A dual-layer substrate circuit module employing non-sealed hybrid high integration

Abstract

This invention relates to the field of integrated circuit technology, specifically to a non-sealed, hybrid, highly integrated dual-layer substrate circuit module, comprising: an upper substrate and a lower substrate, with components soldered onto both substrates. The upper substrate serves as a pulse signal layer, and the lower substrate as a power layer, with the circuit printing thickness of the pulse signal layer being less than that of the power layer. The layout design of the components on the lower substrate is updated based on the interference of the temperature distribution of the lower substrate on the pulse signal in the upper substrate during circuit module operation. This invention further solves the problem of the difficulty in integrating high-frequency pulse signals and high-power electrical signals, improving the stability of the circuit module.

Description

Technical Field

[0001] This invention relates to the field of integrated circuit technology, and more specifically to a dual-layer substrate circuit module employing non-sealed hybrid high integration. Background Technology

[0002] In fields such as power electronics and new energy control (e.g., substation battery management, motor drives), integrated circuits need to simultaneously handle high-power electrical signals and high-frequency pulse signals (e.g., MHz-level PWM control, high-speed communication clocks). Traditional single-layer substrate architectures, due to the mixing of power devices and high-frequency circuits, require the simultaneous placement of power devices (e.g., IGBTs) and high-frequency chips (e.g., FPGAs) on a single substrate, resulting in low wiring density and difficulty in miniaturizing circuit modules. Furthermore, power devices generate significant heat, requiring large-area heat dissipation, while high-frequency pulse signals are sensitive to thermal noise, further increasing the difficulty of circuit module layout and design. Summary of the Invention

[0003] To address the aforementioned problem of integrating high-frequency pulse signals and high-power electrical signals, this invention provides a dual-layer substrate circuit module employing a non-sealed, hybrid, highly integrated design.

[0004] The present invention provides a non-sealed, hybrid, highly integrated dual-layer substrate circuit module using the following technical solution:

[0005] One embodiment of the present invention provides a dual-layer substrate circuit module with non-sealed hybrid high integration. The circuit module includes an upper substrate and a lower substrate, on which components are soldered respectively. The upper substrate serves as a pulse signal layer, and the lower substrate serves as a power layer. The circuit printing thickness of the pulse signal layer is less than that of the power layer.

[0006] Based on the interference of the temperature distribution of the lower substrate on the pulse signal in the upper substrate during the operation of the circuit module, the layout design of the components in the lower substrate is updated. The specific steps include the following:

[0007] The temperature change curve of each region on the lower substrate and the pulse error of the pulse signal line in the upper substrate are obtained under each operating condition of the circuit module. All regions passed under the pulse signal line are recorded as target regions and any target region is recorded as reference regions. Under the same operating condition, the difference between the temperature change curve of the reference region and the temperature change curves of all target regions outside the reference region is recorded as the temperature distribution difference of the reference region under the same operating condition.

[0008] Under the condition where the distribution difference in the reference area is the greatest, the amplitude of the pulse error change caused by the temperature change curve of the reference area is used as the interference coefficient of the reference area.

[0009] Under the condition where the temperature change curves of all target areas have the smallest differences, the estimation error of the pulse signal line is obtained based on the interference coefficient of all target areas. The estimation error is used to correct the interference coefficient of all target areas so that the difference between the estimation error obtained by the corrected interference coefficient and the change trend of the pulse error is minimized. The corrected interference coefficient of all target areas under all conditions is used to assist in updating the layout design of components in the lower substrate.

[0010] Preferably, the specific steps for correcting the interference coefficients of all target regions using the estimation error, so that the difference between the estimation error and the change trend of the impulse error obtained by the corrected interference coefficients is minimized, are as follows:

[0011] D1: The estimation error obtained for the interference coefficient of all target areas, the difference between the trend of the estimation error and the impulse error is recorded as the interference coefficient correction amount;

[0012] D2: Under the condition of greatest distribution difference in the reference area, on the temperature change curve of each target area, the product of the temperature difference at adjacent time points and the interference coefficient of each target area is denoted as the prediction pulse error of each target area at any two adjacent time points; the mean of the prediction pulse errors of each target area at all adjacent time points is denoted as the second mean of each target area; the ratio of the second mean of the reference area to the sum of the second means of all target areas is denoted as the correction ratio y; the interference coefficient of the reference area after correction is positively correlated with the correction coefficient; the correction coefficient is positively correlated with y and negatively correlated with the amount of interference coefficient correction;

[0013] D3: After D2 is completed, D1 is re-executed using the corrected interference coefficients for each target region. This includes: the estimation error obtained for the corrected interference coefficients of all target regions, and the difference between the estimation error and the change trend of the impulse error is recorded as the interference coefficient correction amount.

[0014] Then, the process of executing D2 and D1 is recorded as one elimination process. After repeating the elimination process several times, the corrected interference coefficient obtained when the interference coefficient correction reaches the minimum value is taken as the final interference coefficient of each target area. For the final interference coefficient of all target areas, the difference between the estimation error and the change trend of the impulse error obtained by the final interference coefficient is minimized.

[0015] Preferably, after each elimination process, for any target region, among all the interference coefficients obtained for that target region, the interference coefficients with the largest reduction are selected to form the first interference coefficient sequence for that target region; all the interference coefficients in the first interference coefficient sequence of that target region correspond to the obtained interference coefficient correction amounts to form the first correction amount sequence; the first interference coefficient sequence and the first correction amount sequence are linearly normalized respectively, and the Pearson correlation coefficient between the normalized first interference coefficient sequence and the normalized first correction amount sequence is obtained, which is recorded as the appropriate elimination coefficient for that target region after each elimination process;

[0016] When the elimination suitability coefficient of each target region is less than or equal to the second preset threshold, the elimination process of each target region is suppressed during the next elimination process, specifically including:

[0017] Obtain the suppression correction coefficient, which is positively correlated with the correction coefficient obtained for each target region and negatively correlated with the elimination appropriate coefficient for each target region; in the next elimination process, the corrected interference coefficient of each target region is positively correlated with the suppression correction coefficient.

[0018] Preferably, the specific steps for determining the difference in the changing trends of the estimation error and the pulse error are as follows:

[0019] The estimation error obtained for the interference coefficients of all target areas, or the estimation error obtained for the corrected interference coefficients of all target areas, represents the estimation error of the pulse signal at any adjacent time point.

[0020] The estimation errors of the pulse signals at all adjacent time points constitute the estimation error variation curve; the first-order difference result of the pulse error of the pulse signal line at all time points is denoted as the measurement error variation curve; the estimation error variation curve and the measurement error variation curve are respectively linearly normalized, and the Pearson correlation coefficient between the normalized estimation error variation curve and the measurement error variation curve is denoted as x. The difference in the variation trend between the estimation error and the pulse error is negatively correlated with x.

[0021] Preferably, the specific steps for obtaining the estimation error are as follows:

[0022] For the interference coefficients of all target areas or the corrected interference coefficients of all target areas; under the condition that the temperature change curves of all target areas have the smallest difference, for the temperature change curve of any target area, and for any two adjacent time points on the temperature change curve, obtain the temperature difference between any two adjacent time points, and the product of the temperature difference and the interference coefficient of the target area is recorded as the first product of any target area at any two adjacent time points; the sum of the first products of all target areas at any two adjacent time points is recorded as the estimation error of the pulse signal at any two adjacent time points.

