Method for improving signal transmission stability of function board

Abstract

The invention provides a method for improving signal transmission stability of a function board, which relates to the technical field of electronic information, and comprises the following steps of: presetting an impedance detection point at a key position of a DDR4 signal transmission channel, acquiring a characteristic impedance value and a crosstalk coefficient, dynamically calculating a compensation parameter by combining a microstrip line transmission theory, and generating a dynamic impedance matching matrix; meanwhile, an interlayer coupling coefficient is introduced to correct a compensation matrix, and finally real-time impedance matching and whole-course impedance continuous matching of a signal transmission channel are realized through a programmable impedance network. Thus, the problems of signal reflection and crosstalk caused by impedance mismatch during long-distance transmission of DDR4 high-speed memory signals in a PCB function board can be solved to a certain extent, the integrity and stability of signal transmission are effectively improved, the occurrence rate of signal reflection, overshoot and jitter is reduced, and especially in a complex multi-layer PCB design and a dynamic impedance environment, the signal transmission efficiency is improved. The signal transmission performance is obviously improved, and the reliability requirement of high-speed signal transmission is met.

Description

Technical Field

[0001] The present invention relates to the field of electronic information technology, and more specifically, to a method for improving the signal transmission stability of a functional board. Background Art

[0002] With the rapid development of modern electronic technology, high-speed signal processing technology has become a core area in electronic device design. In particular, DDR (Double Data Rate) memory technology has been widely used in high-performance computing, data centers, and storage systems. As the current mainstream memory technology, DDR4 significantly increases its data transmission rate and signal integrity requirements. In high-frequency, high-speed signal transmission environments, signal integrity (SI) issues directly impact system performance. These issues, including signal reflection, crosstalk, and impedance mismatch, present technical difficulties that urgently need to be addressed in design. To this end, microstrip transmission theory, impedance matching techniques, and parasitic capacitance optimization have gradually become research focuses in the field of PCB (Printed Circuit Board) design. However, as DDR4 signal rates increase, existing technologies still face many challenges in addressing signal transmission stability.

[0003] Existing signal transmission optimization methods mostly reduce signal reflections and crosstalk through fixed impedance matching, electromagnetic shielding materials, or designed stacked structures. These methods perform well in low-speed signal transmission scenarios, but in DDR4 high-speed signal transmission, due to the higher signal frequency, parasitic effects and interlayer coupling effects during transmission are more significant, making traditional static compensation methods difficult to adapt to the ever-changing signal transmission environment. For example, fixed impedance matching methods cannot respond in real time to dynamic impedance deviations caused by environmental changes (such as temperature, load, or frequency changes); at the same time, existing methods lack an effective dynamic compensation mechanism when dealing with coupling problems caused by parasitic capacitance between PCB layers. Especially for complex multi-layer board designs, the interlayer coupling effect directly affects the stability of the signal transmission channel. Existing technologies are unable to achieve continuous impedance matching throughout the entire signal transmission process, resulting in a decrease in signal integrity, manifested as increased signal reflections, overshoot, and severe jitter problems. Summary of the Invention

[0004] To address the above technical problems, the present invention provides a method for improving the stability of signal transmission on a functional board, which can, to a certain extent, solve the signal reflection and crosstalk problems caused by impedance mismatch during long-distance transmission of DDR4 high-speed memory signals on a PCB functional board.

[0005] According to one aspect of the present invention, a method for improving the signal transmission stability of a function board is provided, comprising:

[0006] Preset multiple impedance detection points at key locations of the DDR4 signal transmission channel and collect characteristic impedance values ​​and crosstalk coefficients of the impedance detection points;

[0007] Based on the characteristic impedance value and the crosstalk coefficient, impedance compensation parameters are calculated in combination with microstrip line transmission theory to generate a dynamic impedance matching matrix;

[0008] Measuring parasitic capacitance values ​​between PCB layers, calculating an interlayer coupling coefficient based on the parasitic capacitance values, and using the interlayer coupling coefficient to correct the dynamic impedance matching matrix to obtain a compensation matrix that takes into account the interlayer coupling effect;

[0009] According to the compensation matrix that takes into account the interlayer coupling effect, real-time impedance matching is performed on the DDR4 signal transmission channel through a programmable impedance network, and interpolation operation is performed using the compensation parameters of adjacent impedance detection points to achieve continuous impedance matching throughout the signal transmission channel.

[0010] Furthermore, the impedance detection point adopts a through-hole embedded test probe structure;

[0011] The impedance detection points are arranged alternately on both sides of the differential pair signal lines, that is, a first detection point is set on the positive line side of the differential signal line, and a second detection point is set on the negative line side at a preset distance from the first detection point, and the remaining detection points are set in this way.

[0012] Furthermore, generating the dynamic impedance matching matrix includes:

[0013] An impedance calculation model is constructed based on the frequency characteristics and material parameters during DDR4 high-speed signal transmission to obtain the actual measured characteristic impedance value.

[0014] Calculate the impedance deviation value of the monitoring point and calculate the equivalent impedance correction value using the dynamic coupling strength evaluation function, and calculate the total compensation value;

[0015] The equivalent impedance correction amount and the total compensation value are iteratively optimized to obtain the dynamic impedance matching matrix.

