High transmission rate optical module circuit board and preparation method thereof

Abstract

The present application relates to the technical field of optical module circuit boards, and specifically discloses a high transmission rate optical module circuit board and a preparation method thereof, which obtains a through-hole filling state image collected by an industrial camera, and adopts image analysis and processing technology based on deep learning to extract the filling shape and structural features of the through-hole filling state image respectively, so as to automatically judge whether there is a defect by selecting the feature representation of the through-hole state structural feature based on the through-hole filling state shape feature as a constraint feature. In this way, the through-hole filling state image is automatically collected by an industrial camera, which can significantly reduce the detection time, and can achieve a faster detection speed compared with manual detection or X-ray and ultrasonic detection of each sample. At the same time, it can identify subtle defects, such as tiny bubbles, cracks, delamination, etc., improve the accuracy and reliability of detection, and realize automated thermal conductive through-hole filling quality detection.

Description

Technical Field

[0001] The present application relates to the technical field of optical module circuit boards, and more specifically, to a high transmission rate optical module circuit board and a preparation method thereof. Background Art

[0002] High transmission rate optical modules are usually used in data centers, cloud computing platforms, high-performance computing and other fields, requiring the circuit board to have extremely high transmission rates and excellent electrical performance. These modules will generate a lot of heat during operation, so the circuit board must have good thermal management capabilities. Chinese patent CN110876225A proposes a circuit board, a method for preparing a circuit board and an optical module with the circuit board, which forms an efficient heat conduction path by opening and filling a through hole with heat-conducting metal in the central sub-stack layer, and opening and copper-plated laser holes in the multi-layer structure on both sides, thereby improving the heat dissipation performance of the circuit board.

[0003] In high-transmission-rate optical module circuit boards, the filling quality of thermal vias is crucial. Thermal vias are not only used for electrical connections, but also for transferring heat. If the filling is uneven or defective, it will lead to thermal management failure, which in turn affects the performance and reliability of the circuit board. However, in the above-mentioned patents, traditional filling quality detection of thermal vias usually relies on manual inspection or detection technology, which has certain drawbacks and defects. Specifically, manual visual inspection is inefficient and easily affected by personal experience and fatigue, resulting in unstable test results, prone to misjudgment or missed detection. Although X-ray and ultrasonic testing can detect internal defects, they usually need to be tested sample by sample, and the detection speed is slow. In addition, these traditional methods have limitations in detecting subtle defects, and it is difficult to detect tiny bubbles, cracks and other problems, which affects the accuracy and reliability of detection.

[0004] Therefore, an optimized method for preparing a circuit board for an optical module with a high transmission rate is desired. Summary of the invention

[0005] The present application provides a high transmission rate optical module circuit board and a preparation method thereof, which can significantly reduce the detection time by automatically collecting through-hole filling status images through an industrial camera, and can achieve a faster detection speed than manual detection or X-ray and ultrasonic detection of each sample. At the same time, it can identify subtle defects such as tiny bubbles, cracks, delamination, etc., improve the accuracy and reliability of detection, and realize automated thermal conductive through-hole filling quality detection.

[0006] In a first aspect, a method for preparing a high transmission rate optical module circuit board is provided, comprising: constructing a central sub-stack layer; opening a thermal conductive through hole on the central sub-stack layer, and filling the thermal conductive through hole with a thermal conductive metal to obtain a filled central sub-stack layer; stacking a prepreg and a copper foil on both sides of the filled central sub-stack layer and then laminating to obtain a pressed sub-stack layer; opening a laser hole at a predetermined position of the pressed sub-stack layer, and copper plating the laser hole to obtain the high transmission rate optical module circuit board;

[0007] Wherein, filling the thermally conductive through hole with a thermally conductive metal to obtain a filled central sub-stacked layer further comprises: acquiring a through hole filling state image acquired by an industrial camera;

[0008] Performing filling shape feature extraction and filling structure feature extraction on the through-hole filling state image respectively to obtain a through-hole filling shape feature map and a through-hole filling structure feature map, and extracting a through-hole filling shape description vector from the through-hole filling shape feature map;

[0009] The through-hole filling shape description vector and the through-hole filling structure feature map are subjected to feature selection based on metadata blocks to obtain a filling shape constrained through-hole filling structure feature, including: performing linear transformation and gradient amplitude-based masking processing on the through-hole filling shape description vector to obtain a masked first through-hole filling shape description meta-modal auxiliary feature and a masked second through-hole filling shape description meta-modal auxiliary feature; based on the masked first through-hole filling shape description meta-modal auxiliary feature and the masked second through-hole filling shape description meta-modal auxiliary feature, feature selection is performed on the through-hole filling structure feature map to obtain the filling shape constrained through-hole filling structure feature;

[0010] Based on the filling shape constrained through-hole filling structural features, a detection result is obtained.

[0011] In a possible implementation, performing filling shape feature extraction and filling structure feature extraction on the through-hole filling state image to obtain a through-hole filling shape feature map and a through-hole filling structure feature map, and extracting a through-hole filling shape description vector from the through-hole filling shape feature map, including:

[0012] Inputting the through-hole filling state image into a filling shape feature extractor based on a first hole convolutional neural network model to obtain the through-hole filling shape feature map;

[0013] Inputting the through-hole filling state image into a filling structure feature extractor based on a second hole convolutional neural network model to obtain the through-hole filling structure feature map;

[0014] A global mean pooling process is performed on the through-hole filling shape feature map to obtain the through-hole filling shape description vector.

