Impedance adjusting method and impedance adjusting system suitable for power supply delivery network

Abstract

An impedance adjusting method is suitable for a power supply delivery network comprising a plurality of observation points and a plurality of capacitor positions. The impedance adjustment method includes generating a corresponding weight value according to a current value of each of the observation points, determining an optimal capacitance position from the capacitance positions according to the weight value corresponding to each of the observation points, generating a plurality of curves according to a plurality of self-impedances of the observation points, generating a plurality of intersection point frequencies according to the curves and a target impedance curve corresponding to target impedance, generating an average intersection point frequency according to the weight value corresponding to each observation point and the intersection point frequencies, and selecting a group of candidate capacitors according to the average intersection point frequency, and generating a plurality of adjusted self-impedances according to the group of candidate capacitors and the optimal capacitor position so as to select the optimal capacitor from the group of candidate capacitors.

Description

Technical Field

[0001] The present invention relates to an impedance adjustment method and an impedance adjustment system applicable to a power delivery network, and in particular to an impedance adjustment method and an impedance adjustment system capable of selecting an optimal capacitor position and an optimal capacitor. Background Art

[0002] The power distribution network (PDN) is a transmission medium for the voltage source to provide DC power to the load, and on a printed circuit board, it is usually the transmission path from the voltage regulator module (VRM) to the load circuit. The power distribution network can be composed of the power layer and the ground layer of the printed circuit board, and the cables, connectors, capacitors and other components electrically connected to the power layer and the ground layer can all be part of the power distribution network.

[0003] The parasitic effects of the power delivery network can cause parallel resonance, resulting in high self-impedance at the load circuit location. At the load circuit location, voltage disturbances may occur due to changes in load current and the high self-impedance of the power delivery network.

[0004] In addition, when the self-impedance versus frequency curve of the power delivery network is not flat, the voltage disturbance problem will be aggravated by the distorted wave problem caused by the overlapping of voltage disturbances.

[0005] A server is an electronic device that includes a printed circuit board. The power layer and ground layer of the printed circuit board can constitute a part of the power delivery network. Therefore, in the design of the power delivery network of the printed circuit board of the server, in addition to considering reducing impedance to avoid noise interference, it is also necessary to consider the problem of impedance flattening to avoid distortion. Summary of the invention

[0006] An embodiment provides an impedance adjustment method, which is applicable to a power delivery network, wherein the power delivery network includes a plurality of observation points. The impedance adjustment method includes generating a corresponding weight value according to a current value of each of the observation points, determining an optimal capacitor position from a plurality of capacitor positions of the power delivery network according to the weight value corresponding to each of the observation points, generating a plurality of curves according to a plurality of self-impedances of the observation points, generating a plurality of intersection frequencies according to the curves and a target impedance curve corresponding to a target impedance, generating an average intersection frequency according to the weight value corresponding to each of the observation points and the intersection frequencies, selecting a group of candidate capacitors from a plurality of capacitors according to the average intersection frequency, generating a plurality of adjusted self-impedances according to the group of candidate capacitors and the optimal capacitor position, and selecting an optimal capacitor from the group of candidate capacitors according to the adjusted self-impedances.

[0007] The method further comprises: placing the optimal capacitor into the optimal capacitor position; updating the plurality of adjusted self-impedances; and when at least one of the updated plurality of adjusted self-impedances is greater than the target impedance, determining a current optimal capacitor position from other capacitor positions of the plurality of capacitor positions except the optimal capacitor position according to at least the plurality of weight values.

[0008] The average intersection frequency is generated according to the weight value corresponding to each of the plurality of observation points and the plurality of intersection frequencies, including: obtaining a sum value according to the product of a j-th weight value and a j-th intersection frequency of the plurality of intersection frequencies, wherein the sum value is positively correlated to the average intersection frequency, j is a positive integer greater than 0, and j is less than or equal to the number of the plurality of observation points.

[0009] A self-resonant frequency of each of the group of candidate capacitors is between a times and b times the average intersection frequency, wherein a and b are parameters greater than zero, a is less than b, a is between 0.1 and 5, and b is between 1 and 10.

[0010] Where a is equal to 0.5 and b is equal to 2.