[0016] Furthermore, the calculation of the equivalent impedance correction amount is shown in the following formula:

[0017]

[0018] in,

[0019] K coupling (d) = k0e -γd (1+δsin(ωt))

[0020] Among them, Z correctionis the equivalent impedance correction, L is the signal line self-inductance, M is the mutual inductance coefficient, C is the signal line self-capacitance, C m is the mutual capacitance coefficient, K coupling (d) is the coupling strength coefficient, d is the distance between adjacent signal lines, k0 is the reference coupling coefficient, γ is the spatial attenuation factor, δ is the time-varying modulation coefficient, ω is the angular frequency, and t is the time variable.

[0021] Furthermore, the dynamic impedance matching matrix is ​​shown as follows:

[0022]

[0023] Among them, P ij is the element in row i and column j of the compensation matrix, η is the iterative learning rate, Z comp,i is the total compensation value of the i-th detection point, Z correctoon,m is the equivalent impedance correction value of the mth detection point, and k is the number of iterations.

[0024] Furthermore, calculating the interlayer coupling coefficient includes:

[0025] Mark the strong coupling point through the local detection window to determine whether it is located in the signal line intersection area and determine the angle correction factor;

[0026] With the strong coupling point as the center, the electric field gradient is calculated and the high coupling area is marked;

[0027] Calculating the regional coupling coefficient of the high coupling area, and determining whether to use the skin effect correction factor or the power supply coupling suppression factor for correction according to the detection area;

[0028] Multiplying the angle correction factor, the power supply coupling suppression factor, and the skin effect correction factor to obtain a comprehensive correction coefficient;

[0029] The interlayer coupling coefficient is obtained by multiplying the comprehensive correction coefficient by the regional coupling coefficient.

[0030] Furthermore, using the interlayer coupling coefficient to modify the dynamic impedance matching matrix includes:

[0031] The interlayer coupling coefficient measurement points are aligned with the impedance detection points in the dynamic impedance matching matrix to establish a spatial mapping relationship;

[0032] Perform correction calculation on each impedance detection point in the dynamic impedance matching matrix;

[0033] Conduct continuity checks on the corrected impedance values ​​in the vertical direction and optimize the corrected impedance distribution along the signal transmission path;

[0034] The corrected and optimized impedance values ​​are re-filled into the dynamic impedance matching matrix according to the position sequence of the detection points to generate a compensation matrix that takes into account the interlayer coupling effect.

[0035] Furthermore, the correction calculation includes:

[0036] Determine the equivalent coupling strength of each impedance detection point;

[0037] The original impedance value is corrected according to the equivalent coupling strength. When the equivalent coupling strength is greater than or less than the preset reference value, attenuation or compensation processing is performed respectively; the attenuation or compensation processing adopts a piecewise linear mapping method.

[0038] Furthermore, by writing the compensation matrix into the programmable impedance network, the control unit calculates and adjusts the resistance value of the impedance unit, combining high-speed sampling and step-by-step approximation adjustment method to achieve real-time impedance matching of the DDR4 signal transmission channel.

[0039] Furthermore, the interpolation operation identifies adjacent detection point pairs and divides the interpolation interval, calculates the interpolation parameters based on a distance weighted method, and generates a parameter distribution that meets the impedance continuous matching requirement.

[0040] Compared with the prior art, the method for improving the signal transmission stability of the functional board provided by the present invention presets impedance detection points at key positions of the DDR4 signal transmission channel, collects characteristic impedance values ​​and crosstalk coefficients, dynamically calculates compensation parameters in combination with microstrip line transmission theory, generates a dynamic impedance matching matrix, and introduces interlayer coupling coefficients to correct the compensation matrix. Finally, real-time impedance matching and continuous impedance matching of the entire signal transmission channel are achieved through a programmable impedance network. In this way, the signal reflection and crosstalk problems caused by impedance mismatch during long-distance transmission of DDR4 high-speed memory signals in PCB functional boards can be solved to a certain extent, effectively improving the integrity and stability of signal transmission, reducing the incidence of signal reflection, overshoot and jitter, and significantly improving signal transmission performance, especially in complex multi-layer PCB designs and dynamic impedance environments, meeting the reliability requirements of high-speed signal transmission. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0042] Figure 1 The figure is a flow chart of a method for improving signal transmission stability of a function board according to an embodiment of the present invention.

[0043] Figure 2 This is a flow chart of real-time impedance matching in a method for improving signal transmission stability of a functional board according to an embodiment of the present invention. DETAILED DESCRIPTION

[0044] Below, the exemplary embodiments according to the present invention will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments of the present invention, and it should be understood that the present invention is not limited to the exemplary embodiments described herein.