[0015] In a possible implementation, performing linear transformation and gradient amplitude-based masking processing on the through hole filling shape description vector to obtain a masked first through hole filling shape description meta-modal auxiliary feature and a masked second through hole filling shape description meta-modal auxiliary feature, including:

[0016] Performing a linear transformation on the through hole filling shape description vector based on a first weight modifier matrix and a second weight modifier matrix to obtain a first through hole filling shape description element modal auxiliary feature mapping mode representation vector and a second through hole filling shape description element modal auxiliary feature mapping mode representation vector;

[0017] Based on the gradient amplitude of each position in the first through hole filling shape description element modal auxiliary feature mapping pattern representation vector, performing gradient amplitude-based mask processing on the first through hole filling shape description element modal auxiliary feature mapping pattern representation vector to obtain a masked first through hole filling shape description element modal auxiliary feature mapping pattern representation vector as the masked first through hole filling shape description element modal auxiliary feature;

[0018] Based on the gradient amplitude of each position in the second through-hole filling shape description meta-modal auxiliary feature mapping pattern representation vector, the second through-hole filling shape description meta-modal auxiliary feature mapping pattern representation vector is subjected to gradient amplitude-based mask processing to obtain a masked second through-hole filling shape description meta-modal auxiliary feature mapping pattern representation vector as the masked second through-hole filling shape description meta-modal auxiliary feature.

[0019] In a possible implementation, performing a linear transformation based on a first weight modifier matrix and a second weight modifier matrix on the through hole filling shape description vector to obtain a first through hole filling shape description element modal auxiliary feature mapping mode representation vector and a second through hole filling shape description element modal auxiliary feature mapping mode representation vector, including:

[0020] Multiplying the through hole filling shape description vector with the first weight modifier matrix, adding the obtained first through hole filling shape description modification vector with the first modifier bias vector by position to obtain the first through hole filling shape description element modal auxiliary feature mapping mode representation vector;

[0021] The through hole filling shape description vector is multiplied by the second weight modifier matrix, and the obtained second through hole filling shape description modification vector is positionally added to the second modifier bias vector to obtain the second through hole filling shape description meta-modal auxiliary feature mapping pattern representation vector.

[0022] In a possible implementation, based on the gradient amplitude of each position in the first through hole filling shape description element modal auxiliary feature mapping pattern representation vector, performing gradient amplitude-based mask processing on the first through hole filling shape description element modal auxiliary feature mapping pattern representation vector to obtain a masked first through hole filling shape description element modal auxiliary feature mapping pattern representation vector, including:

[0023] Calculate the square of the difference between the eigenvalues ​​at the th position and the th position in the first through hole filling shape description element modal auxiliary feature mapping mode representation vector to obtain a first through hole filling shape description gradient amplitude representation vector;

[0024] Calculating the absolute value of the square root of each eigenvalue in the first through hole filling shape description gradient amplitude representation vector to obtain a first through hole filling shape description gradient amplitude modulation representation vector;

[0025] The first through-hole filling shape description gradient amplitude modulation representation vector is masked, and in response to the eigenvalue of each position in the first through-hole filling shape description gradient amplitude modulation representation vector being greater than a predetermined threshold, the eigenvalue of the position is set to 1, and the rest are set to 0, to obtain the masked first through-hole filling shape description meta-modal auxiliary feature mapping pattern representation vector.

[0026] In a possible implementation, based on the masked first through-hole filling shape description meta-modal auxiliary feature and the masked second through-hole filling shape description meta-modal auxiliary feature, feature selection is performed on the through-hole filling structure feature map to obtain the filling shape constrained through-hole filling structure feature, including:

[0027] Performing preliminary feature selection on the through-hole filling structure feature map based on the masked first through-hole filling shape description element modal auxiliary feature mapping pattern representation vector and the hyperbolic tangent function to obtain a through-hole filling structure feature map after preliminary selection;

[0028] Based on the masked second through-hole filling shape description meta-modal auxiliary feature mapping pattern representation vector and Sigmoid function, feature selection is performed again on the through-hole filling structure feature map after the initial selection to obtain a filling shape constrained through-hole filling structure feature map as the filling shape constrained through-hole filling structure feature.

[0029] In a possible implementation, performing preliminary feature selection on the through-hole filling structure feature map based on the masked first through-hole filling shape description element modal auxiliary feature mapping mode representation vector and the hyperbolic tangent function to obtain a through-hole filling structure feature map after preliminary selection, including:

[0030] Multiplying the feature values ​​of each position in the masked first through-hole filling shape description element modal auxiliary feature mapping pattern representation vector by each feature matrix along the channel dimension of the through-hole filling structure feature map by position to obtain a through-hole filling structure weighted feature map;

[0031] The hyperbolic tangent function is used to perform feature selection on the weighted feature map of the via filling structure to obtain the initially selected via filling structure feature map.

[0032] In a possible implementation, performing feature selection again on the initially selected through-hole filling structure feature map based on the masked second through-hole filling shape description element modal auxiliary feature mapping pattern representation vector and Sigmoid function to obtain a filling shape constrained through-hole filling structure feature map, including:

[0033] Multiplying the feature values ​​of each position in the masked second through-hole filling shape description element modal auxiliary feature mapping pattern representation vector by each feature matrix along the channel dimension of the through-hole filling structure feature map after preliminary selection by position to obtain a weighted feature map of the through-hole filling structure after preliminary selection;

[0034] The weighted feature map of the through-hole filling structure after the preliminary selection is processed using a Sigmoid function to obtain the filling shape constrained through-hole filling structure feature map.

[0035] In one possible implementation, a detection result is obtained based on the filling shape constrained through-hole filling structure feature, including: inputting the filling shape constrained through-hole filling structure feature map into a classifier-based filling quality detector to obtain a detection result, wherein the detection result is used to indicate whether there is a defect.

[0036] In a second aspect, a high transmission rate optical module circuit board is provided, wherein the high transmission rate optical module circuit board is manufactured by the method for manufacturing the high transmission rate optical module circuit board described in the first aspect.