[0023] Preferably, the amplitude of the pulse error change caused by the temperature change curve of the reference region is used as the interference coefficient of the reference region, and the specific steps include the following:

[0024] For the temperature change curve of the reference region, the temperature difference between any two adjacent time points on the temperature change curve is represented as C1;

[0025] The difference in pulse error between adjacent time points in the pulse error variation curve is denoted as C2. C2 / (C1+c0) is denoted as the error growth rate between any two adjacent time points, where c0 is a preset data. The average error growth rate of all adjacent time points in the temperature variation curve of the reference area is denoted as the interference coefficient of the reference area.

[0026] Preferably, the difference between the temperature change curve of the reference area and the temperature change curves of all target areas outside the reference area is recorded as the temperature distribution difference of the reference area under the same operating condition, and the specific steps include the following:

[0027] The product of the average slope and average temperature of each temperature change curve is recorded as the evaluation index of each temperature change curve. The mean of the evaluation indexes of the temperature change curves of all target areas outside the reference area is recorded as the first mean. The difference between the evaluation index of the temperature change curve of the reference area and the first mean is recorded as the temperature distribution difference of the reference area.

[0028] Preferably, the specific steps for obtaining the condition where the temperature change curves of all target areas have the smallest difference are as follows:

[0029] Under any operating condition, for the temperature distribution difference of each target area, obtain the variance of the temperature distribution difference of all target areas, and record it as the temperature distribution characteristic under each operating condition. Obtain the operating condition with the smallest temperature distribution characteristic, under which the temperature change curve difference of all target areas is minimized.

[0030] Preferably, the specific steps of using the corrected interference coefficients of all target regions under all operating conditions to assist in updating the layout design of components in the lower substrate are as follows:

[0031] The corrected interference coefficient obtained when the difference between the estimation error and the pulse error is minimal is recorded as the final interference coefficient; under each operating condition, for all target areas traversed by each pulse signal line, each target area corresponds to a final interference coefficient; the target area with the largest final interference coefficient is recorded as the area of ​​interest for each pulse signal line under each operating condition; the sum of the interference coefficients of all target areas outside the area of ​​interest is recorded as the interference threshold for each pulse signal line under each operating condition.

[0032] Under all operating conditions, all components within all areas of interest corresponding to each pulse signal line are denoted as the components of each pulse signal line, and the average value of all interference thresholds corresponding to each pulse signal line is denoted as the warning threshold of each pulse signal line.

[0033] For any component below any pulse signal line, obtain the area where the component is located, and obtain the average value of the interference coefficient of the area under all operating conditions, which is recorded as the interference coefficient of the component.

[0034] When rearranging components in the lower substrate, an early warning is issued if a component is located below each pulse signal line, or if the sum of the interference coefficients of all components below each pulse signal line is greater than or equal to the corresponding warning threshold.

[0035] Preferably, the upper substrate and the lower substrate are vertically stacked by support pillars. Both the upper substrate and the lower substrate contain pins, which are connected by soldering copper wires to enable electrical signal communication between the upper substrate and the lower substrate. An interface is installed on the lower substrate to supply power to the entire circuit module and input data corresponding to each operating condition.

[0036] The beneficial effects of the technical solution of the present invention are:

[0037] The upper substrate serves as the pulse signal layer, and the lower substrate serves as the power layer. The circuit printing thickness of the pulse signal layer is less than that of the power layer. The circuit module of this invention vertically separates the power layer and the pulse signal layer, placing them on different substrates. Furthermore, the circuit printing thickness of the pulse signal layer is less than that of the power layer, ensuring isolation between high-frequency signals and low-frequency power supplies. This significantly improves the power density of the circuit module and enhances its output capability for high-frequency pulse signals, thereby improving the overall performance of the circuit module.

[0038] Furthermore, based on the interference of the temperature distribution of the lower substrate on the pulse signal in the upper substrate during circuit module operation, the layout design of components in the lower substrate is updated. This process avoids the problem of a large amount of heat generated by the lower substrate during operation being blocked by the upper substrate and unable to dissipate directly into the air, causing heat accumulation between the lower and upper substrates and significantly interfering with the high-frequency signal pulse in the upper substrate, thus improving the stability of the circuit module.

[0039] Specifically, under the condition where the distribution difference in the reference area is the greatest, the amplitude of the pulse error change caused by the temperature change curve of the reference area is used as the interference coefficient of the reference area; under the condition where the temperature change curve difference in all target areas is the smallest, the estimation error of the pulse signal line is obtained based on the interference coefficient of all target areas, and the interference coefficient of all target areas is corrected using the estimation error, so that the difference between the estimation error obtained by the corrected interference coefficient and the change trend of the pulse error is minimized; the corrected interference coefficient of all target areas under all conditions is used to assist in updating the layout design of components in the lower substrate. This process utilizes the interference coefficients of other target regions outside the reference region to correct the interference coefficient of the reference region, gradually eliminating the interference of other target regions outside the reference region on the pulse signal, rather than directly ignoring the interference of other target regions outside the reference region on the pulse signal. At the same time, when multiple target regions interfere with the pulse signal simultaneously, it can obtain the interference of the temperature rise of a single target region on the pulse signal as accurately as possible. It also allows the interference coefficients of multiple target regions to be combined to jointly evaluate the change in the interference trend of the temperature increase of multiple target regions on the pulse signal line. This ensures that when performing auxiliary layout design, it can consider both the impact of heat generation from a single target region on the layout and the interference of multiple target regions generating heat simultaneously on the same pulse signal line. To a certain extent, it solves the problem of the difficulty in integrating high-frequency pulse signals and high-power electrical signals, and improves the stability of the circuit module. Attached Figure Description

[0040] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0041] Figure 1 An overall structural diagram of a dual-layer substrate circuit module employing non-sealed hybrid high integration is provided in one embodiment of the present invention;

[0042] Figure 2A top view of the lower substrate of a dual-layer substrate circuit module employing a non-sealed hybrid high-integration design, provided as an embodiment of the present invention;

[0043] Figure 3 A bottom view of the upper substrate of a dual-layer substrate circuit module employing a non-sealed hybrid high-integration design, provided as an embodiment of the present invention;

[0044] Figure 4 A flowchart illustrating the steps for updating the component layout design in the lower substrate according to an embodiment of the present invention. Detailed Implementation

[0045] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a non-sealed hybrid highly integrated double-layer substrate circuit module proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0046] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0047] The following description, in conjunction with the accompanying drawings, details a specific solution for a non-sealed, hybrid, highly integrated dual-layer substrate circuit module provided by the present invention.

[0048] Example 1:

[0049] This embodiment provides a non-sealed, hybrid, highly integrated dual-layer substrate circuit module, comprising an upper substrate 2 and a lower substrate 1, wherein components 3 are soldered onto the upper substrate 2 and the lower substrate 1, such as... Figures 1 to 3 As shown. Both the upper substrate 2 and the lower substrate 1 are single-layer boards, and their material is HL832NS organic core board material. In other embodiments, ceramic materials such as silicon nitride can also be used.