[0045] Figure 1 Flowchart of a method for improving the stability of signal transmission of a function board according to an embodiment of the present invention. Figure 1 As shown, the method for improving the signal transmission stability of the function board includes:

[0046] S1: Preset multiple impedance detection points at key locations of the DDR4 signal transmission channel and collect characteristic impedance values ​​and crosstalk coefficients at the impedance detection points, wherein the impedance detection points are staggered on both sides of the differential pair signal lines;

[0047] Preset multiple impedance detection points at key locations of the DDR4 signal transmission channel, and collect characteristic impedance values ​​and crosstalk coefficients at the impedance detection points, specifically including:

[0048] First, on the PCB traces between the DDR4 memory and the controller chip, impedance detection points were set every 20 mm in the signal transmission direction. These impedance detection points used a through-hole embedded test probe structure. The impedance detection points were staggered on both sides of the differential signal lines. That is, a first detection point was set on the positive line side of the differential signal line, a second detection point was set 20 mm away from the first detection point on the negative line side, and the remaining detection points were set in this order. The impedance detection points were connected via a high-precision impedance analyzer, and the characteristic impedance of each detection point was measured using time domain reflectometry (TDR) technology to obtain the characteristic impedance value. Simultaneously, signal crosstalk waveforms were collected between adjacent differential signal lines using the impedance detection points. The crosstalk coefficient in the frequency domain was calculated using a fast Fourier transform. The crosstalk coefficient represents the degree of electromagnetic interference between adjacent signal lines. Specifically, the crosstalk coefficient was obtained by measuring the near-end crosstalk (NEXT) and far-end crosstalk (FEXT) of adjacent signal lines, and the crosstalk attenuation at different frequencies was recorded.

[0049] S2: Based on the characteristic impedance value and the crosstalk coefficient, impedance compensation parameters are calculated in combination with microstrip line transmission theory to generate a dynamic impedance matching matrix. The dynamic impedance matching matrix compensates for the impedance mismatch of the signal transmission channel in real time through a programmable impedance network;

[0050] First, in response to the impedance matching difficulties faced by high-speed DDR4 signals during PCB transmission, this solution has comprehensively improved the traditional microstrip line impedance calculation model. The traditional model has insufficient accuracy when processing high-frequency signals, which is mainly reflected in: not fully considering the frequency-dependent effect, ignoring material loss, and overly simplifying the description of the influence of geometric structure. To solve these problems, on the basis of the traditional microstrip line impedance calculation, taking into account the frequency characteristics and nonlinear changes of material parameters during DDR4 high-speed signal transmission, multiple correction factors are introduced to construct a more accurate calculation model. This model not only considers the basic physical parameters of the PCB dielectric material such as the dielectric constant εr, signal line width w, signal line thickness t, and the distance h between the signal line and the reference plane, but also incorporates the influence of process parameters such as dielectric loss angle tanδ and copper foil roughness Ra. Based on these parameters, an improved characteristic impedance calculation formula is established:

[0051]

[0052] The calculation of the compensation coefficient α fully considers the geometric characteristics and relative position relationship of the signal line:

[0053]

[0054] Where Z0 is the characteristic impedance of the microstrip line, in Ω, ε eff is the effective dielectric constant, dimensionless, h is the vertical distance from the signal line to the reference plane, in mm, w is the width of the signal line, in mm, t is the thickness of the signal line, in mm, and α is the compensation coefficient, dimensionless.

[0055] During high-speed signal transmission, the dielectric constant changes with frequency, so the calculation of the effective dielectric constant εeff needs to consider the frequency dependency. An improved formula for calculating the effective dielectric constant is proposed to more accurately reflect the changes in the dielectric properties of the material under high-frequency conditions:

[0056]

[0057] Among them, β is the frequency compensation coefficient, its value is related to the characteristics of the PCB material, f is the signal operating frequency, ε r is the relative dielectric constant of the PCB substrate, dimensionless. After obtaining the actual measured characteristic impedance value, it needs to be compared with the standard impedance value of the DDR4 differential signal (100Ω). To this end, the impedance deviation value of each test point is defined as:

[0058] ΔZ i =Z measured,i -Z standard

[0059] Where ΔZ iis the impedance deviation value of the i-th detection point, in Ω, Z measured,i is the actual impedance value measured at the i-th detection point, in Ω, Z standard It is the standard characteristic impedance value (100Ω).

[0060] Next, considering the importance of crosstalk in high-speed signal transmission, this solution provides a dynamic coupling strength evaluation function. It can accurately describe how the electromagnetic coupling strength between signal lines changes with distance and time. Specifically, it is shown in the following formula:

[0061] K coupling (d) = k0e -γd (1+δsin(ωt))

[0062] Among them, K coupling (d) is the dimensionless coupling strength coefficient, d is the distance between adjacent signal lines, k0 is the base coupling coefficient, γ is the spatial attenuation factor, δ is the time-varying modulation coefficient, ω is the angular frequency, and t is the time variable. By introducing the time-varying term δsin(ωt), the transient coupling effect during signal switching can be effectively described. Based on this function, the calculation formula for the equivalent impedance correction considering the dynamic coupling effect is established:

[0063]

[0064] Among them, Z correction is the equivalent impedance correction value, the unit is Ω, L is the signal line self-inductance, the unit is H, M is the mutual inductance coefficient, the unit is H, C is the signal line self-capacitance, the unit is F, C m is the mutual capacitance coefficient, unit is F, K coupling (d) is the coupling strength coefficient, dimensionless.

