JP2715887B2 - Crosstalk analysis method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method of analyzing crosstalk of a circuit board, and more particularly to a method of analyzing crosstalk of a circuit board which can be operated at high speed.

[0002]

2. Description of the Related Art Recently, noises such as signal delay time, signal reflection, crosstalk, etc. have become more important in a mounted circuit board such as a printed circuit board as the density and speed have increased.

[0003] As one of these noise countermeasures, there is a multi-chip module (MCM) which is a technique for forming one module using a plurality of chips.

[0004] As another method, there is a method of utilizing a simulation for design. By analyzing a transmission line using a finite element method (FEM) or a boundary element method (BEM), characteristic impedance, crosstalk, and the like are obtained. Is determined, and a design is made based on these values.

[0005] For example, in the literature (Matsui et al., “Electric Characteristics Simulator for Printed Plates,” NTT R & D vol. 38n
o. 10 (1989) pp 1211-1220)
A typical prior art example of a simulation-based design method is disclosed.

In the design CAD using the above-described conventional technology, when analyzing characteristic impedance, crosstalk, etc., first, a finite element method (FEM) and a boundary element method (BEM)
And the like, and then the characteristic impedance, crosstalk, and the like are determined from these values by using TEM (transverse electromagnetic wave mode) approximation to perform the design.

[0007]

However, when the analysis is performed using the above-mentioned conventional design CAD software, there are the following problems.

(1) In the analysis by the finite element method (FEM), at least 100%
The above number of nodes is required. For this, 50MIP
Even when a high-speed EWS (Engineering Workstation) of about S (Million Instruction Per Second) is used, analysis time of 6 seconds or more is required. Further, even if the boundary element method (BEM) is used, almost the same analysis time is required.

(2) In the case of crosstalk analysis, it is necessary to perform an analysis for each pair of nets (referred to as a “pair net”). Therefore, for N nets, N × (N−1) / 2 analyzes are performed. Is required. Therefore, when performing a crosstalk analysis on, for example, 400 nets, it takes about one day or more of calculation time even if a high-speed processing device is used.

From the above, it is practically impossible to apply the crosstalk analysis using the conventional technique to an actual design.

Accordingly, an object of the present invention is to solve the above problems and to extract the number of pairs of nets to be analyzed based on various wiring parameters in crosstalk analysis using a finite element method (FEM) and TEM approximation. An object of the present invention is to provide a crosstalk analysis method capable of high-speed operation.

[0012]

In order to achieve the above object, the present invention provides a wiring design CAD used for designing a circuit board.
In software, a first pair net group is extracted from the layout design data of the circuit board, and a minimum value Smin of the wiring interval among the pair nets constituting the first pair net group is a predetermined wiring interval limit. A second pair net group having a value Sc or less is extracted, and a first division is performed on the wiring of the pair nets constituting the second pair net group,
The wiring aspect ratio (X = ΣL / L) is obtained by obtaining the sum of each part for the first division, where S is the wiring interval, W is the wiring width, and L is the wiring length in each part obtained in the first division.
(S + W)), and the third wiring aspect ratio having a value larger than a predetermined wiring aspect ratio limit value Xc is determined.
Is extracted, and a second division is performed on the wiring of the paired nets constituting the third paired netting. The wirings of the paired nets of the respective parts obtained by the second division are in the same direction and Reverse current mode characteristic impedance Z
ne, Zno is calculated, and the approximate value Cr of the crosstalk calculated using the characteristic impedance in the same direction and the reverse current mode is equal to or greater than a predetermined limit value Cr0 of the approximate value of the crosstalk. A paired net group is extracted, and the capacity of each paired net constituting the fourth paired net group is calculated using the finite element method (FEM).
A crosstalk analysis method characterized by determining crosstalk between wirings using a crosstalk analysis value calculated by EM approximation.

[0013]

SUMMARY OF THE INVENTION The wiring design CAD software according to the present invention has the following crosstalk analysis function.

(Step 1) One pair net is extracted from the layout design data, and the minimum value Smin of the wiring interval in the net pair is extracted. Next, if Smin is equal to or greater than a predetermined constant value Sc, the process returns to the beginning to extract a new pair net, and if Smin is equal to or less than Sc, the next step 2 is performed.

(Step 2) Next, when the wiring constituting the paired net is divided into several parts, and the wiring spacing at each part is S, the wiring width is W, and the wiring length is L, the wiring aspect ratio X is calculated by the following equation (5). X = ΣL / (S + W) (5)

Here, Σ is the sum of each part. If the wiring aspect ratio X is larger than the fixed value Xc in the pair net, the next step 3 is performed. If X is equal to or smaller than Xc, a new pair net is extracted and executed from the beginning. Note that the division of the wiring means that the two wirings forming the net pair are divided into each area when the width or the interval is different in each part.