[0037] The present application provides a high transmission rate optical module circuit board and a preparation method thereof, which obtains a through-hole filling state image collected by an industrial camera, and uses image analysis and processing technology based on deep learning to extract the filling shape and structural features of the through-hole filling state image respectively, so as to automatically judge whether there is a defect by selecting the feature representation of the through-hole state structural feature based on the through-hole filling state shape feature as a constraint feature. In this way, the through-hole filling state image is automatically collected by an industrial camera, which can significantly reduce the detection time, and can achieve a faster detection speed than manual detection or X-ray and ultrasonic detection of each sample. At the same time, it can identify subtle defects, such as tiny bubbles, cracks, delamination, etc., improve the accuracy and reliability of detection, and realize automated thermal conductive through-hole filling quality detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings of the embodiments of the present application are briefly introduced below. Obviously, the drawings described below only relate to some embodiments of the present application, and are not intended to limit the present application.

[0039] Figure 1 It is a schematic flow chart of a method for preparing a high transmission rate optical module circuit board according to an embodiment of the present application.

[0040] Figure 2 This is a schematic flow chart of filling the thermally conductive through hole with a thermally conductive metal to obtain a filled central sub-stacked layer in the method for preparing a high transmission rate optical module circuit board according to an embodiment of the present application.

[0041] Figure 3 In the method for preparing the high transmission rate optical module circuit board of the embodiment of the present application, a heat-conducting metal is filled into the heat-conducting through hole to obtain a data flow diagram of the filled central sub-stacked layer.

[0042] Figure 4 In the preparation method of the high transmission rate optical module circuit board of the embodiment of the present application, the filling shape feature extraction and the filling structure feature extraction are respectively performed on the through-hole filling state image to obtain a through-hole filling shape feature map and a through-hole filling structure feature map, and a schematic flow chart of extracting the through-hole filling shape description vector from the through-hole filling shape feature map is provided.

[0043] Figure 5 In the method for preparing a high transmission rate optical module circuit board according to an embodiment of the present application, a schematic flowchart of performing feature selection based on metadata blocks on the through-hole filling shape description vector and the through-hole filling structure feature map to obtain filling shape constrained through-hole filling structure features. DETAILED DESCRIPTION

[0044] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work also fall within the scope of protection of the present application.

[0045] High-transmission-rate optical modules play an important role in data centers, cloud computing platforms, and high-performance computing applications. They require high-speed data transmission and excellent electrical properties from circuit boards. At the same time, due to the large amount of heat generated during operation, the circuit boards required for these modules must also have an excellent thermal management system to ensure the stable operation of the equipment. Chinese patent CN110876225A introduces an improved design scheme that constructs an efficient heat conduction path by setting and filling through holes of thermally conductive metal in the central sub-stack layer of the circuit board, and copper plating in the multi-layer structure on both sides using laser opening, thereby enhancing the thermal dissipation performance of the circuit board.

[0046] For circuit boards in such high-transmission-rate optical modules, the quality of filling thermal vias is particularly critical. These vias not only serve as electrical connections, but are also an important way for heat conduction. If the filling process is uneven or defective, it may cause the failure of the heat management mechanism, ultimately affecting the functional performance and reliability of the entire circuit board. However, in the traditional methods mentioned in the patent, the filling quality assessment of thermal vias often relies on manual inspection or specific technical means, and these methods have their shortcomings. For example, manual inspection methods are inefficient and easily affected by individual differences and fatigue of operators, resulting in inconsistent inspection results, possible misjudgments or omissions. Although X-ray inspection and ultrasonic inspection can detect internal defects, these two methods usually require each sample to be tested separately, so the detection rate is limited. In addition, for subtle defects such as tiny bubbles or cracks, these traditional methods may not be sensitive enough to detect and cannot guarantee the accuracy and reliability of the test results.

[0047] Based on this, the present application proposes a method for preparing a high transmission rate optical module circuit board. Figure 1 FIG. 1 is a schematic flow chart of a method for preparing a high transmission rate optical module circuit board according to an embodiment of the present application. Figure 1As shown, the preparation method of the high transmission rate optical module circuit board includes: S1, constructing a central sub-stack layer; S2, opening a thermal through hole on the central sub-stack layer, and filling the thermal through hole with a thermally conductive metal to obtain a filled central sub-stack layer; S3, stacking a semi-cured sheet and a copper foil on both sides of the filled central sub-stack layer and then laminating to obtain a pressed sub-stack; S4, opening a laser hole at a predetermined position of the pressed sub-stack, and copper-plating the laser hole to obtain the high transmission rate optical module circuit board.

[0048] Accordingly, in filling the thermally conductive through hole with a thermally conductive metal to obtain the filled central sub-stacked layer, the technical concept of the present application is to obtain a through-hole filling state image captured by an industrial camera, and use image analysis and processing technology based on deep learning to extract the filling shape and structural features of the through-hole filling state image respectively, so as to automatically determine whether there is a defect by selecting the through-hole state structural features based on the through-hole filling state shape features as constraint features. In this way, the detection time can be significantly reduced by automatically collecting through-hole filling state images with industrial cameras, and a faster detection speed can be achieved compared to manual inspection or X-ray and ultrasonic inspection of each sample. At the same time, it is possible to identify subtle defects such as tiny bubbles, cracks, delamination, etc., improve the accuracy and reliability of detection, and realize automated thermal conductive through-hole filling quality detection.