[0011] The method of selecting the optimal capacitor from the group of candidate capacitors according to the plurality of adjusted self-impedances comprises: generating a plurality of impedance-to-frequency curves according to the plurality of adjusted self-impedances; generating a plurality of slope absolute value sum values ​​according to the plurality of impedance-to-frequency curves; generating a plurality of reference values ​​according to the plurality of slope absolute value sum values ​​and the plurality of weight values; and selecting a minimum capacitor corresponding to the plurality of reference values ​​from the group of candidate capacitors as the optimal capacitor.

[0012] Another embodiment provides an impedance adjustment system, which is applicable to a power delivery network, the power delivery network including a plurality of observation points, the impedance adjustment system including a capacitor position selection unit, and a capacitor selection unit. The capacitor position selection unit is used to generate a corresponding weight value according to a current value of each of the observation points, and determine an optimal capacitor position from the plurality of capacitor positions of the power delivery network according to the weight value corresponding to each of the observation points. The capacitor selection unit is used to generate a plurality of curves according to the plurality of self-impedances of the observation points, generate a plurality of intersection frequencies according to the curves and a target impedance curve, generate an average intersection frequency according to the weight value corresponding to each of the observation points and the intersection frequencies, select a group of candidate capacitors from the plurality of capacitors according to the average intersection frequency, generate a plurality of adjusted self-impedances according to the group of candidate capacitors and the optimal capacitor position, and select an optimal capacitor from the group of candidate capacitors according to the adjusted self-impedances.

[0013] The capacitor position selection unit generates the plurality of weight values ​​according to a plurality of currents respectively corresponding to the plurality of observation points.

[0014] A self-resonant frequency of each of the group of candidate capacitors is between a times and b times the average intersection frequency, wherein a and b are parameters greater than zero, a is less than b, a is between 0.1 and 5, and b is between 1 and 10. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 FIG. 4 is a schematic diagram of a simplified power delivery network in an embodiment.

[0016] Figure 2 FIG. 4 is a schematic diagram of a power delivery network in an embodiment.

[0017] Figure 3 FIG. 4 is a schematic diagram of an impedance adjustment system applicable to a power delivery network according to an embodiment.

[0018] Figure 4 FIG. 4 is a flow chart of an impedance adjustment method applicable to a power delivery network in an embodiment.

[0019] Figure 5 FIG. 4 is a schematic diagram of generating a plurality of crossover frequencies in an embodiment.

[0020] Figure 6 FIG. 4 is a schematic diagram of the range of self-resonant frequencies of candidate capacitors in an embodiment.

[0021] Figure 7 for Figure 3 Flow chart of selecting the optimal capacitor in the step.

[0022] Figure 8 FIG. 4 is a schematic diagram of one of the plurality of impedance versus frequency curves in an embodiment.

[0023] Explanation of symbols:

[0024] 10,100: Power Delivery Network

[0025] 200: Impedance adjustment system

[0026] 210: Capacitor position selection unit

[0027] 220: Capacitor selection unit

[0028] 300: Impedance adjustment method

[0029] 305 to 399, 3710 to 3740: Steps

[0030] C1,C2,C3,Cjd: Curve

[0031] fx_avg: average intersection frequency

[0032] fx1,fx2,fx3: intersection frequency

[0033] L1 to L6: Line segment

[0034] Max1 to Max8: local maximum

[0035] Min1 to Min8: local minimum

[0036] P01, P02: Port

[0037] P1 to PJ, Pj: observation point

[0038] Pc1 to PcI, Pci: Capacitor location

[0039] R: Range

[0040] Z11, Z22, Z33: Self-impedance

[0041] Zt: Target impedance DETAILED DESCRIPTION

[0042] Herein, when it is said that the impedance is flattened, it means that the impedance curve in the two-dimensional space of impedance versus frequency is flattened.

[0043] Figure 1 1 is a schematic diagram of a simplified power delivery network 10 in an embodiment. The power delivery network 10 may be a 2-port network having a first port P01 (e.g., coupled to a load device such as an integrated circuit) and a second port P02 (e.g., coupled to a decoupling capacitor), and may perform calculations of equations eq-01 and eq-02:

[0044]

[0045]

[0046] Wherein, Z11 and Z22 may be the self-inductance (also referred to as self-impedance in this article) of the first port P01 and the second port P02, and Z12 may be the mutual inductance of the first port P01 and the second port P02. Z11, Z12, Z22 and Zd may all be elements of the impedance matrix (also referred to as Z-matrix) of the power delivery network 10.