(Step 3) Next, the wiring width at each part of the wiring constituting the paired net is w, the wiring interval is s, the interlayer film thickness is h, the wiring film thickness is t, the wiring length is 1, the wiring length is 1, The relative permittivity is assumed to be εre, and an approximate value Cr of the crosstalk is obtained from Expressions (7) to (10) described later.

The approximate value Cr of the crosstalk is a constant value Cr0.
In the above case, the next step 4 is executed. If Cr is less than Cr0, a new pair net is extracted and executed from the beginning.

(Step 4) Next, capacity calculation is performed for each pair net using FEM. Then, the crosstalk is calculated from the calculation result of the capacitance by using the TEM approximation.

[0020]

Embodiments of the present invention will be described below with reference to the drawings.

A printed circuit board design method according to an embodiment of the present invention will be described. FIG. 1 shows a design flow of a printed circuit board.

In designing a printed circuit board, as shown in FIG. 1, first, a circuit is designed (100), and a net list is extracted from the result (110).

Next, the layout of the LSI chip on the printed circuit board is designed in consideration of the netlist (120).

Next, the wiring on the printed circuit board is automatically or manually designed according to the netlist (layout design of 130).

Next, one net pair is extracted from the layout data designed as described above. Then, the crosstalk between the wirings of this net pair is analyzed (140), and the next paired net is extracted if it is below the allowable criterion, and redesigned if it is above the criterion value.

The crosstalk analysis portion of this embodiment for the above-described printed circuit board design will be described in more detail with reference to FIG. FIG. 2 is a crosstalk analysis part (1) of FIG.
40 is a flowchart in which the processing procedure of (40) is described in detail;

As shown in FIG. 2, the crosstalk analysis part is composed of four steps. Hereinafter, each step will be described.

(Step 1) First, one set of pair nets is extracted from the layout design data, and the minimum value Smin of the wiring interval in the pair nets is calculated (210).

Next, if the minimum value Smin is equal to or larger than the predetermined constant value Sc, the process returns to the beginning and a new pair net is extracted. If the minimum value Smin is equal to or smaller than Sc, the next step 2 is executed. In this embodiment, the value of Sc is 100 μm
Is set.

(Step 2) Next, the wiring constituting the net pair is divided into several parts, and the wiring intervals at each part are S,
When the wiring width is W and the wiring length is L, the wiring aspect ratio X is calculated by the following equation (220). X = ΣL / (S + W) (6)

Here, Σ is the sum of each part. If X is larger than the predetermined value Xc, the next step 3 is executed. If X is equal to or smaller than Xc, a new pair net is extracted and executed from the beginning. In this embodiment, 2 × 10 ³ is used as Xc.

(Step 3) Next, the wiring width at each part of the wiring constituting the paired net is w, the wiring interval is s, the interlayer thickness is h, the wiring thickness is t, and the wiring length is l. Εre, the rise time of the input pulse is tr,
The approximate value Cr of the crosstalk is obtained from the following equation (7) using equation (10) (230).

Zne in equation (9) is an approximate expression of the characteristic impedance of a mode in which current flows in the same direction through two wires (microstrip lines) (referred to as “even mode”). Zno is an approximate expression of the characteristic impedance in a mode in which currents in opposite directions flow through two wires (referred to as “odd mode”).

[0034]

(Equation 2)

If the crosstalk approximate value Cr is equal to or greater than a predetermined fixed value Cr0, the next step 4 is executed. If the crosstalk approximate value Cr is equal to or less than Cr0, a new pair net is extracted and executed from the beginning. In this embodiment, the value of Cr0 is set to 2.5%.

(Step 4) Next, the capacity is calculated for each pair net using the finite element method (FEM) (24).
0).

The calculation of the capacity will be described below in detail with reference to FIG. FIG. 3 is an explanatory diagram of a cross section of a wiring forming a net pair. As shown in FIG.
Wiring 301, second wiring 302, ground electrode 30
3 and an interlayer insulating film 304.

The capacitance between each wiring is determined by the finite element method (FEM).
Analyze using. First, the analysis region shown in FIG. 3 is mesh-divided into several elements. Next, a matrix equation was assembled by simultaneously combining the element equations, a potential was applied to each wiring as a boundary condition, and the potential distribution was analyzed. The charge distribution of each electrode is obtained from the analysis result, and the two-dimensional capacitance Cij between the wirings is obtained.
Is calculated.

Next, the inductance ｛L is obtained by TEM approximation.
ij} (see equation (11)), and then the crosstalk coefficient Ki is calculated by equation (12).

[0040]

(Equation 3)

From this, the crosstalk analysis value Xcr is calculated using Xcr = ΣKi · l (1 is the wiring length). Here, Σ represents the sum regarding each divided area.

If the crosstalk analysis value Xcr thus obtained is 2% or less, the next pair net is extracted,
If it is 2% or more, it is determined that crosstalk is defective, and redesign is performed.

In the capacity analysis using the finite element method, as described above, the number of nodes required to obtain an analysis accuracy of 10% is about 10 nodes.
It is set to 0 or more. In this case, the CPU time required for automatically generating a mesh and analyzing a determinant is, for example, 50 MIPS.
EWS of 6 seconds or more.