[0049] Figure 2 This is a schematic flow chart of filling the thermally conductive through hole with a thermally conductive metal to obtain a filled central sub-stacked layer in the method for preparing a high transmission rate optical module circuit board according to an embodiment of the present application. Figure 3 The method for preparing the high transmission rate optical module circuit board of the embodiment of the present application is to fill the thermal conductive through hole with thermal conductive metal to obtain a data flow diagram of the filled central sub-stacked layer. Figure 2 and Figure 3 As shown, filling the thermally conductive through hole with a thermally conductive metal to obtain a central sub-stacked layer after filling, further comprising: S21, acquiring a through hole filling state image acquired by an industrial camera; S22, performing filling shape feature extraction and filling structure feature extraction on the through hole filling state image respectively to obtain a through hole filling shape feature map and a through hole filling structure feature map, and extracting a through hole filling shape description vector from the through hole filling shape feature map; S23, performing feature selection based on metadata blocks on the through hole filling shape description vector and the through hole filling structure feature map to obtain a filling shape constrained through hole filling structure feature; S24, obtaining a detection result based on the filling shape constrained through hole filling structure feature.

[0050] In the above-mentioned method for preparing a high transmission rate optical module circuit board, the S21, obtains a through-hole filling state image collected by an industrial camera. It should be understood that by automatically collecting through-hole filling state images by an industrial camera, the through-hole filling situation can be obtained quickly and efficiently, and then these images are processed using image analysis technology based on deep learning, which can quickly identify filling defects such as tiny bubbles, cracks or delamination, etc., improve the accuracy and reliability of detection, and realize automated thermal conductive through-hole filling quality detection, thereby ensuring the quality and performance of the circuit board. This method can significantly reduce detection time and improve detection efficiency.

[0051] In the above-mentioned method for preparing a high transmission rate optical module circuit board, the S22 is to extract the filling shape feature and the filling structure feature of the through-hole filling state image to obtain a through-hole filling shape feature map and a through-hole filling structure feature map, and extract a through-hole filling shape description vector from the through-hole filling shape feature map. Optionally, in one embodiment of the present application, Figure 4 In the method for preparing a high transmission rate optical module circuit board of the embodiment of the present application, the filling shape feature extraction and the filling structure feature extraction are respectively performed on the through hole filling state image to obtain a through hole filling shape feature map and a through hole filling structure feature map, and a schematic flow chart of extracting a through hole filling shape description vector from the through hole filling shape feature map. Figure 4 As shown, filling shape feature extraction and filling structure feature extraction are respectively performed on the through-hole filling state image to obtain a through-hole filling shape feature map and a through-hole filling structure feature map, and a through-hole filling shape description vector is extracted from the through-hole filling shape feature map, including: S221, inputting the through-hole filling state image into a filling shape feature extractor based on a first void convolutional neural network model to obtain the through-hole filling shape feature map; S222, inputting the through-hole filling state image into a filling structure feature extractor based on a second void convolutional neural network model to obtain the through-hole filling structure feature map; S223, performing global mean pooling processing on the through-hole filling shape feature map to obtain the through-hole filling shape description vector.

[0052] That is, considering that the hole filling shape in the through-hole filling state image is crucial to ensuring the quality of the circuit board, such as whether the size and shape of the through-hole are in compliance. Based on this, in order to capture and extract key through-hole shape information to improve the accuracy of subsequent defect detection, in the technical solution of the present application, the through-hole filling state image is input into a filling shape feature extractor based on a first hole convolutional neural network model to obtain a through-hole filling shape feature map. That is, the shape features of the through-hole filling can help detect whether the filling is uniform and complete. If the filling is insufficient or there are problems such as bubbles, the electrical performance of the circuit board may be reduced, and even cause open circuit or short circuit failures in extreme cases.

[0053] Then, considering the structural features in the through-hole filling state image, such as the density and uniformity of the filling metal in the through-hole, it is also crucial for through-hole defect detection. Therefore, in the technical solution of the present application, the through-hole filling state image is input into the filling structure feature extractor based on the second hole convolutional neural network model to capture and extract the key filling structure features, and obtain the through-hole filling structure feature map.

[0054] Next, in order to extract the key shape description feature information in the through-hole filling shape feature map and reduce the interference of irrelevant information and noise, in the technical solution of the present application, the through-hole filling shape feature map is subjected to global mean pooling processing to obtain a through-hole filling shape description vector. This can remove some local noise or unimportant detail information, making the obtained features more representative and able to focus more on global information.

[0055] In the above-mentioned method for preparing a high-transmission-rate optical module circuit board, the S23 performs feature selection based on metadata blocks on the through-hole filling shape description vector and the through-hole filling structural feature map to obtain the filling shape constraint through-hole filling structural features. It should be understood that, considering that the through-hole filling shape description vector is used to describe the shape characteristics of the through-hole filling, including but not limited to information such as the outline, edge, size and shape of the through-hole, an overall description of the through-hole filling shape is provided, including outline and edge features. The through-hole filling structural feature map reflects the specific distribution of the filler inside the through-hole, including the distribution density of the filler, the relationship between the filler and the through-hole wall, and provides detailed information about the internal structure of the filler, including the distribution and potential defects of the filler. Moreover, there is a mutual correlation and influence between the two. If the through-hole filling shape description indicates that there are irregular edges in the filling, then in the through-hole filling structural features, those features that reflect edge information become particularly important. Therefore, in order to perform feature selection on the through-hole filling structure feature map based on the through-hole filling shape description information, in the technical solution of the present application, feature selection based on the metadata block is performed on the through-hole filling shape description vector and the through-hole filling structure feature map to obtain a filling shape constrained through-hole filling structure feature map as a filling shape constrained through-hole filling structure feature.