[0047] Z of formula eq-01 11wdecap@2 It can be the adjusted element of the impedance matrix after the second port P02 is set with a decoupling capacitor, and ΔZ of the formula eq-0211wGND@2 It can be the inductance characteristic reduced after the second port P02 is provided with a decoupling capacitor. Wherein, Zd can be the self-inductance value of the provided decoupling capacitor, which can be a constant and can be ignored for the convenience of calculation.

[0048] The above equations eq-01 and eq-02 are only examples. If in an n-port network, the general formula for the mutual inductance of the Ath port of the n ports can be shown as equations eq-03 and eq-04:

[0049]

[0050]

[0051] Wherein, Z1n, Znn, Z1A, ZAn and ZAA may be elements of the impedance matrix (Z-matrix) of the power delivery network. Formulas eq-01, eq-02, eq-03 and eq-04 may be applied to the following step 320 and formulas eq-2.1 and eq-2.2.

[0052] Figure 2 1 is a schematic diagram of a power delivery network 100 in an embodiment. The power delivery network 100 may include a plurality of observation points P1 to PJ and a plurality of capacitor positions Pc1 to PcI, wherein the jth observation point may be labeled as Pcj and the ith capacitor position may be labeled as Pci. Parameters i, j, I and J may be integers greater than zero, 1≤i≤I, and 1≤j≤J. Each of the plurality of observation points P1 to PJ may be a port for coupling a load device, such as an integrated circuit. Each of the plurality of capacitor positions Pc1 to PcI may be a port for coupling a decoupling capacitor.

[0053] Figure 3 2 is a schematic diagram of an impedance adjustment system 200 in an embodiment. The impedance adjustment system 200 may include a capacitor position selection unit 210 and a capacitor selection unit 220. The capacitor position selection unit 210 may be used to select the best capacitor position among the capacitor positions Pc1 to Pc1 of the power delivery network 100. The capacitor selection unit 220 may be used to select the best capacitor in the capacitor library. The best capacitor may be set at the best capacitor position to reduce and flatten the impedance of the power delivery network 100. The capacitor position selection unit 210 and the capacitor selection unit 220 may be respectively set in different software, hardware and / or firmware, or the capacitor position selection unit 210 and the capacitor selection unit 220 may be integrated into the same software, hardware and / or firmware. The related operations of the capacitor position selection unit 210 and the capacitor selection unit 220 may be described below. For example, the capacitor position selection unit 210 and the capacitor selection unit 220 may be set in hardware devices such as circuits, controllers, processors and / or chips.

[0054] Figure 4 1 is a flow chart of an impedance adjustment method 300 in an embodiment. The impedance adjustment method 300 may be used to adjust the power delivery network 100 and includes the following steps:

[0055] Step 305: extract scattering parameters (also called S-parameters) of the power delivery network 100, and generate impedance parameters according to the scattering parameters, wherein the impedance parameters may be elements of an impedance matrix (Z-matrix) of the power delivery network 100;

[0056] Step 310: Generate a plurality of weight values ​​(hereinafter referred to as w1 to wJ) corresponding to a plurality of observation points P1 to PJ respectively;

[0057] Step 320: Determine the best capacitor position from among the plurality of capacitor positions Pc1 to PcI according to at least a plurality of weight values ​​w1 to wJ;

[0058] Step 330: Generate a plurality of curves according to a plurality of self-impedances of a plurality of observation points P1 to PJ;

[0059] Step 340: Generate a plurality of intersection frequencies (hereinafter referred to as fx1 to fxJ) according to the curves and a target impedance curve corresponding to a target impedance (hereinafter referred to as Zt);

[0060] Step 350: Generate an average intersection frequency (hereinafter referred to as fx_avg) according to a plurality of weight values ​​w1 to wJ and a plurality of intersection frequencies fx1 to fxJ;

[0061] Step 360: Select a group of candidate capacitors (hereinafter referred to as C1 to CD, where the dth candidate capacitor may be Cd, d and D are positive integers, and 1≤d≤D) from the plurality of capacitors according to the average crossover frequency fx_avg;

[0062] Step 365: Generate a plurality of adjusted self-impedances according to the candidate capacitors C1 to CD and the optimal capacitor positions;

[0063] Step 370: Selecting an optimal capacitor from the set of candidate capacitors C1 to CD according to a plurality of adjusted self-impedances;

[0064] Step 380: placing the optimal capacitor at the optimal capacitor position;

[0065] Step 385: updating a plurality of adjusted self-impedances;

[0066] Step 390: Check whether at least one of the updated plurality of adjusted self-impedances is greater than the target impedance Zt? If yes, proceed to step 395; if no, proceed to step 399;

[0067] Step 395: Determine the current best capacitor position from among the other capacitor positions of the plurality of capacitor positions Pc1 to PcI except the best capacitor position according to at least a plurality of weight values ​​w1 to wJ, so as to update the best capacitor position, and proceed to step 330; and

[0068] Step 399: End.

