interfacial bonding could increase the risk of voids in the pore walls or thermal cycling cracks. Therefore, material selection is constrained by both electrical performance and reliability boundaries.

[0039] refer to Figure 2 In an alternative filling method, the low-loss insulating filler 6 can form a concentric first dielectric region 61 and a second adhesive dielectric region 62. The first dielectric region 61 surrounds the central signal conductor 7, and the second adhesive dielectric region 62 is located between the first dielectric region 61 and the hole wall of the continuous metal core 1. The first dielectric region 61 preferentially undertakes low-loss electric field transmission, and the second adhesive dielectric region 62 preferentially undertakes hole wall wetting, bonding, and curing stress buffering.

[0040] The interface between the two dielectric regions should be continuous circumferentially, and no through gap should be formed between the central signal conductor 7 and the continuous metal core 1. If the detection finds that the central deviation causes the first dielectric region 61 to disappear partially, the board should not be released solely based on passing the DC insulation test. This radial partitioning can allocate low-loss performance and interface reliability requirements to different regions, while maintaining continuous insulation around the outer periphery of the central signal conductor 7.

[0041] Combination Figure 2It is understood that the central signal hole is drilled in the center of the cured and leveled low-loss insulating filler 6, and a hole copper layer 71 is formed through hole wall activation, chemical copper deposition, and electroplating, thereby constituting the central signal conductor 7. For example, the initial drilling diameter of the central signal hole can be approximately 0.25 mm, and the target thickness of the hole copper layer 71 can be approximately 20 μm to 25 μm as the starting point for process design; the effective outer diameter of the central signal conductor 7... d The measured diameter of the outer surface of the center signal conductor 7 after the hole wall is metallized should be determined based on the hole diameter and the actual copper layer thickness. The nominal diameter of the drill bit cannot be directly used as the outer diameter of the conductor in the electromagnetic model. The center signal conductor 7 receives the radio frequency energy from the upper signal pad 44, transmits it axially, and outputs it to the lower signal pad 54. The minimum insulating wall separates the center signal conductor 7 from the continuous metal core 1. If the hole is eccentric and the local insulating wall is lower than the structural limit, the manufacturing process should lock the subsequent surface treatment and isolate the board to avoid the risk of short circuit between the center signal conductor 7 and the continuous metal core 1 being covered by electroplating or solder mask.

[0042] like Figure 2 As shown, when the dielectric distribution within the main hole section 11 is approximately uniform and end discontinuities are temporarily disregarded, the initial characteristic impedance of the through-core transmission section can be estimated using the following formula:

[0043] In the formula, Z v This is the estimated initial characteristic impedance of the through-core transmission section, in ohms. ε r,eff The effective relative permittivity formed by the cured low-loss insulating filler 6 and adjacent layers is dimensionless. D 1 represents the effective diameter of the main hole segment 11, in millimeters; d The effective outer diameter of the center signal conductor 7 after the hole wall is metallized is expressed in millimeters; ln represents the natural logarithm and should satisfy the following condition. D 1 greater than d and d Greater than 0. This formula is only used to determine... D 1 and d The initial dimensions should be determined by the model, and subsequently, the end compensation hole section 12, upper signal pad 44, lower signal pad 54, C-shaped grounding ring 81, copper layer thickness, and dispersion effects should be incorporated into the three-dimensional electromagnetic field model. Alternatively, calibrated parameter lookup tables or sample regression relationships can be used to determine the initial dimensions, but the inputs and outputs used for lookup or regression should correspond to the same materials, layer thickness, and manufacturing tolerances. Furthermore, the effective relative permittivity should be greater than 0 and obtained according to the aforementioned three-dimensional electromagnetic field model calculation or test kit calibration method.

[0044] like Figure 3As shown, the upper RF signal line 41 is connected to the upper signal pad 44 via the upper transition section 43, and the lower RF signal line 51 is connected to the lower signal pad 54 via the lower transition section 53. Exemplarily, the equivalent diameter of the signal pad... D p The length of the transition section can be calculated from 0.40 mm to 0.55 mm. L t The range can be calculated from 0.5 mm to 1.5 mm. For circular signal pads, D p This refers to the outer diameter of the signal pad; for non-circular signal pads, D p Determined by the diameter of a circle that is equal to the projected area of ​​the signal pad. L t It is the length along the center line of the radio frequency signal propagation, from the position where the geometric change begins on the conventional radio frequency signal line to the connection position of the corresponding signal pad outer edge. D p An excessively large value may increase the parallel capacitance to the end face of the continuous metal core 1. D p If it is too small, it will reduce the hole ring connection and pattern alignment margin; L t Too short a length will cause a sudden change in the signal line width and the gap between the ground conductor. L t Excessive length increases board space occupancy and transmission length. By simultaneously changing the RF signal line width of the transition section, the gap between adjacent ground conductors, and the local clearance profile, the field distribution can be gradually converged from the planar transmission area to the central signal conductor 7, thereby mitigating the local impedance imbalance caused by abrupt access via only circular signal pads.

[0045] Combination Figure 3 It is understood that the upper grounding conductor 42 and the lower grounding conductor 52 respectively form C-shaped grounding rings 81 surrounding the central signal conductor 7 and having openings in the routing direction. Multiple grounding blind holes 8 are arranged on each C-shaped grounding ring 81. For example, 8 to 12 grounding blind holes 8 can be provided on each side, with the diameter of a single hole ranging from 0.10 mm to 0.20 mm, and the diameter of the circle at the center of the hole ranging from 1.8 mm to 2.2 mm. The grounding blind hole 8 receives the return current from the corresponding side grounding conductor and introduces the return current into the continuous metal core 1 nearby. The hole ring on the other side then leads the return current to the lower grounding conductor 52. Multiple holes connected in parallel are beneficial for reducing return inductance and limiting leakage field, but the number of holes and the radius of the circle at the center of the hole must not allow the grounding blind hole 8 to intrude into the end compensation hole section 12, nor should they violate the hole spacing and the laminating resin flow space.