[0050] The circuit module in this embodiment is used for power control, such as grid connection control of substation batteries and power control of large motors. During operation, it needs to process both high-power electrical signals (e.g., current signals during power charging and discharging) and high-frequency pulse signals (e.g., PWM control pulses, clock signals during high-speed serial communication). In this embodiment, when designing the component layout of the upper substrate 2 and the lower substrate 1, components or chips that process high-power electrical signals (e.g., IGBTs, thyristors, optocoupler isolation chips) are placed on the lower substrate 1 as a power layer; components or chips that process or generate high-frequency pulse signals (e.g., high-speed switching transistors, high-speed ADCs / DACs) are placed on the upper substrate 2 as a pulse signal layer.

[0051] Furthermore, the upper substrate 2 and the lower substrate 1 are vertically stacked via support pillars 4. The support pillars 4 are made of ceramic and are glued to the upper substrate 2 and the lower substrate 1 at both ends. Both the upper substrate 2 and the lower substrate 1 contain pins 5, which are connected by copper wire to achieve electrical signal communication between the upper substrate 2 and the lower substrate 1. It should be noted that the stacking method of the upper substrate 2 and the lower substrate 1 is a known technology. For example, the multi-layered hybrid integrated circuit module disclosed in CN218772646U can be used to implement the stacking of the upper substrate 2 and the lower substrate 1 in this embodiment.

[0052] An interface 6 is mounted on the lower substrate 1 for supplying power to the entire circuit module and inputting external data, such as the voltage and current of the battery.

[0053] The lower substrate 1 (i.e., the power layer) employs thick-film circuit printing, such as screen printing to lay a copper circuit layer (i.e., a copper layer); the upper substrate 2 (i.e., the pulse signal layer) employs thin-film circuit printing, such as atomic layer deposition to lay a copper circuit layer (i.e., a copper layer). In this embodiment, the circuit printing thickness (i.e., the copper layer thickness) of the pulse signal layer is 0.5 μm, preferably ranging from 10 nm to 1 μm; the circuit printing thickness (i.e., the copper layer thickness) of the power layer is 40 μm, preferably ranging from 2 μm to 50 μm.

[0054] In this embodiment, the circuit module vertically separates the power layer and the pulse signal layer, placing them on different substrates. The circuit printing thickness of the pulse signal layer is smaller than that of the power layer, ensuring isolation between high-frequency signals and low-frequency power supplies. This significantly improves the power density of the circuit module and enhances its output capability for high-frequency pulse signals, thereby improving the overall performance of the circuit module.

[0055] Example 2:

[0056] This embodiment considers the following problems that arise when the power layer and pulse signal layer of the circuit module in Embodiment 1 are vertically layered: The lower substrate 1, serving as the power layer, generates a large amount of heat during operation. This heat cannot dissipate directly into the air due to the obstruction of the upper substrate 2, which serves as the pulse signal layer. This causes heat to accumulate between the lower substrate 1 and the upper substrate 2, significantly interfering with the high-frequency pulse signal in the upper substrate 2. This interference alters the resistivity of the conductor, causes dielectric constant drift, and ultimately leads to errors in the high-frequency pulse signal of the upper substrate 2 (e.g., delayed or jittered rise or fall edges). Furthermore, the heat generated at different locations on the lower substrate 1 varies depending on the operating conditions of the circuit module (or when switching between different operating conditions), resulting in a non-fixed impact on the pulse signal in the upper substrate 2. Ultimately, this leads to unstable operation of the upper substrate 2 (i.e., the pulse signal layer), or its unpredictable and uncertain operating state.

[0057] This embodiment updates the layout design of components 3 in the lower substrate 1 based on the interference of the temperature distribution of the lower substrate 1 on the pulse signal in the upper substrate 2 during circuit module operation. By updating the layout design of components 3 in the lower substrate 1, the aforementioned problems or their impacts are minimized. Figure 4 As shown, the specific steps are as follows:

[0058] Step S201: Obtain the temperature change curve of each region on the lower substrate 1 under each operating condition of the circuit module, and the pulse error of the pulse signal line in the upper substrate 2.

[0059] In this embodiment, the lower substrate 1 is divided into several regions, and each component 3 on the lower substrate 1 is considered as one region.

[0060] The operating conditions refer to the input data of the circuit module's interface 6 during operation. For example, when the circuit module is used for power control, it needs to be connected to the positive and negative terminals of each battery pack and to the charger to control the charging and discharging of the battery packs in the power grid. The input data includes the voltage or current output by each battery pack or charger. Different input data during operation lead to different operating conditions and heating patterns for components 3 within the circuit module. This embodiment requires collecting temperature change curves for each region under each operating condition (i.e., under each input data). These temperature change curves refer to the average temperature change curve of each region over a preset time period under each operating condition (i.e., under each input data). They describe the temperature rise in each region when the circuit module is operating under each condition, with the horizontal axis representing time and the vertical axis representing temperature magnitude.

[0061] Simultaneously, the pulse error variation curve of each pulse signal line in the upper substrate 2 is collected within a preset time period. The pulse error variation curve represents the change curve of the pulse signal error in the pulse signal line over time. The pulse signal error (abbreviated as pulse error) refers to the difference between the pulse signal in the pulse signal line at each time point and the pulse signal in the pulse signal line at the initial time (i.e., at the first time point of the preset time period). The larger the pulse error, the more obvious the deformation of the pulse signal in the pulse signal line (including the delay of the rising or falling edge, and the jitter of the rising or falling edge).

[0062] Thus, the temperature change curve for each region under each operating condition, as well as the pulse error change curve composed of the pulse error on each pulse signal line, are obtained.

[0063] Step S202: Record all areas passed under the pulse signal line as target areas, and any target area as a reference area; under the same operating condition, the difference between the temperature change curve of the reference area and the temperature change curves of all target areas outside the reference area is recorded as the temperature distribution difference of the reference area under the same operating condition.

[0064] The entire area traversed by each pulse signal line is designated as the target area. The deformation of the pulse signal of the pulse signal line is mainly affected by the heat generated in the target area.

[0065] Specifically, if the difference between the maximum and minimum temperatures in the temperature change curve of any target area under any operating condition is less than 10℃, then the target area will not be considered in subsequent analyses under that operating condition (that is, the target area is considered to not exist or will no longer be considered as the target area under that operating condition).

[0066] Any target area is designated as the reference area. For the temperature change curves of all target areas under the same operating condition, the difference between the temperature change curve of the reference area and the temperature change curves of all target areas outside the reference area (referred to as the temperature distribution difference of the reference area under the same operating condition) describes the difference between the temperature rise of the reference area and the temperature rise of all other target areas. The larger the value, the greater the difference between the temperature rise of the reference area and the temperature rise of all other target areas; the smaller the value, the more similar or identical the temperature rise of the reference area to the temperature rise of all other target areas.

[0067] As an alternative example, methods for obtaining the temperature distribution differences in a reference area under the same operating conditions include:

[0068] The product of the average slope and the average temperature of each temperature change curve is recorded as the evaluation index for each temperature change curve.