[0065] In order to achieve multi-point collaborative compensation, a compensation value calculation model considering multi-point coupling effects was established:

[0066]

[0067] in,

[0068]

[0069] Among them, Z comp,i is the total compensation value of the i-th detection point, in Ω, ΔZ i is the impedance deviation value of the i-th detection point, in Ω, λ ij is the coupling weight coefficient, dimensionless, Z correction,j is the equivalent impedance correction value of the jth detection point, in Ω, n is the total number of detection points, d ijis the physical distance between detection points i and j, in mm, d0 is the characteristic distance (related to the signal wavelength), in mm, ij is the detection point number. The coefficient matrix describes the degree of mutual influence between detection points.

[0070] Based on the above calculation results, an iterative optimization method is used to generate the n×n dimensional dynamic impedance matching matrix P:

[0071]

[0072] Among them, P ij is the element in row i and column j of the compensation matrix, with unit Ω, η is the iterative learning rate, dimensionless, ranging from (0,1), Z comp,i is the total compensation value of the i-th detection point, in Ω, Z correction,m is the equivalent impedance correction value of the mth detection point, in Ω, and k is the number of iterations. This iterative process continues until the matrix elements converge or the preset number of iterations is reached.

[0073] To ensure the practicality of the compensation effect, smoothness constraints are imposed on the compensation matrix. These constraints ensure that the impedance compensation value changes smoothly in space and avoid the generation of new impedance discontinuities:

[0074]

[0075] At the same time, in order to ensure the stability of the compensation scheme, matrix standard deviation control is introduced:

[0076]

[0077] Where σ(P) is the standard deviation of the compensation matrix, P ij is the compensation matrix element, is the average value of all elements of the compensation matrix, n σ is the matrix dimension, ∈ is the maximum allowed standard deviation,.

[0078] When σ(P) is less than the preset threshold ε, it is considered that the compensation matrix has achieved the required uniformity.

[0079] It should be noted that the generation process of the dynamic impedance matching matrix is ​​explained in detail using a specific DDR4 memory signal transmission scenario:

[0080] Suppose that on a multi-layer PCB, the differential signal pair between the DDR4 memory and the main control chip requires impedance matching. Multiple impedance measurement points are pre-set along the signal transmission path. For example, the differential pair signal originates from the main control chip, passes through the multi-layer PCB's vias, and ultimately reaches the memory chip. A time domain reflectometer is used to measure the first measurement point. The measured waveform reveals a significant impedance discontinuity at this location, manifested as a reflection on the rising edge of the waveform. By analyzing this reflected waveform, the characteristic impedance value at that point is obtained and compared with the standard differential impedance value to obtain the first deviation value.

[0081] Next, we observed that another adjacent differential pair of signal lines was causing crosstalk on the target signal. Near-end crosstalk measurements revealed that when the adjacent signal experienced a level jump, coupled noise was generated on the target signal line. The strength of this coupling was closely related to the distance between the two differential pairs; the closer the distance, the stronger the coupling. By analyzing this coupling phenomenon, we derived a coupling coefficient, which reflects the degree of crosstalk.

[0082] After obtaining the impedance deviation values ​​and coupling coefficients for all test points, the compensation matrix is ​​constructed. Taking the first test point as an example, its compensation value must consider not only its own impedance deviation but also the crosstalk from adjacent differential pairs. During the iterative optimization of the compensation matrix, after the initial few iterations, the compensation values ​​of some test points vary significantly. To ensure a smooth impedance transition, gradient constraints are introduced. For example, when the compensation value of a test point varies significantly compared to adjacent points, the optimization algorithm automatically adjusts the value to meet the smooth transition requirements. In the resulting dynamic impedance matching matrix, each element reflects the impedance compensation required at a specific location. The rows and columns of the matrix correspond to different test point locations, and the size of the matrix element represents the degree of impedance adjustment required at that location.

[0083] S3: measuring parasitic capacitance values ​​between PCB layers, calculating an interlayer coupling coefficient based on the parasitic capacitance values, and using the interlayer coupling coefficient to correct the dynamic impedance matching matrix to obtain a compensation matrix that takes interlayer coupling effects into account;