[0046] refer to Figure 3In an alternative grounding connection method, the C-shaped grounding ring 81 can be formed by a combination of arc segments and straight segments, and the upper and lower grounding blind holes 8 can be arranged in a staggered manner with rounded corners along their respective hole centers. When the local board space is insufficient, a combination of metallized grounding grooves and some grounding blind holes 8 can also be used, as long as this combination establishes short return paths to the continuous metal core 1 at both ends of the central signal conductor 7. If a grounded coplanar waveguide is used on the upper side and a microstrip line is used on the lower side, the upper tapered section 43 can simultaneously narrow the gap between the upper RF signal line 41 and the coplanar grounding conductor, and the lower tapered section 53 can introduce a local grounding conductor near the lower signal pad 54 and then connect to the continuous metal core 1 through the grounding blind holes 8. The above replacement changes the surface electric field shaping method, but does not change the main lines of insulation between the central signal conductor 7 and the continuous metal core 1, proximity conduction of the double-ended grounding blind holes 8, and common impedance constraint.

[0047] like Figures 1 to 3 As shown, in the scenario where the radio frequency signal propagates from the top to the bottom, the incident energy passes sequentially through the upper radio frequency signal line 41, the upper transition section 43, the upper signal pad 44, the center signal conductor 7, the lower signal pad 54, the lower transition section 53, and the lower radio frequency signal line 51. The accompanying return current enters the continuous metal core 1 from the upper ground conductor 42 through the upper ground blind hole 8, bypasses the outer periphery of the low-loss insulating filler 6, and then returns to the lower ground conductor 52 through the lower ground blind hole 8.

[0048] The stepped aperture separates the main transmission section and the end compensation section for adjustment, the double-ended grounded blind hole ring shortens the return path, and the planar gradient section slows down the transmission mode conversion. All three work together on the same local electromagnetic boundary, so the RF reflection cannot be simply attributed to a single signal pad or a single aperture.

[0049] In the embodiment where the radio frequency (RF) signal propagates from bottom to top, the RF signal sequentially passes through the lower RF signal line 51, the lower transition section 53, the lower signal pad 54, the center signal conductor 7, the upper signal pad 44, the upper transition section 43, and the upper RF signal line 41. The return current flows from the lower ground conductor 52 through the lower ground blind via 8 into the continuous metal core 1, and then back to the upper ground conductor 42 through the upper ground blind via 8. Since this circuit board has a passive interconnect structure, reverse transmission does not change the electrical connection relationships of each physical structure. When the upper and lower layers are asymmetrical, the dielectric parameters, copper layer thickness, and end structures of the corresponding sides should be used for modeling.

[0050] Combination Figure 4The illustrated process allows for the design, fabrication, and testing of the stepped through-core structure of this invention. The relevant processes include steps S101 to S106. Each step can be performed collaboratively by RF designers, process engineers, and corresponding processing and testing equipment. The data processing is only used to compare design versions and measurement results. The finished circuit board itself remains a passive interconnect structure and does not require a controller or power supply in service.

[0051] like Figure 4 As shown, in step S101, the constraint modeling module 101 receives the target frequency band and the target impedance. Z tar The thickness of the continuous metal core, the thickness of the upper and lower low-loss dielectric layers, and the relative permittivity at the corresponding frequency point within the target frequency band. ε r and the dielectric loss tangent tan δ The copper layer thickness and surface roughness, hole processing tolerance and electroplating tolerance are determined, and a three-dimensional parametric model is established including the upper RF signal line 41, the upper gradient section 43, the upper signal pad 44, the center signal conductor 7, the stepped hole, the double-ended grounding blind hole 8, the lower signal pad 54, the lower gradient section 53 and the lower RF signal line 51.

[0052] For example, 1 GHz to 12 GHz and 50 Ω can be used as the initial calculation frequency band and target impedance, with the aforementioned dimensions as initial variables. However, this frequency band and target impedance should be adjusted according to the specific interface specification. If a material batch lacks available values ​​within the target frequency band... ε r tan δ Alternatively, the thickness data of the medium layer after lamination is used. The constraint modeling module 101 marks the batch as incomplete input and blocks the size determination. At this time, it is allowed to use conservative boundary values ​​with traceable sources to make material characterization test pieces instead of directly putting them into product board manufacturing, so as to prevent the supplier's typical value from being mixed with the effective value of the finished product.

[0053] Combination Figure 4 It can be seen that in step S102, the parameter solving module 102 calls the model established in step S101, and first determines the approximate relationship of the main body segments. D 1 and d The first round ratio, then joint scanning D 2. H c , D p , L t The number and arrangement radius of the grounding blind holes 8. The parameter solving module 102 outputs a window of manufacturable parameters that satisfy the minimum insulation wall thickness, hole spacing, termination position of the grounding blind holes 8, and electroplating capability, rather than a single optimum that only holds true in an ideal model.