Therefore, in the prior art in which all pair nets are analyzed by the finite element method (FEM), about 1 minute of CPU time is required for each pair net pair, and all pair nets (when the number of nets is 400 When the crosstalk determination was made for 400 × 399/2 and about 80,000 combinations), it took about one day and was not suitable for practical use at all.

On the other hand, when the design method of the present invention is used, in Step 1 of extracting a pair net having a minimum value of the wiring interval equal to or less than a predetermined value, about 50%, and a pair having a wiring aspect ratio equal to or more than a predetermined value. About 50 in step 2 to extract the net
%, And about 60% of the pair nets in the step 3 for extracting the pair nets whose crosstalk approximate value is equal to or greater than a predetermined value,
o determination, that is, it was determined that crosstalk was not a problem (good crosstalk).

Each of the determinations in steps 1 to 3 is made only by mathematical formula calculation, and the calculation time is F
It can be ignored compared to the analysis time of EM.

In the method of the present invention, about 10
%, Only the paired nets need to be analyzed by FEM, so the total CPU time is about 3 hours, which is 1 /
8 was greatly reduced. Further, when the crosstalk was determined to be non-defective in the determinations in steps 1 to 3, the crosstalk analysis was further performed, and no defective product was found.

[0048]

As described above, according to the method of the present invention, for a pair net (step 1) in which the minimum value of the wiring interval is equal to or less than a predetermined value, a pair net having a wiring aspect ratio equal to or more than a predetermined value (step 1). 2) Further, a pair net whose crosstalk approximate value is equal to or greater than a predetermined value is extracted (step 3), and only the extracted pair net is subjected to the finite element method (FE).
By calculating the capacitance using M) and determining using the crosstalk analysis value calculated by TEM approximation, the crosstalk of each wiring of the circuit board can be designed to be a certain value or less without reducing the analysis accuracy at all. This has the effect of significantly reducing the calculation time for performing the calculation.

In particular, according to the present invention, the pair nets determined to be non-defective in the above steps 1 to 3 are not analyzed as crosstalk failures.
It can be seen that the present invention significantly improves the analysis efficiency while maintaining the analysis accuracy.

Describing quantitatively the effect of the present invention, according to the processing steps of the present invention, about 10% of net pairs of the entire net pair are extracted, and only the extracted 10% of net pairs are finite elements. By performing the method (FEM) analysis, the calculation time is reduced to about 1/8 of the conventional example.

[Brief description of the drawings]

FIG. 1 is a flowchart showing the configuration of an embodiment of the present invention.

FIG. 2 is a flowchart illustrating a detailed processing procedure of a crosstalk analysis unit according to an embodiment of the present invention.

FIG. 3 is an explanatory diagram of a wiring cross section of a pair net for performing a capacitance analysis according to the present invention.

[Explanation of symbols]

 301 first wiring 303 second wiring 303 ground electrode 304 interlayer insulating film

Claims (2)

(57) [Claims] 1. A wiring design CAD software used for designing a circuit board, wherein a first paired net group is extracted from layout design data of the circuit board, and a wiring among the paired nets constituting the first paired net group is extracted. A second pair net group whose minimum interval value Smin is equal to or smaller than a predetermined wiring interval limit value Sc is extracted, and a first division is performed on the wiring of the pair nets constituting the second pair net group. The wiring aspect ratio (X = ΣL) is obtained by calculating the sum of each part for the first division, where S is the wiring interval, W is the wiring width, and L is the wiring length in each part obtained in the first division. /
(S + W)), and a third pair net group having a wiring aspect ratio larger than a predetermined wiring aspect ratio limit value Xc is extracted, and the pair nets constituting the third pair net group are extracted. A second division is performed on the wiring, and the characteristic impedances Zne and Zno of the same direction and the reverse current mode are calculated for the wiring of the pair net of each part obtained by the second division, and the same direction and the reverse current are calculated. Extracting a fourth pair net group in which the approximate value Cr of the crosstalk calculated using the characteristic impedance of the mode is equal to or greater than a predetermined limit value Cr0 of the approximate value of the crosstalk; Is calculated using the finite element method (FEM) for each paired net that composes the crosstalk, and the crosstalk between wires is calculated using the crosstalk analysis value calculated by TEM approximation. A crosstalk analysis method characterized in that a crosstalk is determined. 2. The wiring of each part obtained by the second division is as follows: wiring width w, wiring interval s, interlayer film thickness h, wiring film thickness t, wiring length l, ratio of interlayer film. The dielectric constant is εre, the rise time of the input pulse is tr, and the characteristic impedances Zne, Zn in the same-direction and reverse-direction current modes are given.
o is calculated by the following equations (3) and (4), whereby the crosstalk approximate value C is calculated using the equations (2) and (1).
The crosstalk analysis method according to claim 1 , wherein r is calculated. (Equation 1)