[0084] When calculating the interlayer coupling coefficient based on the parasitic capacitance value, the parasitic capacitance value of each 5×5 measurement point array on the PCB board is first normalized, and the normalization process refers to dividing the parasitic capacitance value of the measurement point by the maximum parasitic capacitance value in the area; then, a 3×3 local detection window centered on each measurement point is established, and the average parasitic capacitance of the 9 measurement points in the detection window is calculated. When the parasitic capacitance value of the central measurement point exceeds 1.5 times the average value, the measurement point is marked as a strong coupling point; then, it is determined whether the strong coupling point is located in the signal line intersection area. When it is located in the signal line intersection area, an angle correction factor is determined according to the signal line intersection angle θ, and the angle correction factor is equal to 1-sinθ; then, with the strong coupling point as the center, the electric field gradient values ​​of the 8 adjacent measurement points around are calculated, and the electric field gradient value is equal to the parasitic capacitance difference of the adjacent measurement points divided by the measurement angle. measurement point spacing; when the electric field gradient value is greater than 0.5pF / mm, the area is marked as a high coupling area; for the measurement points in the high coupling area, the distance weighted method is used to calculate the regional coupling coefficient, and the weight factor in the distance weighted method exponentially decays with the increase of the distance from the strong coupling point; when the measurement area is located between the signal layer and the power layer, a power coupling suppression factor is introduced, and the power coupling suppression factor is related to the number and distribution position of the decoupling capacitors in the power layer; when the measurement area is located between the signal layer and the ground layer, and the signal line width is greater than 1 / 5 of the thickness of the dielectric layer, a skin effect correction factor is introduced, and the skin effect correction factor is related to the signal frequency and material conductivity; finally, the angle correction factor, the power coupling suppression factor and the skin effect correction factor are multiplied to obtain a comprehensive correction coefficient, and the comprehensive correction coefficient is multiplied by the regional coupling coefficient to obtain the final interlayer coupling coefficient.

[0085] Specifically, when the distance weighted method is used to calculate the regional coupling coefficient, a polar coordinate system is first established with the strong coupling point as the center, and the surrounding measurement points are divided into an inner annular region, a middle annular region and an outer annular region according to the distance from the center point. The radius of the inner annular region is smaller than the radius of the middle annular region, and the radius of the middle annular region is smaller than the radius of the outer annular region; then, a basic weight value is assigned to the measurement points in each annular region, and the basic weight value is inversely proportional to the sequence number of the annular region, the basic weight value of the inner annular region is greater than the basic weight value of the middle annular region, and the basic weight value of the middle annular region is greater than the basic weight value of the outer annular region; then, the actual weight coefficient of each measurement point is calculated, and the actual weight coefficient is equal to the basic weight value multiplied by the distance attenuation value, where the distance attenuation value increases with the increase of the distance from the measurement point to the center point. Add and decrease, when the distance between the measuring point and the center point is d, the distance attenuation value is equal to e to the power of negative d / d0, where d0 is a characteristic distance related to the PCB board layer spacing; then, the parasitic capacitance values ​​of the measuring points in each annular area are weighted and summed, and the result of the weighted summation is divided by the sum of the actual weight coefficients in the annular area to obtain the equivalent parasitic capacitance value of the annular area; when there are multiple strong coupling points in a certain annular area, the equivalent parasitic capacitance values ​​corresponding to the strong coupling points are arithmetic averaged; further, the equivalent parasitic capacitance values ​​of the three annular areas are weighted averaged according to a preset area ratio, and the area ratio is determined in a step-by-step decreasing manner from the inner layer to the outer layer; finally, the weighted averaged equivalent parasitic capacitance value is divided by the standard parasitic capacitance value specified in the PCB design specification to obtain the coupling coefficient of the area.

[0086] The calculation process of the power coupling suppression factor is as follows: first, the area between the signal layer and the power layer is gridded, and the size of the grid division is smaller than the width of the signal line. The parasitic capacitance value of each grid area is associated with the distribution position of the decoupling capacitor in the area; then, the position coordinates and capacitance value of each decoupling capacitor are obtained, and a spatial distribution map of the decoupling capacitor is established. The spatial distribution map reflects the distribution density of the decoupling capacitors on the PCB board; then, a power noise propagation path is established for each grid area. The propagation path includes the horizontal distance from the signal line to the nearest decoupling capacitor and the vertical distance from the decoupling capacitor to the power layer; then, the power coupling is calculated based on the propagation path. Attenuation value, when the propagation path distance is longer, the attenuation value is larger. When the propagation path passes through multiple decoupling capacitors, the attenuation values ​​are superimposed and calculated. Furthermore, the PCB board area is divided according to the spatial distribution density of the decoupling capacitors. When the decoupling capacitor distribution density in the area is higher than the average density, the coupling attenuation value of the area increases. When the decoupling capacitor distribution density in the area is lower than the average density, the coupling attenuation value of the area decreases. Finally, the coupling attenuation value of the grid area is weighted averaged with the coupling attenuation value of the adjacent grids to obtain the power supply coupling suppression factor. When the grid area is located at the intersection of signal lines, the larger value of the coupling attenuation value of the adjacent grids is taken as the power supply coupling suppression factor of the area.

[0087] When calculating the skin effect correction factor based on the change in effective cross-sectional area, a standard cross-sectional area baseline value for the signal line is first established. The standard cross-sectional area baseline value is equal to the product of the signal line width and thickness. Then, the effective cross-sectional area after accounting for the skin effect is obtained, and the ratio of the effective cross-sectional area to the standard cross-sectional area baseline value is used as the cross-sectional area change rate. Next, a corresponding relationship between the cross-sectional area change rate and current distribution is established. When the cross-sectional area change rate is less than a preset threshold, it indicates that the current is mainly concentrated on the surface of the signal line, and the skin effect is significant. Subsequently, the signal line is divided into regions according to its burial depth. The burial depth refers to the distance from the center of the signal line to the reference plane. The greater the burial depth, the more significant the impact of the skin effect on the impedance. Furthermore, the cross-sectional area change rate and the burial depth are combined for calculation. When the signal frequency is the same, the region with a larger burial depth has a larger correction amplitude. When the burial depth is the same, the signal with a higher frequency has a larger correction amplitude. Finally, based on the combined calculation results, the skin effect correction factor is obtained. When the signal line is located at a corner, the correction factor needs to take into account the current density distortion at the corner.