[0054] For example, when increasing D 2. When it helps to reduce the capacitive effect of the orifice but results in insufficient termination position of the grounding blind hole 8, the parameter solving module 102 is retained. D The manufacturing limit of 2 is then compared to a smaller value. D p ,Change H c or extend L t The combination of these factors is considered; if none of the combinations satisfy the insulation and hole spacing conditions, a design instruction to redistribute the board space is output without relaxing the structural safety boundary. This allows electromagnetic compensation and manufacturing constraints to converge within the same calculation cycle.

[0055] like Figure 4 As shown, in step S103, the metal core processing equipment, based on the controlled hole processing parameter table and common axis reference output in step S102, first forms the main hole segment 11 on the continuous metal core 1, and then forms the larger diameter end compensation hole segment 12 coaxially from at least one side, preferably from both sides, and performs deburring, cleaning, and hole diameter inspection. If the actual... D 1. D 2 or H c If the dimensions exceed the manufacturable parameter window given in step S102, the process acquisition module 103 records the measurement results and isolates the plate, preventing the hole shape deviation from being masked by increasing the amount of filler material. If the dimensions are still within the parameter window but close to the boundary, the process acquisition module 103 retains the measured values ​​for subsequent model backfilling. This step ensures that the end compensation geometry has a measurable solid boundary while preventing burrs from the continuous metal core 1 from piercing the low-loss insulating filler 6.

[0056] Combination Figure 2 and Figure 4 As can be seen, in step S104, the filling equipment performs vacuum-assisted injection and segmented degassing on the stepped holes, so that the low-loss insulating filler 6 fills the main hole segment 11 and the end compensation hole segment 12; after curing and leveling, the upper low-loss dielectric layer 2, the lower low-loss dielectric layer 3 and the corresponding copper foil are stacked on both sides of the continuous metal core 1 and pressed together. For example, the gas in the main hole segment 11 can be discharged through a low vacuum stage first, and then the filling material can enter the cavity from one end or both ends; the specific pressure, temperature and time are determined according to the verified process data of the selected material, and this document does not limit unverified production conditions. If non-destructive inspection or witness sectioning shows through-holes, interface cracks or end shrinkage, the detection comparison module 104 blocks the processing of the center signal hole; uncured workpieces that can be reworked are returned to the cleaning and refilling process, and cured workpieces that cannot be reliably reworked are isolated, thereby preventing through-hole defects from developing into weak points in pressure resistance or thermal cycling.

[0057] like Figures 2 to 4 As shown, in step S105, the drilling equipment forms a signal hole in the center of the low-loss insulating filler 6 based on the reference target and stepped hole axis of the continuous metal core 1; the hole wall is cleaned, activated, chemically deposited copper and electroplated to form a hole copper layer 71 and a central signal conductor 7, which are compatible with the material system. Subsequently, grounding blind holes 8 terminating in the continuous metal core 1 are formed from the predetermined positions of the upper grounding conductor 42 and the lower grounding conductor 52, respectively, and the grounding blind holes 8 are made to be connected to the continuous metal core 1 by metallization of the hole wall; the pattern transfer equipment simultaneously forms the upper radio frequency signal line 41, the lower radio frequency signal line 51, the upper gradient section 43, the lower gradient section 53, the upper signal pad 44, the lower signal pad 54 and the C-shaped grounding ring 81. If the factory adopts a process of imaging before electroplating or multi-hole metallization, the local order of the steps can be adjusted, but the center signal conductor 7 must be insulated from the continuous metal core 1, the grounding blind hole 8 must be connected to the continuous metal core 1, and the two types of conductive networks should be identified differently in the manufacturing data and electrical test procedures.

[0058] refer to Figure 3 and Figure 4 In the electrical inspection of step S105, the detection comparison module 104 receives the insulation test results from the center signal conductor network to the continuous metal core 1, the conduction resistance test results from the upper grounding conductor 42 and the lower grounding conductor 52 to the continuous metal core 1, and the hole position measurement results, respectively. (Center deviation) e This refers to the distance between the axis of the center signal conductor 7 and the axis of the stepped aperture within a cross-section perpendicular to the axis of the stepped aperture, measured in millimeters. It can be obtained through slicing or by a suitable non-destructive imaging method. For example, the center deviation can be... e The ratio to the minimum allowable insulation wall thickness is used as a process monitoring quantity. The initial process warning threshold is selected within the range of 10% to 15%, and this ratio is required not to exceed the selected warning threshold. This threshold is not a universal release standard; its final value is determined jointly by the target withstand voltage, target frequency band, aperture tolerance, and process capability. When the structural release boundary is exceeded, the detection comparison module 104 locks subsequent surface treatment and isolates the board. When the warning boundary is reached but the structural limits are still met, data is recorded and cross-sectional sampling of adjacent test pieces is increased, without masking the physical eccentricity through software parameter modification. This graded handling can distinguish between process trend drift and structural failure.

[0059] like Figure 3As shown, a RF solder mask window 10 can be provided in the critical RF line area to prevent the upper RF signal line 41, lower RF signal line 51, upper gradient section 43, and lower gradient section 53 from being covered by the unverified solder mask layer; the solder pads can be selectively surface treated to meet assembly requirements. A copper layer with controlled surface roughness reduces the additional path of skin current along the rough interface, and the RF solder mask window 10 reduces the disturbance of the covering medium on the effective relative permittivity and dielectric loss tangent. Alternatively, a thin covering medium characterized by the target frequency band can be retained on the RF signal lines, but the parameter solving module 102 should specify its thickness, relative permittivity, etc. ε r and the dielectric loss tangent tan δ Incorporate into the model; do not directly equate the surface treatment results of ordinary line areas with the loss performance of long-distance radio frequency signal lines.