[0069] The average value of the evaluation index of the temperature change curves of all target areas outside the reference area is recorded as the first average value, and the difference between the evaluation index of the temperature change curve of the reference area and the first average value is recorded as the temperature distribution difference of the reference area.

[0070] As another example, methods for obtaining the temperature distribution differences in a reference area under the same operating conditions include:

[0071] In this preferred example, for each time point on each temperature change curve, a time window of length K1 is constructed centered on each time point. In this embodiment, K1 is one-tenth (rounded up) of the total number of time points on each temperature change curve. Specifically, if K1 is not odd, then K1 is incremented by one to obtain the length of the time window. In other embodiments, the preferred value of K1 is between one-tenth and one-fifth of the total number of time points on each temperature change curve.

[0072] Obtain the temperature sequence of all time points on each temperature change curve within each time window, and obtain the values ​​of multiple indicators of the temperature sequence, specifically including average temperature and average slope.

[0073] Within the same time window, the average value of the temperature change curves of all target areas outside the reference area for the same indicator is obtained and denoted as the first mean of each indicator. The difference between the temperature change curve of the reference area for each indicator and the first mean of each indicator is calculated, and the ratio of this difference to the first mean of each indicator is denoted as the first difference of each indicator. The purpose of using the first mean of each indicator as the denominator to calculate the ratio is to remove the dimensions and order of magnitude of the first difference.

[0074] For all time windows, multiple first differences can be obtained for each indicator, and the largest first difference is recorded as the reference difference for each indicator.

[0075] The maximum value of the reference difference of all indicators is recorded as the temperature distribution difference of the reference area, and the time window corresponding to this maximum value is recorded as the reference time period of the temperature distribution difference.

[0076] This example demonstrates the most significant temperature differences between the reference region and other target regions (i.e., target regions outside the reference region) within a local timeframe.

[0077] Thus, for each target area, a temperature distribution difference has been obtained using the above method.

[0078] Step S203: Under the condition of the greatest distribution difference in the reference area, the amplitude of the pulse error change caused by the temperature change curve of the reference area is used as the interference coefficient of the reference area.

[0079] The above describes the temperature distribution differences of the reference area under each operating condition. The operating condition with the largest temperature distribution difference in the reference area is identified, meaning that the temperature change in the reference area differs most significantly from the temperature changes in other target areas under this condition. In this case, the deformation of the pulse signal in the upper substrate 2 is mainly affected by the reference area. Therefore, this embodiment ignores the influence of temperature changes in other target areas on the pulse signal line and uses the amplitude of the pulse error change caused by the temperature change curve of the reference area as the interference coefficient of the reference area. A larger interference coefficient indicates that the temperature change in the reference area can significantly cause changes in the pulse error on the pulse signal line.

[0080] As an example, the amplitude of the pulse error change caused by the temperature change curve of the reference area is used as the interference coefficient of the reference area. The specific methods are as follows:

[0081] For the temperature change curve of the reference area, for any two adjacent time points t1 and t2 on the temperature change curve, t2 is greater than t1. The difference between the temperature of t2 and the temperature of t1 is recorded as the temperature difference between adjacent time points, denoted as C1.

[0082] The difference between the pulse error at time t2 and the pulse error at time t1 in the pulse error variation curve is denoted as C2. C2 / (C1+c0) is denoted as the error growth rate between any two adjacent time points. Here, c0 is a preset data to avoid the denominator being 0. In this embodiment, c0 is equal to the maximum temperature difference between adjacent time points in the temperature variation curve.

[0083] In the temperature change curve of the reference region, the mean of the error growth rate at all adjacent time points represents the error change brought to the pulse signal by the temperature change, and is denoted as the interference coefficient of the reference region.

[0084] As another example, the amplitude of the pulse error change caused by the temperature change curve of the reference region is used as the interference coefficient of the reference region. The specific methods include the following:

[0085] Obtain the reference time period corresponding to the temperature distribution difference in the reference area.

[0086] For any two adjacent time points within the reference time period, obtain the error growth rate of any two adjacent time points. The average error growth rate of all adjacent time points within the reference time period is recorded as the interference coefficient of the reference area.

[0087] Thus, for each target area, an interference coefficient is obtained using the above method.

[0088] Step S204: Under the condition that the temperature change curves of all target areas have the smallest difference, obtain the estimation error of the pulse signal line based on the interference coefficient of all target areas.

[0089] For any target area under any operating condition, and for each pulse signal line, the pulse signal in the pulse signal line is simultaneously interfered with by all target areas. It is impossible to directly evaluate the interference of each target area on the pulse signal, which is not conducive to the subsequent redesign of the components 3 in the lower substrate 1 based on the interference of each target area on the pulse signal.

[0090] In the above process, for any target area, i.e., the reference area, based on the working condition where the temperature distribution difference is the greatest, the interference of the reference area on the pulse signal line (interference coefficient of the reference area) is initially evaluated by ignoring the interference of other target areas on the pulse signal. Among them, for the value of the temperature distribution difference, the larger the value, the more accurate the above interference coefficient is. However, in practice, it is impossible to guarantee that the value is large enough. The deformation of the pulse signal line needs to consider the interference coefficients of multiple target areas at the same time. This means that the interference coefficient obtained above can only roughly evaluate the interference of each target area on the pulse signal line. Therefore, the above interference coefficient needs to be corrected.

[0091] Specifically, the operating condition with the smallest difference in temperature change curves among all target areas is obtained. Under this condition, the interference of all target areas on the pulse signal line is the same or similar, and the interference of each target area is not negligible. This best highlights the interference of multiple target areas on the pulse signal line when they heat up simultaneously.

[0092] As an example, the method for obtaining the operating condition with the smallest difference in temperature change curves across all target areas includes:

[0093] Under any operating condition, for the temperature distribution difference of each target area (obtained using step S202 above), obtain the variance of the temperature distribution difference of all target areas, and record it as the temperature distribution characteristic under each operating condition. Obtain the operating condition with the smallest temperature distribution characteristic, under which the temperature change curve difference of all target areas is minimized.

[0094] Furthermore, under the condition of minimal difference in temperature change curves across all target areas, for any target area's temperature change curve and any two adjacent time points on that curve, the temperature difference between these two adjacent time points is obtained. The product of this temperature difference and the interference coefficient of the target area is denoted as the first product of any target area at any two adjacent time points. The sum of the first products of all target areas at any two adjacent time points is denoted as the pulse signal estimation error at any two adjacent time points, used to describe the change in pulse signal error when the temperatures of all target areas simultaneously cause interference on the pulse signal line at adjacent time points.

[0095] The estimation errors of the pulse signals at all adjacent time points constitute the estimation error variation curve.

[0096] The pulse error variation curve is processed by first-order difference (the pulse error at the next time point is subtracted from the pulse error at the previous time point in the adjacent time points of the pulse error variation curve) to obtain the measurement error variation curve.

[0097] The estimation error variation curve and the measurement error variation curve are both linearly normalized. The Pearson correlation coefficient between the normalized estimation error variation curve and the measurement error variation curve is denoted as x. exp(-|x|) is denoted as the interference coefficient correction. Here, exp() represents an exponential function with the natural constant as the base. The interference coefficient correction describes the difference in the variation trend between the estimation error and the impulse error.