[0088] On the other hand, when the interlayer coupling coefficient is used to correct the dynamic impedance matching matrix, a spatial mapping relationship is first established, and the interlayer coupling coefficient measurement point on the PCB board is positionally matched with the impedance detection point in the dynamic impedance matching matrix, and the position correspondence includes horizontal position correspondence and vertical position correspondence; when the measurement point and the impedance detection point do not coincide in spatial position, a local area grid is established with the measurement point as the center, and the interlayer coupling coefficient in the area is interpolated using the area weight method. The area weight method determines the interpolation weight according to the distance from the target point to the surrounding measurement points; then, a correction calculation is performed on each impedance detection point in the dynamic impedance matching matrix. The correction calculation first determines the equivalent coupling strength of the point, and the equivalent coupling strength The original impedance value is obtained by weighting the interlayer coupling coefficients directly above and below the point according to the ratio of vertical distance. Next, the original impedance value is corrected according to the equivalent coupling strength. When the equivalent coupling strength is greater than the preset reference value, the original impedance value is attenuated according to the ratio of the coupling strength, and the attenuation process adopts a piecewise linear mapping method. When the equivalent coupling strength is less than the preset reference value, the original impedance value is compensated according to the ratio of the coupling strength, and the compensation process also adopts a piecewise linear mapping method. Subsequently, the corrected impedance value is checked for continuity in the horizontal direction, and the impedance difference between adjacent detection points is used as the judgment basis. When the impedance difference exceeds the preset change rate, a transition point is added between the two detection points, and the impedance value of the transition point is determined by linear interpolation.

[0089] Furthermore, the corrected impedance value is checked for continuity in the vertical direction using the same inspection and correction methods as those used in the horizontal direction. The corrected impedance distribution is then optimized along the signal transmission path. The optimization process first identifies the starting and ending points of signal transmission, and divides the transmission path between these two points into several segments. The allowable rate of change of the impedance value within each segment is determined by the signal rise time. Finally, the corrected and optimized impedance values ​​are re-entered into the dynamic impedance matching matrix in the order of the detection points to generate a compensation matrix that takes into account the interlayer coupling effect. When the error between the correction result and the target indicator exceeds a preset range, the above correction steps need to be repeated, and the equivalent coupling strength and impedance correction value need to be recalculated each time the steps are repeated.

[0090] It should be noted that the piecewise linear mapping method of attenuation processing means: when the equivalent coupling strength is greater than the preset reference value, the variation interval of the equivalent coupling strength is divided into multiple sub-intervals, each sub-interval corresponds to a different attenuation proportional coefficient, and the attenuation proportional coefficient increases with the increase of the equivalent coupling strength; within each sub-interval, there is a linear relationship between the attenuation proportional coefficient and the equivalent coupling strength, and the attenuation proportional coefficients of adjacent sub-intervals remain continuous at the dividing point; the original impedance value is multiplied by the corresponding attenuation proportional coefficient to obtain the attenuated impedance value; when the equivalent coupling strength is at the dividing point of the sub-interval, the average value of the attenuation proportional coefficients of the previous and next sub-intervals is used for calculation.

[0091] The piecewise linear mapping method of compensation processing means: when the equivalent coupling strength is less than a preset reference value, the variation interval of the equivalent coupling strength is divided into multiple sub-intervals, each sub-interval corresponds to a different compensation proportional coefficient, and the compensation proportional coefficient increases as the equivalent coupling strength decreases; within each sub-interval, there is a linear relationship between the compensation proportional coefficient and the equivalent coupling strength, and the compensation proportional coefficients of adjacent sub-intervals remain continuous at the dividing point; the original impedance value is multiplied by the corresponding compensation proportional coefficient to obtain the compensated impedance value; when the equivalent coupling strength is at the sub-interval dividing point, the average value of the compensation proportional coefficients of the previous and next sub-intervals is used for calculation.

[0092] S4: Based on the compensation matrix that takes into account the interlayer coupling effect, real-time impedance matching is performed on the DDR4 signal transmission channel through a programmable impedance network, and compensation parameters of adjacent impedance detection points are used for interpolation operation to achieve continuous impedance matching throughout the signal transmission channel.