[0060] Combination Figure 4 and Figure 8 As can be seen, in step S106, the tester sequentially performs open-circuit and short-circuit checks, insulation tests, cross-sectional area checks, time-domain reflectometry tests, and two-port scattering parameter tests on the board and the same test strip. The test strip and the product board use the same stacking, material batch, copper layer treatment, and via metallization process, and may include equal-length planar through-line, back-to-back through-pass transition structure composed of two through-pass transitions in opposite directions, and calibration or de-embedding structures; the test ports and reference surfaces are arranged according to... Figure 8 The settings are configured to differentiate between planar RF signal line losses, the effects of clamps and connections, and the effects of through-core transitions. The detection comparison module 104 will then perform the actual... D 1. D 2. H c The dielectric layer thickness, via copper layer thickness, and material batch data are correlated with the controlled design version, and local impedance deviations (high or low) and scattering parameter deviations are sent to the correction output module 105. If both the equal-length planar through-line and the back-to-back through-core transition structure deviate simultaneously, the material parameters and RF signal line width are checked first. If the equal-length planar through-line is within the project limits but the back-to-back through-core transition structure deviates, the conductivity connection and center deviation of the end compensation hole section 12, the grounding blind hole 8, and the continuous metal core 1 are checked first to avoid attributing all differences to the stepped hole.

[0061] like Figure 4As shown, the correction output module 105 generates the next batch of controlled correction amounts based on the comparison results of step S106. When the transition region exhibits a relatively low local impedance trend, a combination of increasing the effective diameter of the end compensation hole segment 12, decreasing the equivalent diameter of the signal pad, decreasing the depth of the end compensation hole segment 12, or adjusting the ground conductor clearance can be compared, provided that the manufacturing boundary is met. When exhibiting a relatively high local impedance trend, a combination of decreasing the effective diameter of the end compensation hole segment 12, moderately increasing the equivalent diameter of the signal pad, adjusting the depth of the end compensation hole segment 12, or shortening the ground return distance can be compared. Each correction retains at least one unchanged benchmark test kit, and the traceability recording module 106 records the input data, model version, processing documents, measurement results, and approval status. If data is missing or the versions are inconsistent, the traceability recording module 106 blocks the release of new documents, and the system is downgraded to using the most recently completed controlled version with structural and electrical verification, without splicing local parameters from different sources.

[0062] Combination Figure 4 and Figure 8 As can be seen, this embodiment does not use unverified fixed return loss or insertion loss values ​​as the product conclusion. Instead, it establishes RF electrical release limits according to the project interface specifications, and uses structural safety limits as a prerequisite. Even if the RF response is temporarily within the project limits, if the center signal conductor 7 is found to be short-circuited to the continuous metal core 1, or if the low-loss insulation filler 6 has a through-hole, or if the grounding blind hole 8 is not reliably connected to the continuous metal core 1, the detection comparison module 104 should still determine that the structure is unqualified. Conversely, a board with an intact structure but an RF response exceeding the limit cannot be released on the grounds that the DC continuity is normal. The dual constraints of structural safety limits and RF electrical limits establish a traceable causal relationship between manufacturing deviations, insulation reliability, and RF performance.

[0063] like Figure 5 As shown, the three parallel embodiments in this section are all used to achieve insulated through-core transmission between the upper and lower surface radio frequency lines in a continuous metal core 1 environment, and all follow the principles of electrical isolation between the center signal conductor 7 and the continuous metal core 1, proximity-based conduction between the surface ground conductor and the continuous metal core 1, and the participation of measured parameters of the fabricated board in impedance collaborative control. The differences between the embodiments are not only in the size variations, but also in the different ways in which the dielectric boundary, the outer metal conductor boundary, or the thermal-RF layout relationship is formed.

[0064] The first parallel embodiment employs a radial dual-medium structure. For example... Figure 5As shown in the first combined cross-section, the low-loss insulating filler 6 is radially divided into a first dielectric region 61 and a second adhesive dielectric region 62 along the main hole section 11 and the end compensation hole section 12. The first dielectric region 61 surrounds the central signal conductor 7 and forms the main electric field passage area. The second adhesive dielectric region 62 is located between the first dielectric region 61 and the hole wall of the continuous metal core 1, and is used to improve the wetting, curing shrinkage, or interfacial bonding conditions between the first dielectric material and the metal hole wall. During manufacturing, the target impedance is first determined... Z tar Candidate windows for radial thickness are obtained by calculating the relative permittivity and dielectric loss tangent of the two media at corresponding frequency points within the target frequency band. Two media regions are then formed with the same central axis. During testing, cross-sectional inspection or applicable non-destructive testing methods are used to confirm that no through-hole defects have formed between the two media regions and between them and the central signal conductor 7 and the continuous metal core 1.

[0065] Combination Figure 5 It can be seen that in this radial dual-dielectric structure, the first dielectric region 61 is not required to fill the entire stepped hole. Therefore, while maintaining controlled dielectric loss in the central high-frequency electric field region, the second bonding dielectric region 62 can undertake the process task of bonding with the hole wall of the continuous metal core 1. The effective relative permittivity and interface position of the two dielectrics together change the distributed capacitance of the main hole segment 11. The nominal relative permittivity of the first dielectric material cannot be substituted into the single-dielectric approximation formula alone.