[0056] Optionally, in one embodiment of the present application, Figure 5 This is a schematic flow chart of performing feature selection based on metadata blocks on the through-hole filling shape description vector and the through-hole filling structure feature map in the method for preparing the high transmission rate optical module circuit board of the embodiment of the present application to obtain the filling shape constraint through-hole filling structure feature. Figure 5As shown, the through-hole filling shape description vector and the through-hole filling structure feature map are subjected to feature selection based on metadata blocks to obtain the filling shape constrained through-hole filling structure feature, including: S231, linear transformation and gradient amplitude-based masking processing are performed on the through-hole filling shape description vector to obtain a masked first through-hole filling shape description meta-modal auxiliary feature and a masked second through-hole filling shape description meta-modal auxiliary feature; S232, based on the masked first through-hole filling shape description meta-modal auxiliary feature and the masked second through-hole filling shape description meta-modal auxiliary feature, feature selection is performed on the through-hole filling structure feature map to obtain the filling shape constrained through-hole filling structure feature;

[0057] Optionally, in one embodiment of the present application, the through-hole filling shape description vector is subjected to a linear transformation and a gradient amplitude-based masking process to obtain a masked first through-hole filling shape description meta-modal auxiliary feature and a masked second through-hole filling shape description meta-modal auxiliary feature, including: performing a linear transformation on the through-hole filling shape description vector based on a first weight modifier matrix and a second weight modifier matrix to obtain a first through-hole filling shape description meta-modal auxiliary feature mapping pattern representation vector and a second through-hole filling shape description meta-modal auxiliary feature mapping pattern representation vector; based on the gradient amplitude of each position in the first through-hole filling shape description meta-modal auxiliary feature mapping pattern representation vector, The first through-hole filling shape description meta-modal auxiliary feature mapping pattern representation vector is subjected to gradient amplitude-based masking processing to obtain a masked first through-hole filling shape description meta-modal auxiliary feature mapping pattern representation vector as the masked first through-hole filling shape description meta-modal auxiliary feature mapping pattern representation vector; based on the gradient amplitude of each position in the second through-hole filling shape description meta-modal auxiliary feature mapping pattern representation vector, the second through-hole filling shape description meta-modal auxiliary feature mapping pattern representation vector is subjected to gradient amplitude-based masking processing to obtain a masked second through-hole filling shape description meta-modal auxiliary feature mapping pattern representation vector as the masked second through-hole filling shape description meta-modal auxiliary feature.

[0058] Optionally, in one embodiment of the present application, the through-hole filling shape description vector is linearly transformed based on the first weight modifier matrix and the second weight modifier matrix to obtain a first through-hole filling shape description meta-modal auxiliary feature mapping pattern representation vector and a second through-hole filling shape description meta-modal auxiliary feature mapping pattern representation vector, including: multiplying the through-hole filling shape description vector with the first weight modifier matrix, adding the obtained first through-hole filling shape description modification vector and the first modifier bias vector by position to obtain the first through-hole filling shape description meta-modal auxiliary feature mapping pattern representation vector; multiplying the through-hole filling shape description vector with the second weight modifier matrix, adding the obtained second through-hole filling shape description modification vector and the second modifier bias vector by position to obtain the second through-hole filling shape description meta-modal auxiliary feature mapping pattern representation vector.

[0059] Optionally, in one embodiment of the present application, based on the gradient amplitude of each position in the first through-hole filling shape description meta-modal auxiliary feature mapping pattern representation vector, the first through-hole filling shape description meta-modal auxiliary feature mapping pattern representation vector is subjected to a mask processing based on the gradient amplitude to obtain a masked first through-hole filling shape description meta-modal auxiliary feature mapping pattern representation vector, including: calculating the first through-hole filling shape description meta-modal auxiliary feature mapping pattern representation vector; Position and The square of the difference between the eigenvalues ​​of the positions is calculated to obtain the first through-hole filling shape description gradient amplitude representation vector; the absolute value of the square root of each eigenvalue in the first through-hole filling shape description gradient amplitude representation vector is calculated to obtain the first through-hole filling shape description gradient amplitude modulation representation vector; the first through-hole filling shape description gradient amplitude modulation representation vector is masked, and in response to the eigenvalue of each position in the first through-hole filling shape description gradient amplitude modulation representation vector being greater than a predetermined threshold, the eigenvalue of the position is set to 1, and the rest are set to 0, so as to obtain the masked first through-hole filling shape description meta-modal auxiliary feature mapping pattern representation vector.

[0060] Optionally, in one embodiment of the present application, based on the masked first through-hole filling shape description meta-modal auxiliary feature and the masked second through-hole filling shape description meta-modal auxiliary feature, feature selection is performed on the through-hole filling structure feature map to obtain the filling shape constrained through-hole filling structure feature, including: performing preliminary feature selection on the through-hole filling structure feature map based on the masked first through-hole filling shape description meta-modal auxiliary feature mapping pattern representation vector and hyperbolic tangent function to obtain a through-hole filling structure feature map after preliminary selection; based on the masked second through-hole filling shape description meta-modal auxiliary feature mapping pattern representation vector and Sigmoid function, performing feature selection again on the through-hole filling structure feature map after preliminary selection to obtain a filling shape constrained through-hole filling structure feature map as the filling shape constrained through-hole filling structure feature.

[0061] Optionally, in one embodiment of the present application, preliminary feature selection is performed on the through-hole filling structure feature map based on the masked first through-hole filling shape description meta-modal auxiliary feature mapping pattern representation vector and the hyperbolic tangent function to obtain the through-hole filling structure feature map after preliminary selection, including: multiplying the eigenvalues ​​of each position in the masked first through-hole filling shape description meta-modal auxiliary feature mapping pattern representation vector by each feature matrix of the through-hole filling structure feature map along the channel dimension by position to obtain a through-hole filling structure weighted feature map; and using the hyperbolic tangent function to perform feature selection on the through-hole filling structure weighted feature map to obtain the through-hole filling structure feature map after preliminary selection.

[0062] Optionally, in one embodiment of the present application, feature selection is performed again on the initially selected via filling structure feature map based on the masked second via filling shape description meta-modal auxiliary feature mapping pattern representation vector and the Sigmoid function to obtain a filling shape constrained via filling structure feature map, including: positionally multiplying the eigenvalues ​​of each position in the masked second via filling shape description meta-modal auxiliary feature mapping pattern representation vector with each feature matrix of the initially selected via filling structure feature map along the channel dimension to obtain a initially selected via filling structure weighted feature map; and using the Sigmoid function to process the initially selected via filling structure weighted feature map to obtain the filling shape constrained via filling structure feature map.