[0069] In the impedance adjustment method 300, the impedance parameter generated in step 305 can be used in step 320. Steps 310 and 320 can be performed by the capacitor position selection unit 210 to select the best capacitor position. Steps 330 to 370 can be performed by the capacitor selection unit 220 to select the best capacitor.

[0070] In step 390, if the updated majority of adjusted self-impedances are not greater than the target impedance Zt in the predetermined frequency range, the impedance of the power delivery network 100 has been adjusted to an acceptable range. In other words, the purpose of performing the impedance adjustment method 300 has been achieved, so the process can be terminated by entering step 399. In step 390, if the updated majority of adjusted self-impedances are still greater than the target impedance Zt in the predetermined frequency range, step 395 can be entered to update the optimal capacitor position, and the current optimal capacitor can be selected again to place the optimal capacitor in the updated optimal capacitor position to further adjust the impedance of the power delivery network 100.

[0071] The details of each step of the impedance adjustment method 300 are as follows. In step 310, corresponding weight values ​​w1 to wJ can be generated according to the current values ​​of the plurality of observation points P1 to PJ, wherein the jth weight value wj of the jth observation point Pj can be as described in equation eq-1:

[0072]

[0073] The current Ij may be the current value of the jth observation point. Therefore, w1 to wJ may be generated accordingly. For example, if J=3, and the current values ​​I1, I2 and I3 of the observation points P1 to P3 are 1 ampere, 2 amperes and 3 amperes respectively, the weight values ​​w1 to w3 may be 1 / 6, 2 / 6 (i.e., 1 / 3) and 3 / 6 (i.e., 1 / 2) respectively.

[0074] In step 320, the current best capacitor position can be determined from the plurality of capacitor positions Pc1 to PcI according to equations eq-2.1 and eq-2.2. Equations eq-2.1 and eq-2.2 can be used to calculate the importance of the capacitor position.

[0075]

[0076] importancePDN@porti = argmax i {wjximportance j@porti}...eq-2.2;

[0077] Zii and Zij in equation eq-2.1 may be elements of the impedance matrix (Z-matrix) of the power delivery network 100. Equation eq-2.1 may be described as, for the jth observation point Pj, short-circuiting the capacitor positions Pc1 to PcI one by one to find the i-th capacitor position Pci with the greatest influence, and the i-th capacitor position Pci may be the capacitor position with the highest importance for the jth observation point Pj. Equation eq-2.1 may be related to the above equation eq-02.

[0078] In equation eq-2.2, each weight value wj (1≤j≤J) of the weight values ​​w1 to wJ can be further considered according to the result of equation eq-2.1, and the argmax function can be used to obtain the i-th capacitor position Pci with the greatest impact on the entire power delivery network 100. The capacitor position Pci corresponding to the i value generated by equation eq-2.2 can be the optimal capacitor position determined in step 320.

[0079] To illustrate step 330 and step 340, Figure 5 Give an example. Figure 5 This is only an example and the embodiments are not limited thereto. Figure 5 FIG. 4 is a schematic diagram of generating a plurality of crossover frequencies fx1 to fx3 in an embodiment. Figure 5 In the example, the parameter J can be 3. In other words, there are three observation points P1 to P3 in this example. The self-impedances Z11, Z22 and Z33 of the three observation points P1 to P3 can be represented by three curves C1, C2 and C3 in the two-dimensional space of impedance versus frequency. Figure 5 As shown, the intersection points of the three curves C1, C2 and C3 with the target impedance curve corresponding to the target impedance Zt are respectively the intersection points fx1, fx2 and fx3. Therefore, since the power delivery network 100 has J observation points P1 to PJ, a plurality of intersection frequencies fx1 to fxJ can be generated in step 340.