[0066] For example, the first dielectric region 61 can be configured as a continuous cylinder or a segmented cylinder that mates with the end compensation hole segment 12, and the second adhesive dielectric region 62 can be configured as a circumferentially continuous annular layer. In an alternative embodiment, a gentle radial transition interface can be used between the two dielectric regions instead of a geometrically abrupt interface. This structure reduces the mutual constraint between low-loss performance and metal interface bonding performance when both are satisfied by the same filler material through material functional partitioning, and allows subsequent corrections to dielectric parameters to correspond to the central electric field region and the outer peripheral adhesive region respectively.

[0067] The second parallel embodiment employs a structure combining a continuous aluminum metal core with a conductive outer jacket. For example... Figure 5 As shown in the second combined cross-section, the continuous metal core 1 can be made of aluminum or an aluminum alloy conductive metal, and a conductive jacket 13 is provided at the hole walls of the main hole section 11 and the end compensation hole section 12. The inner surface of the conductive jacket 13 forms the radio frequency outer conductor boundary surrounding the low-loss insulating filler 6 and the central signal conductor 7, and its outer surface forms an electrical connection with the continuous metal core 1; the upper and lower grounding blind holes 8 terminate at the solid area of ​​the continuous metal core 1 outside the conductive jacket 13 or the corresponding side through the C-shaped grounding ring 81, and are connected to the corresponding structure. This structure does not rely on the premise that the aluminum hole wall can directly obtain the same metallization adhesion conditions as the copper hole wall, but rather establishes a detectable continuous metal boundary through the independent conductive jacket 13.

[0068] refer to Figure 5 The conductive jacket 13 can be a continuous conductive layer formed after hole wall treatment, or it can be a conductive cylindrical component that mechanically fits with a stepped hole and achieves grounding connection at both ends. When using a dissimilar metal interface, a barrier layer or sealing structure can be set according to the metal combination to suppress the influence of interface corrosion on long-term contact resistance, but the specific material grade is not limited. Manufacturing inspection should verify the axial continuity of the conductive jacket 13, the conductivity between the conductive jacket 13 and the continuous metal core 1, and the insulation between the conductive jacket 13 and the center signal conductor 7. If the thickness of the conductive jacket 13 changes the effective inner diameter of the main hole segment 11, the parameter solution should use the effective inner diameter after the conductive jacket 13 is formed, and not the original machined hole diameter of the continuous metal core 1. Thus, while retaining the mass and thermal diffusion selection space of the aluminum continuous metal core 1, the vertical transition outer conductor boundary and the conductive connection interface of the grounding blind hole 8 can still have a clear and traceable physical basis.

[0069] The third parallel embodiment employs a structure that coordinates radio frequency through-core transition with a local heat dissipation hole array. For example... Figure 5 As shown in the third combined cross-section, when the upper circuit layer 4 or the lower circuit layer 5 carries an RF device with a grounded heat dissipation surface, multiple conductive heat dissipation blind holes 9 are provided in the grounded heat dissipation area of ​​the RF device. The conductive heat dissipation blind holes 9 pass through the corresponding low-loss dielectric layer and terminate at the continuous metal core 1 and are connected to the continuous metal core 1; the center signal conductor 7, the end compensation hole segment 12, the grounding blind hole 8, and the C-shaped grounding ring 81 are arranged in the corresponding area of ​​the RF port. The conductive heat dissipation blind holes 9 are used to transfer heat from the RF device and provide DC or low-frequency grounding connection, while the grounding blind holes 8 are mainly used to constrain the near-end return current of the through-core RF transition; although both are connected to the continuous metal core 1, their number, spacing, and termination position should be determined according to their different functions.

[0070] Combination Figure 5It is understood that the layout input of this structure includes the outer edge of the end compensation hole section 12, the opening reserved for the RF signal line by the C-shaped grounding ring 81, the physical space occupied by the conductive heat dissipation blind hole array 9, and the minimum solid connection width of the continuous metal core 1. The designer first ensures that each grounding blind hole 8 terminates in the continuous metal area outside the end compensation hole section 12, and then makes the conductive heat dissipation blind holes 9 avoid the electric field concentration areas of the upper gradient section 43, the lower gradient section 53, the upper signal pad 44, and the lower signal pad 54; if the RF device is located on one side, the conductive heat dissipation blind hole array 9 can be set only on that side, while the grounding blind holes 8 on the upper and lower sides still form a double-ended return path around the central signal conductor 7. In an alternative embodiment, the conductive heat dissipation blind hole 9 can be a metallized hole wall, a conductive filled hole, or a combination of both, as long as it remains insulated from the central signal conductor 7 and forms a detectable thermal and electrical connection with the continuous metal core 1. This structure avoids replacing the thermal path design by increasing the number of grounding blind holes 8, and also avoids the conductive heat dissipation blind hole array 9 from intruding into the end compensation area and unintentionally changing the impedance boundary, so that the heat diffusion path and the RF return path share the continuous metal core 1, while still being able to be checked separately.

[0071] like Figure 6 As shown, the manufacturing and inspection closed loop employs a state transition method for anomalies: trigger judgment, panel locking or process degradation, record feedback, and reset upon meeting certain conditions. Locking means prohibiting the corresponding panel or batch from continuing into subsequent irreversible processes or release stages; degradation means maintaining the most recently verified controlled version or reverting the process to a preceding process that can be re-inspected. Locking objects can be single panels, the same panel, or in-process batches using the same material batch and procedure version. Any anomaly record itself does not constitute a release basis; quality personnel only perform a reset when the cause of the anomaly has been addressed and the verification result re-enters the controlled parameter window.