[0098] The larger the interference coefficient correction (i.e., the greater the difference between the estimation error and the pulse error change trend), the more unreliable the interference coefficient becomes. When evaluating the interference of each target area on the pulse signal line, ignoring the interference of other target areas, it is impossible to use the interference coefficients of all target areas to describe and predict the common interference trend of temperature growth on the pulse signal line in all target areas. In other words, the interference coefficient is less reliable. It cannot be guaranteed that the interference coefficient of each target area can accurately evaluate the error change trend of the pulse signal caused by the heating of a single target area, nor can it be guaranteed that the interference coefficients of all target areas can be combined to jointly evaluate the interference trend of temperature growth on the pulse signal line in multiple target areas.

[0099] The smaller the interference coefficient correction, the more reliable the interference coefficient is. When obtaining the interference coefficient of each target area, after ignoring the interference of other target areas, the interference coefficient of all target areas can be used to describe and predict the temperature interference of all target areas on the pulse signal line.

[0100] It should be noted that, in the subsequent auxiliary layout design of this embodiment, not only the impact of heat generation in a single target area on the layout must be considered, but also the interference to the same pulse signal line when multiple target areas generate heat simultaneously. Therefore, the interference coefficient correction amount calculated above needs to be used to correct the interference coefficient of each target area.

[0101] It should be further explained that the purpose of linearly normalizing the aforementioned estimation error change curve and measurement error change curve is that this implementation is mainly used to analyze the changing trend of the pulse signal error when the temperature changes (described by the interference coefficient correction amount), without focusing on the magnitude of the temperature or pulse signal error change. Therefore, by removing the dimensions and orders of magnitude through linear normalization, the normalized estimation error change curve and measurement error change curve only retain the characteristics of the changing trend, which helps to more reliably evaluate the interference coefficient correction amount.

[0102] At this point, for each pulse signal line, the interference coefficient correction amount has been obtained based on the interference coefficient of all target areas. Next, the interference coefficient correction amount is used to correct the interference coefficient of each target area.

[0103] Step S205: Correct the interference coefficients of all target areas using the estimation error, so that the difference between the estimation error and the change trend of the impulse error obtained by the corrected interference coefficient is minimized.

[0104] The above obtained the interference coefficient for each target region, and any target region is denoted as the reference region. Under the condition where the distribution difference in the reference region is greatest, the following processing is performed:

[0105] On the temperature change curve of each target area, the product of the temperature difference between any two adjacent time points and the interference coefficient of each target area is denoted as the second product of each target area at any two adjacent time points. The second product represents the predicted pulse error when the interference of the target area on the pulse signal line is estimated based on the evaluated interference coefficient.

[0106] The mean of the second product of each target region at all adjacent time points is denoted as the second mean of each target region.

[0107] The ratio of the second mean of the reference area to the sum of the second means of all target areas is denoted as the correction ratio y. y represents the strength of the interference on the pulse signal line from the reference area compared to the interference from all target areas, calculated based on the interference coefficients evaluated above, under the condition where the distribution difference in the reference area is greatest. A smaller y indicates that the interference from the reference area is less than that from all target areas, meaning the interference from target areas outside the reference area cannot be ignored. In this case, a significant correction adjustment to the interference coefficient of the reference area is needed (or, in other words, a smaller correction coefficient). A larger y indicates that the interference from the reference area is greater and more significant than that from all target areas, meaning the interference from target areas outside the reference area can be ignored. In this case, the interference coefficient of the reference area does not need adjustment or only requires fine-tuning (or, in other words, a larger correction coefficient).

[0108] Specifically, the interference coefficient after the reference region correction is positively correlated with y and negatively correlated with the amount of interference coefficient correction. The larger the amount of interference coefficient correction, the less reliable the interference coefficient of all target regions, including the reference region. In this case, the interference coefficient of all target regions needs to be corrected more significantly, that is, the smaller the interference coefficient after the reference region correction.

[0109] As an example, the method for calculating the interference coefficient after correction of the reference area is as follows:

[0110] Let (1+yw) / 2 be the correction coefficient, and let the product of the correction coefficient and the original interference coefficient be the interference coefficient after correction in the reference area.

[0111] This completes the correction of the interference coefficient for each target area.

[0112] After obtaining the corrected interference coefficient for each target region, the interference coefficient correction amount is obtained again according to step S204. Then, based on the corrected interference coefficient and the newly obtained interference coefficient correction amount, step S205 is executed again to further correct the interference coefficient. Steps S204 to S205 are repeated several times (e.g., 30 times). Each time it is repeated, an interference coefficient correction amount is obtained. When the interference coefficient correction amount reaches the minimum value, the obtained corrected interference coefficient is used as the final interference coefficient for each target region. That is, the interference coefficient correction amount obtained using the final interference coefficient of each target region reaches the minimum value (see step S204 for the specific method of obtaining the interference coefficient correction amount).

[0113] The above process continuously feeds the estimated interference coefficient back into the calculation of the correction coefficient, thereby achieving multiple corrections to the interference coefficient until the correction amount reaches its minimum. During each correction (i.e., each re-execution of step S205), the interference coefficient of the reference area is corrected using the interference coefficients of other target areas outside the reference area. This gradually eliminates the interference of other target areas outside the reference area on the pulse signal, rather than directly ignoring the interference. This ensures that even when multiple target areas simultaneously interfere with the pulse signal, the interference from the temperature rise of a single target area can be obtained as accurately as possible. When the interference coefficient correction reaches its minimum, the interference coefficients of multiple target areas can be combined to jointly assess the changing trend of temperature increase in multiple target areas affecting the pulse signal line. In subsequent auxiliary layout design, the impact of heat generation from a single target area on the layout can be considered, as well as the interference from multiple target areas simultaneously generating heat on the same pulse signal line.

[0114] Step S206: Use the corrected interference coefficients of all target areas under all operating conditions to assist in updating the layout design of components 3 in the lower substrate 1.

[0115] In the above process, under each operating condition, for all target areas traversed by each pulse signal line, each target area corresponds to a final interference coefficient; the target area with the largest final interference coefficient is recorded as the region of interest; the sum of the interference coefficients of all target areas outside the region of interest is recorded as the interference threshold.

[0116] Thus, for each operating condition, a region of interest and an interference threshold are calculated for each pulse signal line.

[0117] Under all operating conditions, all components 3 within all areas of interest corresponding to each pulse signal line are denoted as component 3 of each pulse signal line, and the average value of all interference thresholds corresponding to each pulse signal line is denoted as the warning threshold of each pulse signal line.

[0118] In other embodiments, the product of the average of all interference thresholds corresponding to each pulse signal line and the preset scaling factor is recorded as the warning threshold of each pulse signal line; the preset scaling factor in this embodiment is 0.78, and the preferred value range of the preset scaling factor is [0.3, 1].

[0119] For any component 3 located below any pulse signal line, obtain the region where the component 3 is located, and obtain the average value of the interference coefficient of the region under all operating conditions, which is recorded as the interference coefficient of component 3.