[0080] In step 350, the sum value may be obtained by multiplying the jth weight value wj of the plurality of weight values ​​w1 to wJ and the jth intersection frequency fxj of the plurality of intersection frequencies fx1 to fxJ. Among them, the total value For example, as shown in equation eq-3, the average intersection frequency fx_avg can be generated according to the weight values ​​w1 to wJ generated by equation eq-1 and the intersection frequencies fx1 to fxJ generated in step 340.

[0081]

[0082] In step 360, a group of candidate capacitors C1 to CD may be selected from the plurality of capacitors in the capacitor bank according to the average crossover frequency fx_avg generated by equation eg-3. The self-resonance frequency (SRF) of each of the candidate capacitors C1 to CD may be between a times and b times the average crossover frequency fx_avg, where a and b are parameters greater than zero, a is less than b, a may be between 0.1 and 5, and b may be between 1 and 10. Figure 6 is a schematic diagram of the range of self-resonant frequencies of candidate capacitors. Figure 6 As shown, the self-resonant frequency of the candidate capacitor may be within a range R. According to an embodiment, a may be equal to 0.5, and b may be equal to 2, however, a and b may be adjusted according to requirements, and the larger the range R, the more the number of candidate capacitors may be.

[0083] If only one candidate capacitor is selected in step 360 (ie, D is equal to 1), steps 365 and 370 can be omitted and the process proceeds to step 380. If a plurality of candidate capacitors are selected in step 360 (ie, D is greater than 1), steps 365 and 370 can be executed.

[0084] In steps 365 and 370, an optimal capacitor may be selected from the candidate capacitors C1 to CD, as described below. In step 365, the D candidate capacitors C1 to CD may be respectively set at the optimal capacitor positions determined in step 320 to generate an adjusted self-impedance Zjj' for the jth observation point Pj, as described in equation eq-4:

[0085]

[0086] For example, if D is equal to 5 (that is, there are five candidate capacitors C1 to C5), then according to formula eq-4, five adjusted self-impedances can be generated for each observation point. For example, for the j-th observation point Pj, five adjusted self-impedances Zjj'_1, Zjj'_2, Zjj'_3, Zjj'_4 and Zjj'_5 can be generated.

[0087] In equation eq-4, Zjj may be the original self-impedance of the j-th observation point Pj. Zd may be the impedance of the d-th candidate capacitor Cd among the candidate capacitors C1 to CD, and Zij and Zji may be elements of the impedance matrix (called Z-matrix) of the power delivery network 100. In other words, equation eq-4 may generate the adjusted impedance of the observation point generated after each of the candidate capacitors C1 to CD is coupled in parallel to the power delivery network 100.

[0088] As shown in equation eq-4, the jth observation point Pj can generate D adjusted self-impedances Zjj'_1 to Zjj'_D, so the J observation points can correspond to J×D adjusted self-impedances Z11'_1 to ZJJ'_D. Among them, the adjusted self-impedances Z11'_1 to ZJJ'_D can include:

[0089] Z11'_1 to Z11'_D,

[0090] Z22'_1 to Z22'_D,

[0091] Z33'_1 to Z33'_D... and

[0092] ZJJ'_1 to ZJJ'_D, there are J×D adjusted self-impedances in total.

[0093] In step 370, an optimal capacitor may be selected from the candidate capacitors C1 to CD according to the plurality of adjusted self-impedances Z11'_1 to ZJJ'_D. The process may be as follows: Figure 7 shown. Figure 7 for Figure 3 In step 370, a flow chart for selecting an optimal capacitor may include the following steps.

[0094] Step 3710: Generate a plurality of impedance versus frequency curves (hereinafter referred to as C11 to CJD) according to a plurality of adjusted self-impedances Z11′_1 to ZJJ′_D;

[0095] Step 3720: Generate a plurality of slope absolute value sum values ​​(hereinafter referred to as m11 to mJD) according to the plurality of impedance versus frequency curves C11 to CJD;

[0096] Step 3730: Generate a plurality of reference values ​​according to a plurality of slope absolute value sums m11 to mJD and a plurality of weight values ​​w1 to wJ; and

[0097] Step 3740: Select the smallest capacitor corresponding to the majority of reference values ​​from among the candidate capacitors C1 to CD as the optimal capacitor.