[0093] When real-time impedance matching is performed on the DDR4 signal transmission channel through a programmable impedance network, the compensation parameters in the compensation matrix that takes into account the interlayer coupling effect are first written into the configuration register group of the programmable impedance network in address order. The configuration register group includes an impedance control register and a status register. The impedance control register is used to store the target impedance value, and the status register is used to feedback the impedance adjustment state. Then, the control unit reads the compensation parameters in the configuration register group and calculates the resistance adjustment amount of each variable impedance unit according to the compensation parameters. The variable impedance unit adopts a digitally controlled adjustable resistor array structure. Then, the calculated resistance adjustment amount is converted into an impedance control code, and the control code is written into the control register of the variable impedance unit in sequence through the serial data line. When the writing is completed, all variable impedance units are synchronously switched to the new resistance value. Subsequently, a high-speed sampling circuit is used to perform impedance sampling on the key points on the signal transmission channel. The sampling rate of the sampling circuit is not less than four times the signal frequency. , the sampling data is sent to the impedance calculation module after digital filtering; when the deviation between the actual impedance value obtained by sampling and the target impedance value exceeds the set range, the control unit immediately starts the impedance adjustment process, and the adjustment process adopts a step-by-step approximation method, and the step value of each adjustment is determined by the size of the deviation; if the signal quality index is detected to be deteriorated during the adjustment process, it will automatically fall back to the last stable state and re-adjust in small steps; when the external working conditions change, such as when the temperature changes exceed the preset range or the signal frequency is adjusted, it is necessary to reload the compensation parameters under the corresponding working conditions and re-perform the impedance matching process according to the above steps; finally, the impedance matching effect is verified by real-time monitoring of signal integrity indicators, which include signal rise time, overshoot amplitude and jitter parameters. When these indicators meet the design requirements, the current impedance matching state is locked; if the indicators do not meet the requirements, the compensation parameter calculation and impedance adjustment are required to be re-performed until the optimal matching state is reached, such as Figure 2 shown.

[0094] When using the compensation parameters of adjacent impedance detection points for interpolation calculation, first identify adjacent impedance detection point pairs on the transmission line, and use the physical area between each pair of detection points as the interpolation calculation interval. The selection of the detection point pairs is based on the signal transmission direction for directional identification. When the transmission line has a branch structure, each branch needs to be divided into detection point pairs separately. Then, the compensation parameter values ​​of the detection point pairs are read from the compensation matrix, and the sub-interval division interval is determined according to the signal transmission characteristics. The sub-interval division interval is no more than one tenth of the signal wavelength, and the boundary position of each sub-interval is recorded as an interpolation node. Then, local interpolation parameter calculation is performed on each sub-interval. The calculation process adopts a distance-based weighted method. Specifically, when calculating the parameters of a certain interpolation node, the ratio of the distance from the node to the two detection points is used as the weight coefficient, and the detection point with a closer distance obtains a larger weight value.

[0095] Subsequently, the interpolation parameters are smoothed using a five-point sliding average method. The smoothing process progresses gradually from the starting end to the ending end of the interpolation interval. When the parameter change rate between adjacent interpolation nodes exceeds a preset threshold, the smoothing window is automatically expanded to a seven-point smoothing. When the interpolation interval contains a bend in the transmission line, three transition interpolation nodes are added at the bend position. The parameter calculation of the transition interpolation nodes requires the bend angle to be used as a correction factor. Furthermore, a monotonicity analysis is performed on the parameter sequences of all interpolation nodes. The analysis process includes calculating the parameter difference and change trend of adjacent nodes. When parameter fluctuations occur, the weight distribution of the interpolation nodes in this area needs to be readjusted. Finally, the smoothed interpolation parameter sequence is written into the data buffer of the programmable impedance network in address order. The writing process adopts a double buffering mechanism. When a data write error is detected, a data check and retransmission mechanism is activated to ensure the accurate loading of the interpolation parameters. If the interpolation result of a certain area still does not meet the smooth transition requirement, the interpolation calculation process for that area needs to be repeated until a parameter distribution that meets the impedance continuous matching requirement is obtained.

[0096] In summary, a method for improving the signal transmission stability of a functional board based on an embodiment of the present invention is illustrated. The method presets impedance detection points at key positions in the DDR4 signal transmission channel, collects characteristic impedance values ​​and crosstalk coefficients, dynamically calculates compensation parameters in combination with microstrip line transmission theory, generates a dynamic impedance matching matrix, and introduces interlayer coupling coefficients to correct the compensation matrix. Finally, real-time impedance matching and continuous impedance matching of the entire signal transmission channel are achieved through a programmable impedance network. In this way, the signal reflection and crosstalk problems caused by impedance mismatch during long-distance transmission of DDR4 high-speed memory signals in PCB functional boards can be solved to a certain extent, effectively improving the integrity and stability of signal transmission, reducing the incidence of signal reflection, overshoot, and jitter, and significantly improving signal transmission performance, especially in complex multi-layer PCB designs and dynamic impedance environments, meeting the reliability requirements of high-speed signal transmission.

[0097] Here, those skilled in the art will appreciate that the specific operations of each step in the above method for improving the stability of signal transmission of the functional board have been described in detail above. Figure 1 and Figure 2 The method for improving the signal transmission stability of the functional board has been described in detail, and therefore, its repeated description will be omitted.