[0120] When the layout designer rearranges the components 3 in the lower substrate 1, if the components 3 are located below each corresponding pulse signal line, or if the sum of the interference coefficients of all components 3 below each pulse signal line is greater than or equal to the corresponding warning threshold, a warning will be issued to remind the layout designer that the current layout may interfere with the pulse signal.

[0121] The layout designers completed the layout of component 3 based on the above warning prompts.

[0122] Specifically, for all components in areas not traversed by the pulse signal line (or outside the target area), components with a maximum temperature and minimum temperature difference greater than 10°C will not be moved during relocation. Other components will remain in their original positions during relocation, or will be given a warning if they are located below any pulse signal line.

[0123] In this embodiment, the interference coefficient of the target area can be used to describe the interference of the component 3's temperature rise on the pulse signal in each target area, and can also describe the interference of multiple target areas on the pulse signal under the combined effect of multiple target areas by combining the interference coefficients of multiple target areas. This allows for auxiliary layout from both the position of a single component 3 and the distribution of multiple components 3, which helps to further improve the performance of the circuit module.

[0124] Example 3:

[0125] In the above embodiment one, the influence of the reference region on the pulse signal is evaluated by ignoring the interference of other target regions outside the reference region. Then, the interference coefficient is gradually updated and corrected based on the combined influence of the interference coefficients evaluated by all target regions on the pulse signal line, gradually eliminating the influence of other target regions (or making the influence of other target regions gradually negligible). However, when gradually eliminating the influence in embodiment one, over-elimination may occur, causing the above-mentioned interference coefficient correction amount to fail to reach the minimum value quickly, or to fail to reach the global minimum value. Over-elimination means that after obtaining the corrected interference coefficients of all target regions each time, only the interference coefficients of some target regions can ignore the interference of other target regions, so that the interference coefficient of each target region can reliably describe the interference of a single target region on the signal line; while the interference coefficients of other target regions cannot ignore the interference of other target regions; and when obtaining the corrected interference coefficients of all target regions again, the interference coefficients of some target regions will over-eliminate the interference of other target regions.

[0126] In this embodiment, the process of re-acquiring the interference coefficient of each target region (i.e., correcting the interference coefficient of each target region each time) and obtaining the interference coefficient correction amount based on the re-acquired interference coefficients of all target regions (i.e., the corrected interference coefficients) is referred to as an elimination process. In each elimination process, the corrected interference coefficients of all target regions are recalculated, and the interference coefficient correction amount is obtained based on the corrected interference coefficients of all target regions.

[0127] After more than 10 elimination processes, each time an elimination process is performed, for any target area, multiple interference coefficients are obtained for that target area. These interference coefficients are arranged in the order they were obtained, and the interference coefficients with the largest reduction are selected from them to form the first interference coefficient sequence of that target area.

[0128] For each interference coefficient in the first interference coefficient sequence, the interference coefficient can be used to obtain the interference coefficient correction amount (see step S204 for details). That is, each interference coefficient in the first interference coefficient sequence corresponds to an interference coefficient correction amount. All interference coefficients in the first interference coefficient sequence of the target region correspond to the obtained interference coefficient correction amounts to form a first correction amount sequence. The first interference coefficient sequence and the first correction amount sequence are linearly normalized respectively to obtain the Pearson correlation coefficient between the normalized first interference coefficient sequence and the normalized first correction amount sequence. This is recorded as the appropriate elimination coefficient for the target region after each elimination process.

[0129] After each elimination process, the elimination appropriate coefficient for each target area is obtained using the method described above.

[0130] The smaller the elimination appropriate coefficient, the less significant the correction amount of the interference coefficient is as the elimination process progresses and the interference coefficient decreases. This suggests that each target area may have excessively eliminated interference from other target areas.

[0131] When the elimination appropriate coefficient of each target region is less than or equal to the second preset threshold th2, the elimination process of each target region is suppressed in the next elimination process.

[0132] As an example, in the next elimination process, the elimination process for each target region is suppressed, including the following methods:

[0133] In the next elimination process, the following steps will be taken:

[0134] The correction coefficient obtained for each target region is denoted as Q (see step S205 for details). The suppression correction coefficient is obtained. The suppression correction coefficient is positively correlated with Q and negatively correlated with the elimination appropriate coefficient of each target region.

[0135] Then, the product of the suppression correction coefficient and the interference coefficient before correction for each target region (that is, the interference coefficient after correction for each region in the previous elimination process) is recorded as the interference coefficient after correction for each target region in the next elimination process.

[0136] In the above process, the more excessively each target region eliminates the influence of other target regions (i.e., the smaller the elimination appropriate coefficient), the larger the suppression correction coefficient will be in the next elimination process. This will prevent the interference coefficient of each target region from becoming too small in the next elimination process, thus preventing each target region from further excessively eliminating the influence of other target regions in the next elimination process.

[0137] It should be noted that when the elimination appropriate coefficient of each target region is greater than the second preset threshold th2, the elimination process of each target region is also suppressed in the next elimination process, but the suppression correction coefficient used is set to 0 (equivalent to no suppression).

[0138] This embodiment uses th2=0.3 as an example for description, and the preferred value range of th2 is [0,0.5].

[0139] As an example, the suppression correction coefficient is obtained, which is positively correlated with Q and negatively correlated with the elimination appropriate coefficient of each target region. The formula is: Q1=(1-q)×Q, where Q1 represents the suppression correction coefficient.

[0140] As an example, these interference coefficients are arranged in the order they were obtained, and the interference coefficients with the largest reduction are selected from them to form the first interference coefficient sequence for each target region. The method includes:

[0141] Arrange these interference coefficients in the order they were obtained to obtain the first sequence.

[0142] Transform all interference coefficients in the first sequence into a tree structure. Specifically, obtain the first interference coefficient a1 in the first sequence and use a1 as the root node of the tree structure. Obtain all interference coefficients in the first sequence that are less than a1 and follow the position corresponding to a1, and use them as child nodes of the root node, denoted as the first-level nodes of the tree structure. For any interference coefficient a2 corresponding to any node in the first level, obtain all interference coefficients in the first sequence that are less than a2 and follow the position corresponding to a2, and use them as child nodes of the first level, denoted as the second-level nodes of the tree structure. For any interference coefficient a3 corresponding to any node in the second level, obtain all interference coefficients in the first sequence that are less than a3 and follow the position corresponding to a3, and use them as child nodes of the second level, denoted as the third-level nodes. Continue in this manner until all interference coefficients in the first sequence are used as nodes in the tree structure.

[0143] For a path from the root node to any leaf node in a tree structure, the absolute value of the difference between the interference coefficients of two adjacent nodes in the path is obtained, and the sum of the absolute values ​​of the differences between all adjacent nodes in the path is recorded as the reduction in the interference coefficient of the path.

[0144] Obtain the path with the largest reduction in interference coefficient. The interference coefficients of all nodes in this path constitute the first interference coefficient sequence for each target region.

[0145] In this embodiment, the first interference coefficient sequence contains the interference coefficients with the most obvious decreasing trend among all interference coefficients. Specifically, if the number of nodes in the above path is less than or equal to 5, then the path is not considered.

[0146] In other examples, to reduce computational cost, all the minimum values ​​in the first sequence can be directly used as the first interference coefficient sequence.