[0098] The sum of the absolute values ​​of the slopes described in steps 3710 and 3720 will be referred to as Figure 8 For example, in step 3730, weight values ​​w1 to wJ may be used to perform weighted calculation on the slope absolute value sum values ​​m11 to mJD to generate a plurality of reference values. Wherein, wj may be multiplied by mj1 to mjD, respectively. Therefore, a plurality of reference values ​​described in step 3730 may be generated:

[0099] w1×m11, w1×m12, …, w1×m1D,

[0100] w2×m21, w2×m22, …, w2×m2D,

[0101] …、

[0102] wJ×mJ1, wJ×mJ2, …, wJ×mJD, a total of J×D reference values.

[0103] In step 3730, by using weight values ​​w1 to wJ, the slope absolute value sum values ​​m11 to mJD are weighted, and the importance of observation points P1 to PJ can be considered. In step 3740, by selecting the capacitor corresponding to the smallest of the plurality of reference values, the capacitor that makes the impedance curve of the power delivery network 100 flattest can be selected. The principle can be as follows: Figure 8 As described in the examples.

[0104] Figure 8 FIG. 4 is a schematic diagram of a curve Cjd among a plurality of impedance versus frequency curves C11 to CJD in an embodiment. Figure 8 As an example, it is used to illustrate how to generate the sum of the absolute values ​​of the slopes in steps 3710 to 3720, and to illustrate that the smaller the sum of the absolute values ​​of the slopes is, the flatter the impedance curve is.

[0105] Figure 8 In the figure, the sum of the absolute values ​​of the slopes of the curve Cjd can be expressed as mjd. Figure 8 In , points Min1, Min2, Min3 ... to Min8 may correspond to local minimums, and points Max1, Max2, Max3 ... to Max8 may correspond to local maximums. Therefore, line segments L1, L2, L3 ... to L16 may be formed between points Min1 to Min8 and points Max1 to Max8. Among them, L1, L3, L5, L7 ... to L15 may be descending line segments, and L2, L4, L6, L8 ... to L16 may be ascending line segments. In Figure 8 In the example of , each of the line segments L1, L2, L3, ... to L16 may have a slope, wherein the slope of the kth line segment may be mk, where k is an integer greater than 0. Figure 8 In the example, 1≤k≤16, Figure 8 The number of line segments is 16, which is an example and the embodiment is not limited thereto. Figure 8 In the equation eq-5, the absolute values ​​of the slopes m1 to m16 can be taken and added to generate the slope absolute value sum mjd, which can be described as follows:

[0106] mid=∑ k＝1 |mk|=|m1|+|m2|+|m3|+......eq-5;

[0107] The smaller the slope absolute value sum mjd is, the flatter the corresponding curve Cjd is in the two-dimensional space of impedance versus frequency. Figure 7 In steps 3730 and 3740, the smallest capacitance corresponding to a plurality of reference values ​​may be selected as the optimal capacitance.

[0108] Figure 8 In the above, the number of points corresponding to the local minimum and the local maximum, and the number of line segments between the points corresponding to the local minimum and the local maximum, may depend on the accuracy of measuring the observation curve Cjd. If the accuracy is high, the number of line segments may be large, and if the accuracy is low, the number of line segments may be small. However, the principle that the lower the sum of the absolute values ​​of the slopes, the flatter the curve, can still be applied to determine the flatness of the curve.

[0109] In summary, the embodiment provides an impedance adjustment system 200 and an impedance adjustment method 300, which can select the best capacitor position from the plurality of capacitor positions Pc1 to PcI of the power delivery network 100, select a group of candidate capacitors C1 to CD from the plurality of capacitors in the capacitor library according to the average crossover frequency fx_avg, and then select the best capacitor from the candidate capacitors C1 to CD according to the flatness of the impedance curve to couple to the best capacitor position. If the impedance of the power delivery network 100 is still too high after the best capacitor is coupled to the best capacitor position, a loop process (such as Figure 4 In step 395), the current best capacitor and the best capacitor position are further selected to couple the current best capacitor to the current best capacitor position until the impedance curve of the power delivery network 100 is adjusted to be not too high and sufficiently flat. Therefore, the embodiment is helpful for adjusting the impedance of the power delivery network 100, improving the hardware design of the power delivery network 100, and avoiding voltage disturbance and distortion.

[0110] The impedance adjustment system 200 and the impedance adjustment method 300 of the embodiment can be applied to the hardware design of the printed circuit board of the server, so that the server has a high-quality power delivery network to enhance the stability of the server, so that the server can be applied to artificial intelligence (AI) computing, edge computing, and can also be applied to 5G communication servers, cloud servers, or Internet of Vehicles servers.