[0063] In summary, in the embodiment of the present application, the through-hole filling shape description vector is used as the meta-modal feature vector, and the through-hole filling shape description vector and the through-hole filling structure feature map are subjected to feature selection based on the metadata block, which can be expressed by the formula: in, is the shape description vector for the via fill, and are the first weight modifier matrix and the second weight modifier matrix respectively, and are the first modifier bias vector and the second modifier bias vector respectively, and The first through hole filling shape description meta-modal auxiliary feature mapping mode representation vector and the second through hole filling shape description meta-modal auxiliary feature mapping mode representation vector, and The first through hole filling shape description element modal auxiliary feature mapping mode representation vector and The eigenvalues ​​at each position, and The second through hole filling shape description element modal auxiliary feature mapping mode representation vector and The eigenvalues ​​at the positions, For masking operation, is the predetermined threshold, is the eigenvalue of each position in the masked first through-hole filling shape description element modal auxiliary feature mapping mode representation vector, is the eigenvalue of each position in the masked second through-hole filling shape description element modal auxiliary feature mapping mode representation vector, is the masked first through-hole filling shape description element modal auxiliary feature mapping mode representation vector, is the characteristic diagram of the through hole filling structure, In order to use the feature vector to weight the feature map, is the hyperbolic tangent function, is the characteristic diagram of the through hole filling structure after the initial selection, is the masked second through-hole filling shape description element modal auxiliary feature mapping mode representation vector, is the activation function, is the filling shape constrained through-hole filling structure feature map.

[0064] In detail, first, the through-hole filling shape description vector is subjected to a linear transformation based on the first weight modifier matrix and the second weight modifier matrix to obtain the first through-hole filling shape description meta-modal auxiliary feature mapping mode representation vector and the second through-hole filling shape description meta-modal auxiliary feature mapping mode representation vector. That is, using two different weight modifier matrices to implement a linear transformation on the meta-modal feature vector (through-hole filling shape description vector) can increase the representation range of the feature, allowing the model to more comprehensively analyze the relationship between the through-hole filling shape description vector and the through-hole filling structure feature map (main modal feature map) from multiple perspectives. Next, based on the gradient amplitude of the first through-hole filling shape description meta-modal auxiliary feature mapping mode representation vector and the second through-hole filling shape description meta-modal auxiliary feature mapping mode representation vector, the two vectors are masked to obtain the masked first through-hole filling shape description meta-modal auxiliary feature mapping mode representation vector and the masked second through-hole filling shape description meta-modal auxiliary feature mapping mode representation vector. It should be understood that the gradient amplitude characterizes the variation amplitude of each element in the feature mapping pattern representation vector. The use of masking operation can emphasize those features with significant variations, so that in the subsequent feature selection stage, more focus can be placed on those key features. Subsequently, the masked first through-hole filling shape is used to describe the meta-modal auxiliary feature mapping pattern representation vector and is combined with the hyperbolic tangent function to perform preliminary feature selection on the through-hole filling structure feature map. In particular, the hyperbolic tangent function is a commonly used activation function, and its output value range is between -1 and +1, which enables it to clearly distinguish different areas within the feature map and preferentially identify features that are significantly related to the meta-modal features during the preliminary selection process. Afterwards, in order to further highlight the key and important features in the feature map, based on the preliminary selection, the masked second through-hole filling shape is used to describe the meta-modal auxiliary feature mapping pattern representation vector and the Sigmoid function to select the features after the preliminary selection again. In particular, the Sigmoid function is also a common activation function, and its output value range is from 0 to 1, which makes it particularly suitable for probability prediction and binary classification decision scenarios. Therefore, the Sigmoid function can help identify the truly critical features and construct the final filling shape constrained through-hole filling structure feature map based on this. In this way, the through-hole filling shape description vector can be used as a constraint condition to extract and capture the most representative and discriminative features in the through-hole filling structure feature map.

[0065] In the above-mentioned method for preparing a high transmission rate optical module circuit board, the S24, based on the filling shape constraint through-hole filling structural features, obtains a test result. Optionally, in one embodiment of the present application, based on the filling shape constraint through-hole filling structural features, obtains a test result, including: inputting the filling shape constraint through-hole filling structural feature map into a filling quality detector based on a classifier to obtain a test result, and the test result is used to indicate whether there is a defect. That is, using the through-hole filling shape description vector as a meta-modal feature vector, the through-hole filling shape description vector and the through-hole filling structural feature map are subjected to feature selection to obtain the filling shape constraint through-hole filling structural features for classification processing, so as to automatically determine whether there is a defect. In this way, the through-hole filling state image is automatically collected by an industrial camera, which can significantly reduce the detection time, and can achieve a faster detection speed than manual detection or X-ray and ultrasonic detection performed on each sample. At the same time, it is possible to identify subtle defects, such as tiny bubbles, cracks, delamination, etc., improve the accuracy and reliability of detection, and realize automated thermal conductive through-hole filling quality detection.