[0147] Example 4:

[0148] As an example, in step S201 of embodiment one, the method for obtaining the operating condition is as follows:

[0149] When the circuit module is working normally, it collects input data in real time (e.g., once every 1 second). In this embodiment, each input data is regarded as a vector. For example, when the circuit module is used to control the charging and discharging of the battery pack, it combines the output voltage and current of each battery pack, the output voltage and current of the charger, and other data (e.g., remaining power, usage time, set warning voltage and current, etc.) to form the vector corresponding to the input data. In particular, if some data in the input data has no value (e.g., the charger is not connected to the circuit module), then its value is set to 0.

[0150] All the input data obtained so far are whitened (in this embodiment, the ZCA whitening algorithm is used for whitening), the purpose of which is to remove the dimensions of the input data and normalize the input data.

[0151] Then, mean-shift clustering is performed on all input data after whitening to obtain all categories. The mean of all whitened output data in each category is obtained, and the whitened output data with the smallest Euclidean distance from the mean in each category is recorded as the category center. The input data corresponding to each category center is used as each working condition.

[0152] As another example, the operating conditions can be obtained by manually collecting input data from the circuit module at appropriate times when it is working normally. For example, input data can be collected when the battery charge is 100%, 30%, and 10%, or when the power grid to which the battery bank is connected fluctuates (e.g., when a load is switched on or off, or when a distributed power source is switched on or off).

[0153] As an example, in step S201 of embodiment one, the method for collecting the temperature change curve of each region includes:

[0154] In a laboratory environment, a circuit module was randomly selected for testing. The upper substrate 2 and the lower substrate 1 in the circuit module were separated (no longer stacked on top of each other), and the pins 5 of the upper substrate 2 and the lower substrate 1 were connected with enameled wire.

[0155] Input data for each operating condition is connected from interface 6 to the lower substrate 1 at room temperature (e.g., 25°C). Infrared images of the lower substrate 1 are acquired using infrared thermometry (e.g., using a thermal infrared camera facing the lower substrate 1). The grayscale value of each pixel in the infrared image represents the temperature at each location in the lower substrate 1 (including each location of each component 3).

[0156] Within ten minutes of receiving the input data (i.e., within the preset time period described in Example 1), the temperature at each location is collected every second (using infrared thermometry technology) and the average temperature of all locations within each region is recorded. This average temperature of each region at each time point is recorded, and the average temperature of each region at all time points is recorded as the temperature change curve that constitutes the average temperature over time.

[0157] After obtaining the temperature change curves for all regions, the pulse error change curve for each pulse signal line is obtained, specifically including:

[0158] After cooling the lower substrate 1 to room temperature, the upper substrate 2 and the lower substrate 1 are re-stacked, and then the input data for each operating condition is reconnected. A test signal (e.g., a 1MHz square wave signal) is input to one end of each pulse signal line using a signal generator, and the response pulse signal at the other end of each pulse signal line is acquired using an oscilloscope. The pulse signal lines include signal lines connecting to an external crystal oscillator (e.g., clock signal lines for serial communication) and signal lines transmitting PWM signals.

[0159] Within ten minutes of the start of the input data access (i.e., within the preset time period described in Example 1), a pulse signal (e.g., a pulse signal with a duration of 0.1 milliseconds) is collected every second as a time point, which is simply referred to as each pulse signal.

[0160] Obtain the average duration of the rising edge, the average duration of the falling edge, and the average duration of the high level between adjacent rising and falling edges for each pulse signal. Record the average of the average duration of the rising edge, the average duration of the falling edge, and the average duration of the high level as the first pulse evaluation result.

[0161] In some embodiments, the variance of the signal value during each rising edge and the variance of the signal value during each falling edge are obtained, and all variances corresponding to all rising and falling edges are used as the first pulse evaluation result.

[0162] The pulse signal acquired at the first time point is recorded as the reference signal. The difference between the first pulse evaluation result of each pulse signal and the first pulse evaluation result of the reference signal is obtained. The ratio of this difference to the first pulse evaluation result of the reference signal is recorded as the pulse error of the pulse signal acquired at each time point. The pulse errors acquired at all time points constitute the pulse error variation curve.

[0163] The pulse error variation curves of all pulse signal lines were obtained using the method described above.

[0164] Specifically, if the maximum value of the pulse error in the pulse error variation curve of any pulse signal line is less than 0.01, it means that the pulse error of the pulse signal line is close to 0 and can be ignored. In this case, the pulse signal line is not considered when implementing Example 1 (that is, the pulse signal line is regarded as not existing).

[0165] As an example, in step S202 of Embodiment 1, the method for obtaining the average temperature and average slope is as follows:

[0166] For any temperature range variation curve, or the temperature sequence of a temperature range variation curve within any time window.

[0167] Obtain the temperature values ​​corresponding to all time points, and record the average temperature as the mean value of all time points.

[0168] The temperature difference between any two adjacent time points is obtained, and the average of the temperature differences between all adjacent time points is denoted as the average slope.

[0169] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A non-sealed, hybrid, highly integrated dual-layer substrate circuit module, the circuit module comprising an upper substrate (2) and a lower substrate (1), wherein components (3) are soldered onto the upper substrate (2) and the lower substrate (1), characterized in that, The upper substrate (2) serves as the pulse signal layer, and the lower substrate (1) serves as the power layer. The circuit printing thickness of the pulse signal layer is less than that of the power layer. Based on the interference of the temperature distribution of the lower substrate (1) on the pulse signal in the upper substrate (2) during the operation of the circuit module, the layout design of the components (3) in the lower substrate (1) is updated, including the following specific steps: The temperature change curve of each region on the lower substrate (1) of the circuit module under each working condition is obtained, as well as the pulse error of the pulse signal line in the upper substrate (2); all regions passed under the pulse signal line are recorded as target regions, and any target region is recorded as reference region; Under the same operating condition, the difference between the temperature change curve of the reference area and the temperature change curves of all target areas outside the reference area is denoted as the temperature distribution difference of the reference area under the same operating condition. Under the condition where the distribution difference in the reference area is the greatest, the amplitude of the pulse error change caused by the temperature change curve of the reference area is used as the interference coefficient of the reference area. Under the condition where the temperature change curves of all target areas have the smallest difference, the estimation error of the pulse signal line is obtained based on the interference coefficient of all target areas. The estimation error is used to correct the interference coefficient of all target areas so that the difference between the estimation error obtained by the corrected interference coefficient and the change trend of the pulse error is the smallest. The corrected interference coefficient of all target areas under all conditions is used to assist in updating the layout design of components (3) in the lower substrate (1).