[0072] like Figure 6 As shown, the first category involves material or process parameters exceeding limits. After the process acquisition module 103 obtains the thickness of the dielectric layer after lamination, batch dielectric parameters, and the thickness of the copper layer in the hole or the actual size of the stepped hole, the testing equipment and quality personnel will compare it with... Figure 4 The input parameter window established in step S101 is compared with the current controlled drawing. For example, when the actual dielectric parameter falls outside the original parameter scanning range, or when the dimensions of the main hole segment 11 and the end compensation hole segment 12 exceed the tolerance zone of the corresponding drawing, batch locking is triggered to prevent it from continuing to be released according to the original RF structure parameters; the relevant measured values, board number and version are recorded and fed back to the parameter solving module 102.

[0073] After the designers recalculate using the measured values, the panel is reset in a controlled deviation manner only if the new calculation results still meet the insulation boundary and pass the re-inspection; otherwise, the material is replaced, the processing instructions are revised, or the panel is remanufactured.

[0074] The second category is non-responsive processing or process deviation. If the machining of stepped holes, filling and curing, electroplating of center signal holes, or metallization of grounding blind holes (8 processes) does not return a completion receipt, or if the recorded machining trajectory, curing process, or electroplating result exceeds the approved process parameter window, the manufacturing control terminal will change the corresponding board from allowed to proceed to the next process to hold for review, and stop subsequent irreversible machining. Equipment operators check the program version, tooling references, and equipment status; if it can be re-executed without damaging the existing insulation structure and metal interface, it can be retried once after authorization by quality personnel, and the results can be re-collected; if it cannot be reliably reworked, it should be isolated. Only when the new completion receipt is consistent with the physical inspection results, and the record chain covers the reason and result of the retry, will the process be resumed.

[0075] The third category is structural detection anomalies. (Combined with...) Figure 6 If, during cross-sectional inspection, applicable non-destructive testing, or electrical inspection, a through-hole is found in the low-loss insulating filler 6, the center deviation of the center signal conductor 7 causes the minimum insulating wall to enter the prohibited area, the grounding blind hole 8 is not conductive with the continuous metal core 1, or the conductive jacket 13 shows axial discontinuity, the testing equipment generates an abnormal result, and quality personnel lock the corresponding board and block the RF test release. For defects that only involve reworkable surface patterns and do not damage the basic structure, the process can be returned to the corresponding pattern process; for defects involving short circuits between the center signal conductor 7 and the continuous metal core 1, through-insulation defects, or defects where the continuity of the hole copper layer 71 cannot be confirmed, physical treatment is not replaced by plating or software parameter correction. The reset condition is that the same structural inspection, insulation test, and continuity test are completed again after rework, and a unique correlation is established between the records before and after rework; those that cannot meet the conditions should be kept isolated.

[0076] The fourth category includes data acquisition failures, communication interruptions, and file verification failures. When material parameters, drilling files, outer layer graphic files, cross-sectional results, or RF measurement files are transferred between devices, if the verification results are inconsistent, the file version does not match the board production flow record, a complete record is not received within the specified return period, or conflicting duplicate values ​​appear in the same data packet, the receiving end will refuse to write the data into the release record and will retain the previous confirmed version. Operators can retransmit using a controlled local import method, but must not overwrite the original file with manually copied single values. After retransmission, both the sending and receiving ends will verify the file identifier, board number, and completeness. Only when the verification is consistent and missing fields are filled in will the abnormal state be reset to prevent incorrect versions of hole machining parameter tables or material data from entering impedance compensation calculations.

[0077] The fifth category includes test signal acquisition failures or RF results exceeding the limits. For example... Figure 6 and Figure 8As shown, when the calibration structure check fails, the test ticket connection status is abnormal, the time-domain reflectometry waveform is not repeatable, or the peak-valley impedance and scattering parameters of the back-to-back through-hole transition structure exceed the given limits, the tester first locks the RF release status of that batch. If the failure is due to fixture, connection, or calibration, the connection status is adjusted, recalibrated, and the same test ticket is retested; if the retest is still abnormal, the results of the equal-length planar straight-through line, the back-to-back through-hole transition structure, and the calibration structure are compared to locate the problem in the measurement chain, the planar RF signal line, or the through-hole transition. For example, the time-domain reflectometry test allow window can be set according to the target impedance. Z tar A certain percentage is determined; 10% is only used as an example of the initial design threshold, not as a conclusion drawn from existing measurements. Release is only restored when the measurement chain verification passes and the retest results return to within the project limits; if it is determined to be a systematic impedance deviation, it is returned to normal. Figure 4 In steps S101 and S102, parameters are corrected and new samples are made, allowing the out-of-limit finished products to be released without data smoothing.

[0078] The sixth category includes inconsistent status, repeated startups, or abnormal resets. When the same board number simultaneously displays both pending and completed status, the current equipment program version is inconsistent with the controlled version, or the operator issues a restart command for a completed step, the process control terminal refuses to execute and freezes the status change of that board number. Manufacturing personnel check the baseline, production flow record, equipment receipt, and physical process traces on the board. The traceability record module 106 retains repeated commands without overwriting the first record. Status restoration requires confirmation of a unique and reliable status by personnel with release responsibilities, followed by verification of the previous process output by the inspection stage; the freeze is only lifted when the preceding physical object matches the record. Resets after abnormal power-on of the equipment also first restore the read-only record and controlled version before allowing the loading of processing commands to avoid repeated drilling, repeated electroplating, or incorrect pattern processing caused by abnormal resets.