[0066] Preferably, in an example of the present application, inputting the filling shape constraint via filling structure feature map into a classifier-based filling quality detector to obtain a detection result comprises:

[0067] Calculating the sum of the absolute values ​​of each eigenvalue of the filling shape constrained through-hole filling structure feature map to obtain a first filling shape constrained through-hole filling structure spatial structure value, and calculating the square root of the sum of the squares of each eigenvalue of the filling shape constrained through-hole filling structure feature map to obtain a second filling shape constrained through-hole filling structure spatial structure value;

[0068] multiplying each eigenvalue of the fill shape constrained via filling structure feature map by the first fill shape constrained via filling structure spatial structure value and the second fill shape constrained via filling structure spatial structure value respectively to obtain a first fill shape constrained via filling structure structure reference value and a second fill shape constrained via filling structure structure reference value corresponding to each eigenvalue;

[0069] Multiplying each eigenvalue of the filling shape constrained via filling structure feature graph by the scale of the filling shape constrained via filling structure feature graph and the square root of the scale to obtain a first filling shape constrained via filling structure scale transformation value and a second filling shape constrained via filling structure scale transformation value corresponding to each eigenvalue;

[0070] Dividing the first filling shape constrained via filling structure structure reference value by the difference between the first filling shape constrained via filling structure spatial structure value and the first filling shape constrained via filling structure scale transformation value to obtain a first filling shape constrained via filling structure transformation adjustment value;

[0071] Dividing the second filling shape constrained via filling structure structure reference value by the difference between the second filling shape constrained via filling structure spatial structure value and the second filling shape constrained via filling structure scale transformation value to obtain a second filling shape constrained via filling structure transformation adjustment value;

[0072] calculating a weighted sum of the first filling shape constrained via filling structure transformation adjustment value and the second filling shape constrained via filling structure transformation adjustment value to obtain each eigenvalue of the optimized filling shape constrained via filling structure feature graph; and

[0073] The optimized filling shape constraint via filling structure feature map is input into a filling quality detector based on a classifier to obtain a detection result.

[0074] Here, the filling shape constrains the via filling structure feature map, denoted as The optimization expression is:

[0075]

[0076]

[0077]

[0078]

[0079]

[0080]

[0081]

[0082]

[0083]

[0084] in, A through hole filling structure feature map constrained by the filling shape, represents the set of real numbers, , , Represent the width, length and height of the feature map respectively, The first characterization diagram representing the fill shape constraint via filling structure feature map The eigenvalues ​​at the positions, represents the scale of the feature map, represents the first filling shape constraint through hole filling structure structure reference value, represents the second filling shape constraint through-hole filling structure structure reference value, It means point multiplication by position. It means adding by position. represents the second filling shape constraint through hole filling structure transformation adjustment feature map, represents the second filling shape constraint through-hole filling structure transformation adjustment value, represents the first filling shape constraint through hole filling structure transformation adjustment feature map, represents the first filling shape constraint through-hole filling structure transformation adjustment value, represents the weighted hyperparameter, Feature graph showing the optimized fill shape constraint via filling structure.

[0085] That is, the present application takes into account that the through-hole filling shape description vector and the through-hole filling structure feature map respectively represent the image semantic features of different expression dimensions of the through-hole filling state image in the local-global spatial dimension. When performing feature selection based on metadata blocks, the differences in the semantic representation of meta-images in different spatial dimensions will also have differences in the spatial distribution structures of the features to be fused. Therefore, it is expected to improve the convergence and generalization effects of the complex spatial structure classification regression based on different spatial distribution structures of the filling shape constrained through-hole filling structure feature map.

[0086] Therefore, in the preferred example, for the spatial structure information of the feature set of the filling shape constraint through-hole filling structure feature map in the high-dimensional space, the scale-based frame transformation of each feature value of the filling shape constraint through-hole filling structure feature map is performed by using the class norm space structured representation of the filling shape constraint through-hole filling structure feature map as a reference window, and the spatial structure-based frame attention weight adjustment of each feature value of the filling shape constraint through-hole filling structure feature map is realized to ensure the invariance of the spatial transformation (translation, scaling and rotation) of the filling shape constraint through-hole filling structure feature map under the interaction of feature space, thereby improving the convergence and generalization effect of the classification regression of the feature set of the filling shape constraint through-hole filling structure feature map under the complex spatial structure representation, and improving the accuracy of the detection result obtained by inputting it into the filling quality detector based on the classifier. In this way, the automatic acquisition of the through-hole filling state image by the industrial camera can significantly reduce the detection time, and can achieve faster detection speed than manual detection or X-ray and ultrasonic detection of each sample. At the same time, it is able to identify subtle defects such as tiny bubbles, cracks, delamination, etc., which improves the accuracy and reliability of detection and realizes automated thermal via filling quality detection.

[0087] In summary, the preparation method of the high transmission rate optical module circuit board according to the embodiment of the present application is explained, which obtains the through-hole filling state image collected by the industrial camera, and adopts the image analysis and processing technology based on deep learning to respectively extract the filling shape and structural features of the through-hole filling state image, so as to automatically judge whether there is a defect by selecting the feature representation of the through-hole state structural feature based on the through-hole filling state shape feature as the constraint feature. In this way, the through-hole filling state image is automatically collected by the industrial camera, which can significantly reduce the detection time, and can achieve a faster detection speed than manual detection or X-ray and ultrasonic detection of each sample. At the same time, it can identify subtle defects, such as tiny bubbles, cracks, delamination, etc., improve the accuracy and reliability of detection, and realize automated thermal conductive through-hole filling quality detection.

[0088] Specifically, a high transmission rate optical module circuit board is also provided, and the high transmission rate optical module circuit board is manufactured by the aforementioned method for manufacturing the high transmission rate optical module circuit board.

[0089] The basic principle of the present invention is described above in conjunction with specific embodiments. However, it should be pointed out that the advantages, strengths, effects, etc. mentioned in the present invention are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. must be possessed by each embodiment of the present invention. In addition, the specific details of the above embodiments are only for the purpose of illustration and facilitation of understanding, rather than limitation, and the above details do not limit the present invention to being implemented by adopting the above specific details.

[0090] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, please refer to the relevant description of other embodiments. In the several embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the module division is only a logical function division, and there may be other division methods in actual implementation. The modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment.

[0091] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0092] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference to a figure in a claim should not be considered as limiting the claim to which it relates.

[0093] In addition, it is obvious that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units stated in the device claim can also be implemented by one unit through software or hardware.