2. The dual-layer substrate circuit module with non-sealed hybrid high integration according to claim 1, characterized in that, The specific steps involved in correcting the interference coefficients of all target regions using estimation errors to minimize the difference between the estimation error and the trend of the impulse error in the corrected interference coefficients are as follows: D1: The estimation error obtained for the interference coefficient of all target areas, the difference between the trend of the estimation error and the impulse error is recorded as the interference coefficient correction amount; D2: Under the condition of greatest distribution difference in the reference area, on the temperature change curve of each target area, the product of the temperature difference at adjacent time points and the interference coefficient of each target area is denoted as the prediction pulse error of each target area at any two adjacent time points; the mean of the prediction pulse errors of each target area at all adjacent time points is denoted as the second mean of each target area; the ratio of the second mean of the reference area to the sum of the second means of all target areas is denoted as the correction ratio y; the interference coefficient of the reference area after correction is positively correlated with the correction coefficient; the correction coefficient is positively correlated with y and negatively correlated with the amount of interference coefficient correction; D3: After D2 is completed, D1 is re-executed using the corrected interference coefficients for each target region. This includes: the estimation error obtained for the corrected interference coefficients of all target regions, and the difference between the estimation error and the change trend of the impulse error is recorded as the interference coefficient correction amount. Then, the process of executing D2 and D1 is recorded as one elimination process. After repeating the elimination process several times, the corrected interference coefficient obtained when the interference coefficient correction reaches the minimum value is taken as the final interference coefficient of each target area. For the final interference coefficient of all target areas, the difference between the estimation error and the change trend of the impulse error obtained by the final interference coefficient is minimized.

3. The dual-layer substrate circuit module with non-sealed hybrid high integration according to claim 2, characterized in that, After each elimination process, for any target region, among all the interference coefficients obtained for that target region, the interference coefficients with the largest reduction are selected to form the first interference coefficient sequence for that target region. All the interference coefficients in the first interference coefficient sequence of that target region correspond to the obtained interference coefficient correction amounts to form the first correction amount sequence. The first interference coefficient sequence and the first correction amount sequence are linearly normalized respectively, and the Pearson correlation coefficient between the normalized first interference coefficient sequence and the normalized first correction amount sequence is obtained. This is recorded as the appropriate elimination coefficient for that target region after each elimination process. When the elimination suitability coefficient of each target region is less than or equal to the second preset threshold, the elimination process of each target region is suppressed during the next elimination process, specifically including: Obtain the suppression correction coefficient, which is positively correlated with the correction coefficient obtained for each target region and negatively correlated with the elimination appropriate coefficient for each target region; In the next elimination process, the corrected interference coefficient of each target region is positively correlated with the suppression correction coefficient.

4. The dual-layer substrate circuit module with non-sealed hybrid high integration according to claim 2, characterized in that, The specific steps regarding the difference in the changing trends of the estimation error and the impulse error are as follows: The estimation error obtained for the interference coefficients of all target areas, or the estimation error obtained for the corrected interference coefficients of all target areas, represents the estimation error of the pulse signal at any adjacent time point. The estimation errors of the pulse signals at all adjacent time points constitute the estimation error variation curve; the first-order difference result of the pulse error of the pulse signal line at all time points is denoted as the measurement error variation curve; the estimation error variation curve and the measurement error variation curve are respectively linearly normalized, and the Pearson correlation coefficient between the normalized estimation error variation curve and the measurement error variation curve is denoted as x. The difference in the variation trend between the estimation error and the pulse error is negatively correlated with x.

5. A dual-layer substrate circuit module employing non-sealed hybrid high integration as described in claim 1 or 4, characterized in that, The specific steps for obtaining the estimation error are as follows: Under the condition that the temperature change curves of all target areas have the smallest difference, for any target area's temperature change curve and any two adjacent time points on that temperature change curve, obtain the temperature difference between any two adjacent time points; for the interference coefficient of the target area or the corrected interference coefficient of the target area, the product of the temperature difference and the interference coefficient of the target area is recorded as the first product of any target area at any two adjacent time points; the sum of the first products of all target areas at any two adjacent time points is recorded as the estimation error of the pulse signal at any two adjacent time points.

6. The dual-layer substrate circuit module with non-sealed hybrid high integration according to claim 1, characterized in that, The amplitude of the pulse error change caused by the temperature change curve of the reference region is used as the interference coefficient of the reference region. The specific steps include the following: For the temperature change curve of the reference region, the temperature difference between any two adjacent time points on the temperature change curve is represented as C1. The difference in pulse error between adjacent time points in the pulse error variation curve is denoted as C2. C2 / (C1+c0) is denoted as the error growth rate between any two adjacent time points, where c0 is a preset data. The average error growth rate of all adjacent time points in the temperature variation curve of the reference area is denoted as the interference coefficient of the reference area.

7. The dual-layer substrate circuit module with non-sealed hybrid high integration according to claim 1, characterized in that, The difference between the temperature change curve of the reference area and the temperature change curves of all target areas outside the reference area is recorded as the temperature distribution difference of the reference area under the same operating condition. The specific steps include the following: The product of the average slope and average temperature of each temperature change curve is recorded as the evaluation index of each temperature change curve. The mean of the evaluation indexes of the temperature change curves of all target areas outside the reference area is recorded as the first mean. The difference between the evaluation index of the temperature change curve of the reference area and the first mean is recorded as the temperature distribution difference of the reference area.

8. The dual-layer substrate circuit module with non-sealed hybrid high integration according to claim 1, characterized in that, The specific steps for obtaining the operating condition with the smallest difference in temperature change curves across all target regions are as follows: Under any operating condition, for the temperature distribution difference of each target area, obtain the variance of the temperature distribution difference of all target areas, and record it as the temperature distribution characteristic under each operating condition. Obtain the operating condition with the smallest temperature distribution characteristic, under which the temperature change curve difference of all target areas is minimized.

9. A dual-layer substrate circuit module employing non-sealed hybrid high integration as described in claim 1, characterized in that, The specific steps for using the corrected interference coefficients of all target regions under all operating conditions to assist in updating the layout design of components (3) in the lower substrate (1) are as follows: The corrected interference coefficient obtained when the difference between the variation trend of the estimation error and the pulse error is minimal is recorded as the final interference coefficient; under each operating condition, for all target areas traversed by each pulse signal line, each target area corresponds to a final interference coefficient; The target area with the largest final interference coefficient is recorded as the area of ​​interest for each pulse signal line under each operating condition; The sum of the interference coefficients of all target areas outside the area of ​​interest is recorded as the interference threshold of each pulse signal line under each operating condition; Under all operating conditions, all components (3) in all areas of interest corresponding to each pulse signal line are recorded as components of each pulse signal line, and the average value of all interference thresholds corresponding to each pulse signal line is recorded as the warning threshold of each pulse signal line. For any component (3) under any pulse signal line, obtain the area where the component (3) is located, obtain the average value of the interference coefficient corresponding to the area under all operating conditions, and record it as the interference coefficient of the component (3); When rearranging the components (3) in the lower substrate (1), if the components (3) are located below each pulse signal line, or if the sum of the interference coefficients of all components (3) below each pulse signal line is greater than or equal to the corresponding warning threshold, a warning is issued.

10. A dual-layer substrate circuit module employing non-sealed hybrid high integration as described in claim 1, characterized in that, The upper substrate (2) and the lower substrate (1) are stacked vertically by support pillars (4). Both the upper substrate (2) and the lower substrate (1) contain pins (5). The pins (5) in the upper substrate (2) and the lower substrate (1) are connected by copper wire to realize the electrical signal connection between the upper substrate (2) and the lower substrate (1). An interface (6) is installed on the lower substrate (1) to supply power to the entire circuit module and input the input data corresponding to each working condition.

Images