[0111] The above description is only a preferred embodiment of the present invention. All equivalent changes and modifications made according to the scope of the patent application of the present invention should fall within the scope of the present invention.

Claims

1. An impedance adjustment method, It is characterized in that Applicable to a power delivery network, the power delivery network comprising a plurality of observation points, the impedance adjustment method comprising: Generating a corresponding weight value according to a current value of each of the plurality of observation points; Determine an optimal capacitor position from among a plurality of capacitor positions of the power delivery network according to the weight value corresponding to each of the plurality of observation points; generating a plurality of curves according to a plurality of self-impedances of the plurality of observation points; generating a plurality of intersection frequencies according to the plurality of curves and a target impedance curve corresponding to a target impedance; Generating an average intersection frequency according to the weight value corresponding to each of the plurality of observation points and the plurality of intersection frequencies; selecting a group of candidate capacitors from the plurality of capacitors according to the average crossover frequency; generating a plurality of adjusted self-impedances according to the set of candidate capacitors and the optimal capacitor positions; and An optimal capacitor is selected from the set of candidate capacitors based on the plurality of adjusted self-impedances.

2. The impedance adjustment method according to claim 1, It is characterized in that Also includes: placing the optimal capacitor into the optimal capacitor position; updating the plurality of adjusted self-impedances; When at least one of the updated plurality of adjusted self-impedances is greater than the target impedance, a current optimal capacitor position is determined from other capacitor positions of the plurality of capacitor positions except the optimal capacitor position according to at least the plurality of weight values.

3. The impedance adjustment method according to claim 1, It is characterized in that The method of generating the average intersection frequency according to the weight value corresponding to each of the plurality of observation points and the plurality of intersection frequencies comprises: A sum value is obtained according to the product of a j-th weight value and a j-th intersection frequency of the plurality of intersection frequencies, wherein the sum value is positively correlated to the average intersection frequency, j is a positive integer greater than 0, and j is less than or equal to the number of the plurality of observation points.

4. The impedance adjustment method according to claim 1, It is characterized in that A self-resonant frequency of each of the group of candidate capacitors is between a times and b times the average intersection frequency, wherein a and b are parameters greater than zero, a is less than b, a is between 0.1 and 5, and b is between 1 and 10.

5. The impedance adjustment method according to claim 4, It is characterized in that Where a is equal to 0.5 and b is equal to 2.

6. The power delivery network impedance adjustment method as claimed in claim 1, It is characterized in that The method further comprises: selecting the optimal capacitor from the group of candidate capacitors according to the plurality of adjusted self-impedances, comprising: generating a plurality of impedance versus frequency curves according to the plurality of adjusted self-impedances; According to the plurality of impedance versus frequency curves, a plurality of slope absolute value sum values ​​are generated; Generate a plurality of reference values ​​according to the plurality of slope absolute value sums and the plurality of weight values; and From the group of candidate capacitors, a smallest capacitor corresponding to the plurality of reference values ​​is selected as the optimal capacitor.

7. An impedance adjustment system, It is characterized in that Applicable to a power delivery network, the power delivery network comprising a plurality of observation points, the impedance adjustment system comprising: a capacitor position selection unit, configured to generate a corresponding weight value according to a current value of each of the plurality of observation points, and determine an optimal capacitor position from among a plurality of capacitor positions of the power delivery network according to the weight value corresponding to each of the plurality of observation points; and A capacitor selection unit is used to generate a plurality of curves according to the plurality of self-impedances of the plurality of observation points, generate a plurality of intersection frequencies according to the plurality of curves and a target impedance curve, generate an average intersection frequency according to the weight values ​​corresponding to each of the plurality of observation points and the plurality of intersection frequencies, select a group of candidate capacitors from the plurality of capacitors according to the average intersection frequency, generate a plurality of adjusted self-impedances according to the group of candidate capacitors and the optimal capacitor position, and select an optimal capacitor from the group of candidate capacitors according to the plurality of adjusted self-impedances.

8. The impedance adjustment system according to claim 7, It is characterized in that The capacitor position selection unit generates the plurality of weight values ​​according to a plurality of currents respectively corresponding to the plurality of observation points.

9. The impedance adjustment system according to claim 7, It is characterized in that A self-resonant frequency of each of the group of candidate capacitors is between a times and b times the average intersection frequency, wherein a and b are parameters greater than zero, a is less than b, a is between 0.1 and 5, and b is between 1 and 10.

Images