Claims

1. A method for improving the signal transmission stability of a function board, characterized in that: include: Preset multiple impedance detection points at key locations of the DDR4 signal transmission channel and collect characteristic impedance values ​​and crosstalk coefficients of the impedance detection points; Based on the characteristic impedance value and the crosstalk coefficient, impedance compensation parameters are calculated in combination with microstrip line transmission theory to generate a dynamic impedance matching matrix; Measuring parasitic capacitance values ​​between PCB layers, calculating an interlayer coupling coefficient based on the parasitic capacitance values, and using the interlayer coupling coefficient to correct the dynamic impedance matching matrix to obtain a compensation matrix that takes into account the interlayer coupling effect; According to the compensation matrix that takes into account the interlayer coupling effect, real-time impedance matching is performed on the DDR4 signal transmission channel through a programmable impedance network, and interpolation operation is performed using the compensation parameters of adjacent impedance detection points to achieve continuous impedance matching throughout the signal transmission channel.

2. The method for improving the signal transmission stability of a functional board according to claim 1, characterized in that: The impedance detection point adopts a through-hole embedded test probe structure; The impedance detection points are arranged alternately on both sides of the differential pair signal lines, that is, a first detection point is set on the positive line side of the differential signal line, and a second detection point is set on the negative line side at a preset distance from the first detection point, and the remaining detection points are set in this way.

3. The method for improving the signal transmission stability of a functional board according to claim 1, characterized in that: Generating the dynamic impedance matching matrix includes: An impedance calculation model is constructed based on the frequency characteristics and material parameters of DDR4 high-speed signal transmission to obtain the actual measured characteristic impedance value. Calculate the impedance deviation value of the monitoring point and calculate the equivalent impedance correction value using the dynamic coupling strength evaluation function, and calculate the total compensation value; The equivalent impedance correction amount and the total compensation value are iteratively optimized to obtain the dynamic impedance matching matrix.

4. The method for improving the signal transmission stability of a functional board according to claim 3, characterized in that: The calculation of the equivalent impedance correction is shown in the following formula: in, K coupling (d)=k0e -γd (1+δsin(ωt)) Among them, Z correction is the equivalent impedance correction, L is the signal line self-inductance, M is the mutual inductance coefficient, C is the signal line self-capacitance, C m is the mutual capacitance coefficient, K coupling (d) is the coupling strength coefficient, d is the distance between adjacent signal lines, k0 is the reference coupling coefficient, γ is the spatial attenuation factor, δ is the time-varying modulation coefficient, ω is the angular frequency, and t is the time variable.

5. The method for improving the signal transmission stability of a functional board according to claim 3, characterized in that: The dynamic impedance matching matrix is ​​shown below: Among them, P ij is the element in row i and column j of the compensation matrix, η is the iterative learning rate, Z comp,i is the total compensation value of the i-th detection point, Z correction,m is the equivalent impedance correction value of the mth detection point, and k is the number of iterations.

6. The method for improving the signal transmission stability of a functional board according to claim 1, characterized in that: Calculating the interlayer coupling coefficient includes: Mark the strong coupling point through the local detection window to determine whether it is located in the signal line intersection area and determine the angle correction factor; With the strong coupling point as the center, the electric field gradient is calculated and the high coupling area is marked; Calculating the regional coupling coefficient of the high coupling area, and determining whether to use the skin effect correction factor or the power supply coupling suppression factor for correction according to the detection area; Multiplying the angle correction factor, the power supply coupling suppression factor, and the skin effect correction factor to obtain a comprehensive correction coefficient; The interlayer coupling coefficient is obtained by multiplying the comprehensive correction coefficient by the regional coupling coefficient.

7. The method for improving the signal transmission stability of a functional board according to claim 1, characterized in that: Correcting the dynamic impedance matching matrix using the interlayer coupling coefficient includes: The interlayer coupling coefficient measurement points are aligned with the impedance detection points in the dynamic impedance matching matrix to establish a spatial mapping relationship; Perform correction calculation on each impedance detection point in the dynamic impedance matching matrix; Conduct continuity checks on the corrected impedance values ​​in the vertical direction and optimize the corrected impedance distribution along the signal transmission path; The corrected and optimized impedance values ​​are re-filled into the dynamic impedance matching matrix according to the position sequence of the detection points to generate a compensation matrix that takes into account the interlayer coupling effect.

8. The method for improving the signal transmission stability of a functional board according to claim 7, characterized in that: include: The correction calculation includes: Determine the equivalent coupling strength of each impedance detection point; The original impedance value is corrected according to the equivalent coupling strength. When the equivalent coupling strength is greater than or less than the preset reference value, attenuation or compensation processing is performed respectively; the attenuation or compensation processing adopts a piecewise linear mapping method.

9. The method for improving the signal transmission stability of a functional board according to claim 1, characterized in that: By writing the compensation matrix into the programmable impedance network, the control unit calculates and adjusts the resistance value of the impedance unit, combining high-speed sampling and step-by-step approximation adjustment methods to achieve real-time impedance matching of the DDR4 signal transmission channel.

10. The method for improving the signal transmission stability of a functional board according to claim 1, characterized in that: The interpolation operation identifies adjacent detection point pairs and divides the interpolation interval, calculates the interpolation parameters based on the distance weighting method, and generates a parameter distribution that meets the impedance continuous matching requirement.

Images