[0079] like Figure 7 As shown in the figure, modules 101 to 106 are used to illustrate the auxiliary recording relationship between the structural design, manufacturing data, and test results of the present invention, and do not constitute the composition of the finished double-sided metal-based circuit board. This auxiliary recording relationship may include a constraint modeling module 101, a parameter solving module 102, a process acquisition module 103, a test comparison module 104, a correction output module 105, and a traceability recording module 106; each module can be encapsulated in the same manufacturing design and testing program, or it can be carried separately on the design terminal, process terminal, and testing terminal that exchange controlled documents with each other.

[0080] Constraint modeling module 101 receives target frequency band and target impedance. Z tarThe parameter set, including the material dielectric parameters, dielectric layer thickness, and manufacturing tolerances, is established, comprising the main hole segment 11, end compensation hole segment 12, center signal conductor 7, upper signal pad 44, lower signal pad 54, upper gradient segment 43, lower gradient segment 53, and grounding blind hole 8. The parameter solving module 102 then compares the candidate combinations within the constraints of insulation wall thickness, hole spacing, and the termination position of the grounding blind hole 8, and outputs the result. Figure 4 The hole processing parameter table, graphic parameters and version identifier are called in steps S102 to S105; when the input is missing or the constraints conflict, the parameter solving module 102 does not output the production command, but sends the status to the traceability recording module 106 and keeps the previous controlled version.

[0081] The process acquisition module 103 obtains dielectric layer thickness, dielectric parameters, stepped via dimensions, via copper layer thickness, center deviation, and process feedback from manufacturing equipment or verified measurement documents. After unifying the unit, board number, batch, and version, it sends the data to the detection and comparison module 104. The detection and comparison module 104 then processes the data... Figure 8 The structural, electrical, and RF data obtained from the test ticket are compared with the corresponding controlled parameter windows, and the output determines whether to allow sequential transfer, maintain verification, or lock out isolation. If the acquisition file verification fails, the process acquisition module 103 retains the original file and requests re-acquisition; before the new file verification is completed, the detection comparison module 104 must not use missing or speculative values ​​to replace the measured input. The correction output module 105 only forms geometric correction candidates for the next design or subsequent batch based on at least two of the actual dielectric layer thickness, dielectric parameters, via copper layer thickness, and center deviation after the detection comparison module 104 confirms the data is complete, and returns them to the constraint modeling module 101 for re-checking, without directly rewriting the board's pass / fail status. The traceability record module 106 saves the input source, calculation version, manufacturing receipt, judgment result, lockout and reset actions, and approved output files, ensuring that each correction candidate corresponds to a specific material batch and test ticket. Correction calculations can be achieved by using table lookup, parameter scanning, or fitting relationships calibrated by samples. However, the outputs must be reviewed by the constraint modeling module 101 and verified by entity samples before a new controlled version can be formed.

[0082] Figure 7A data processing device 200 capable of supporting the aforementioned auxiliary recording relationships is also shown. The data processing device 200 includes a processor 201, a memory 202, a communication bus 203, a communication interface 204, and a human-machine interface 205. The processor 201 reads instructions and controlled parameters from the memory 202 via the communication bus 203, receives data from manufacturing equipment receipts, measurement documents, or data transmitted from the local production network via the communication interface 204, and presents the version, locking reason, and reset condition via the human-machine interface 205. The processor 201 can sequentially call modules 101 to 106 to complete constraint modeling, parameter solving, process data acquisition, detection comparison, correction candidate generation, and traceability recording. When the communication interface 204 is interrupted, the data processing device 200 retains the last confirmed version and prohibits the release of new processing documents until the connection is restored and the document integrity verification is passed before continuing processing.

[0083] Data processing equipment 200 can be a manufacturing design computer, process control terminal, or testing data processing terminal, and its number is not limited. When multiple data processing devices share modules, each device is associated with input and output by board number, batch number, and version identifier; when a single data processing device carries all modules, the modules exchange results through the controlled data area in memory 202. The manual reset command received by human-machine interface 205 only takes effect when the test comparison module 104 has written the verification result, so as to avoid bypassing structural, electrical, or radio frequency re-inspection based solely on manual confirmation.

[0084] The aforementioned auxiliary recording and data processing equipment is only used for the structural design, manufacturing inspection, and subsequent batch parameter correction of the circuit board of the present invention. It is not part of the finished double-sided metal-based circuit board and does not change the physical structure and connection relationship between the continuous metal core 1, the stepped hole, the insulating filler 6, the center signal conductor 7, and the grounding blind hole 8.

[0085] The above description is merely a specific embodiment of the present invention, and the scope of protection of the present invention is defined by the claims. Any equivalent substitutions or modifications made by those skilled in the art without departing from the inventive concept should fall within the scope of protection of the present invention.

Claims

1. A low-loss double-sided metal-based circuit board for radio frequency signals, comprising a continuous metal core, with a dielectric layer and a circuit layer sequentially disposed on each side of the continuous metal core, the circuit layer on each side including a radio frequency signal line, a signal pad, and a ground conductor; the continuous metal core has a through-hole, the through-hole being filled with an insulating filler, and a central signal conductor insulated from the continuous metal core is disposed within the insulating filler, the two ends of the central signal conductor being respectively connected to the signal pads on both sides; characterized in that: The through hole is a stepped hole, including a main hole section and an end compensation hole section located at at least one end of the main hole section. The end compensation hole section extends inward from the surface of the continuous metal core to the main hole section and has a diameter larger than that of the main hole section. Each signal pad is connected to the RF signal line on the same side via a planar gradient section; Multiple grounding blind holes are provided in the dielectric layer on each side, surrounding the stepped hole. Each grounding blind hole connects the grounding conductor on that side to the continuous metal core.

2. The low-loss double-sided metal-based circuit board for radio frequency signals according to claim 1, characterized in that: There are two end compensation holes, located at both ends of the main hole section and coaxial with the main hole section. The main hole section is located in the middle of the thickness of the continuous metal core. Multiple grounding blind holes on each side are located radially outside the end compensation hole section on that side, and terminate in the continuous metal area outside the end compensation hole section on that side where the continuous metal core is located. The diameter of the circle at the center of the grounding blind hole is larger than the diameter of the end compensation hole section.

3. The low-loss double-sided metal-based circuit board for radio frequency signals according to claim 2, characterized in that: The grounding conductors on each side form a C-shaped grounding loop around the central signal conductor. The C-shaped grounding loop has an opening on that side where the RF signal line enters the signal pad. Multiple grounding blind vias on that side are arranged along the C-shaped grounding loop. The planar gradient section on this side connects to the signal pad through an opening. The line width of the planar gradient section, the gap between the planar gradient section and the adjacent ground conductor, and the local clearance profile of the ground conductor gradually change from the RF signal line to the signal pad.

4. The low-loss double-sided metal-based circuit board for radio frequency signals according to claim 3, characterized in that: The center signal conductor is a metallized via conductor coaxial with the stepped via, and the copper layer of the metallized via conductor is connected to the signal pads on both sides. The insulating filler fills the main hole segment and the end compensation hole segment and fits against the hole walls of the main hole segment and the end compensation hole segment. The insulating filler forms a circumferentially continuous insulating wall between the central signal conductor and the continuous metal core along the entire crossing length of the central signal conductor. The continuous metal core remains electrically conductive around the stepped aperture and is not divided into insulated islands.

5. The low-loss double-sided metal-based circuit board for radio frequency signals according to claim 2, characterized in that: The insulating filler includes a first dielectric region and a second adhesive dielectric region arranged coaxially. The first dielectric region surrounds the central signal conductor, and the second adhesive dielectric region is located between the first dielectric region and the hole walls of the main hole section and the end compensation hole section. The second adhesive dielectric region is a circumferentially continuous annular layer. The interface between the first dielectric region and the second adhesive dielectric region is continuous in the circumferential direction, and neither the first dielectric region nor the second adhesive dielectric region forms a through gap between the central signal conductor and the continuous metal core.

6. The low-loss double-sided metal-based circuit board for radio frequency signals according to claim 2, characterized in that: The continuous metal core is made of aluminum. The main hole section and the end compensation hole section are provided with a conductive sleeve that is continuous along the stepped hole axis. The inner surface of the conductive sleeve is surrounded by an insulating filler and a central signal conductor. The outer surface of the conductive sleeve is electrically connected to the continuous metal core. The conductive sleeve is insulated from the central signal conductor. The grounding blind hole on each side terminates at least one of the conductive jacket and the solid region of the continuous metal core located outside the conductive jacket, and is in communication with the continuous metal core.

7. The low-loss double-sided metal-based circuit board for radio frequency signals according to claim 2, characterized in that: At least one side of the circuit layer includes a grounding and heat dissipation area for radio frequency devices. The dielectric layer on this side is provided with a plurality of conductive heat dissipation blind holes. Each conductive heat dissipation blind hole passes through the dielectric layer on this side from the grounding and heat dissipation area of ​​the radio frequency devices and terminates at the continuous metal core, thereby connecting the grounding and heat dissipation area of ​​the radio frequency devices with the continuous metal core and is insulated from the center signal conductor. The conductive heat dissipation blind via is located on the periphery of the end compensation hole section and avoids the planar gradient section and the area where the signal pads are located on that side; the conductive heat dissipation blind via adopts at least one of the following forms: metallized hole wall and conductive filler.

8. The low-loss double-sided metal-based circuit board for radio frequency signals according to claim 2, characterized in that: The RF signal lines and planar gradient sections on each side are located within the RF solder mask window area and are not covered by the solder mask layer. The equivalent diameter of the signal pad is 0.40 mm to 0.55 mm, and the length of the planar gradient segment along the center line of the RF signal propagation is 0.5 mm to 1.5 mm. The number of grounding blind holes on each side is 8 to 12, the diameter of a single grounding blind hole is 0.10 mm to 0.20 mm, and the diameter of the circle arranged at the center of the grounding blind hole is 1.8 mm to 2.2 mm.

9. The low-loss double-sided metal-based circuit board for radio frequency signals according to claim 8, characterized in that: The depth of the end compensation hole section from the surface of the continuous metal core to the interface between the end compensation hole section and the main hole section is 0.10 mm to 0.15 mm. The center signal conductor is a metallized through-hole conductor coaxial with the stepped via, and the thickness of the copper layer of the metallized through-hole conductor is 20 μm to 25 μm; the equivalent diameter of the signal pad is larger than the outer diameter of the center signal conductor and smaller than the diameter of the end compensation hole section; the grounding blind via is a blind via with metallized hole wall.

10. A low-loss double-sided metal-based circuit board for radio frequency signals according to claim 9, characterized in that: The continuous metal core is a copper-based metal core with a thickness of 0.8 mm, the dielectric layers on both sides are dielectric layers with a thickness of 0.203 mm after lamination, and the circuit layers on both sides are formed by copper foil with a thickness of 18 μm. The diameter of the main hole section is 1.10 mm, and the diameter of the end compensation hole section is 1.30 mm; the design target impedance of the RF signal line is 50 Ω.

Images