[0094] Finally, it should be noted that the above description has been given for the purpose of illustration and description. In addition, the above embodiments are only used to illustrate the technical solution of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for preparing a high transmission rate optical module circuit board, characterized in that: include: Constructing a central sub-stack layer; opening a thermal through hole on the central sub-stack layer, and filling the thermal through hole with a thermally conductive metal to obtain a filled central sub-stack layer; stacking a prepreg and a copper foil on both sides of the filled central sub-stack layer and then laminating to obtain a pressed sub-stack layer; opening a laser hole at a predetermined position of the pressed sub-stack layer, and copper plating the laser hole to obtain the high transmission rate optical module circuit board; Wherein, filling the thermally conductive through hole with a thermally conductive metal to obtain a filled central sub-stacked layer further comprises: acquiring a through hole filling state image acquired by an industrial camera; Performing filling shape feature extraction and filling structure feature extraction on the through-hole filling state image respectively to obtain a through-hole filling shape feature map and a through-hole filling structure feature map, and extracting a through-hole filling shape description vector from the through-hole filling shape feature map; The through-hole filling shape description vector and the through-hole filling structure feature map are subjected to feature selection based on metadata blocks to obtain a filling shape constrained through-hole filling structure feature, including: performing linear transformation and gradient amplitude-based masking processing on the through-hole filling shape description vector to obtain a masked first through-hole filling shape description meta-modal auxiliary feature and a masked second through-hole filling shape description meta-modal auxiliary feature; based on the masked first through-hole filling shape description meta-modal auxiliary feature and the masked second through-hole filling shape description meta-modal auxiliary feature, feature selection is performed on the through-hole filling structure feature map to obtain the filling shape constrained through-hole filling structure feature; Based on the filling shape constraint through-hole filling structure feature, obtaining a detection result, including: inputting the filling shape constraint through-hole filling structure feature map into a filling quality detector based on a classifier to obtain a detection result, wherein the detection result is used to indicate whether there is a defect; The method further comprises: performing feature selection on the through-hole filling structure feature map based on the masked first through-hole filling shape description meta-modal auxiliary feature and the masked second through-hole filling shape description meta-modal auxiliary feature to obtain the filling shape constrained through-hole filling structure feature, including: Performing preliminary feature selection on the through-hole filling structure feature map based on the masked first through-hole filling shape description element modal auxiliary feature mapping pattern representation vector and the hyperbolic tangent function to obtain a through-hole filling structure feature map after preliminary selection; Based on the masked second through-hole filling shape description meta-modal auxiliary feature mapping pattern representation vector and Sigmoid function, feature selection is performed again on the through-hole filling structure feature map after the initial selection to obtain a filling shape constrained through-hole filling structure feature map as the filling shape constrained through-hole filling structure feature.

2. The method for preparing a high transmission rate optical module circuit board according to claim 1, characterized in that: Performing filling shape feature extraction and filling structure feature extraction on the through-hole filling state image to obtain a through-hole filling shape feature map and a through-hole filling structure feature map, and extracting a through-hole filling shape description vector from the through-hole filling shape feature map, including: Inputting the through-hole filling state image into a filling shape feature extractor based on a first hole convolutional neural network model to obtain the through-hole filling shape feature map; Inputting the through-hole filling state image into a filling structure feature extractor based on a second hole convolutional neural network model to obtain the through-hole filling structure feature map; A global mean pooling process is performed on the through-hole filling shape feature map to obtain the through-hole filling shape description vector.

3. The method for preparing a high transmission rate optical module circuit board according to claim 2, characterized in that: The through hole filling shape description vector is subjected to linear transformation and gradient amplitude-based masking processing to obtain a masked first through hole filling shape description meta-modal auxiliary feature and a masked second through hole filling shape description meta-modal auxiliary feature, including: Performing a linear transformation on the through hole filling shape description vector based on a first weight modifier matrix and a second weight modifier matrix to obtain a first through hole filling shape description element modal auxiliary feature mapping mode representation vector and a second through hole filling shape description element modal auxiliary feature mapping mode representation vector; Based on the gradient amplitude of each position in the first through hole filling shape description element modal auxiliary feature mapping pattern representation vector, performing gradient amplitude-based mask processing on the first through hole filling shape description element modal auxiliary feature mapping pattern representation vector to obtain a masked first through hole filling shape description element modal auxiliary feature mapping pattern representation vector as the masked first through hole filling shape description element modal auxiliary feature; Based on the gradient amplitude of each position in the second through-hole filling shape description meta-modal auxiliary feature mapping pattern representation vector, the second through-hole filling shape description meta-modal auxiliary feature mapping pattern representation vector is subjected to gradient amplitude-based mask processing to obtain a masked second through-hole filling shape description meta-modal auxiliary feature mapping pattern representation vector as the masked second through-hole filling shape description meta-modal auxiliary feature.

4. The method for preparing a high transmission rate optical module circuit board according to claim 3, characterized in that: The through hole filling shape description vector is subjected to a linear transformation based on a first weight modifier matrix and a second weight modifier matrix to obtain a first through hole filling shape description element modal auxiliary feature mapping mode representation vector and a second through hole filling shape description element modal auxiliary feature mapping mode representation vector, comprising: Multiplying the through hole filling shape description vector with the first weight modifier matrix, adding the obtained first through hole filling shape description modification vector with the first modifier bias vector by position to obtain the first through hole filling shape description element modal auxiliary feature mapping mode representation vector; The through hole filling shape description vector is multiplied by the second weight modifier matrix, and the obtained second through hole filling shape description modification vector is positionally added to the second modifier bias vector to obtain the second through hole filling shape description meta-modal auxiliary feature mapping pattern representation vector.

5. The method for preparing a high transmission rate optical module circuit board according to claim 4, characterized in that: Based on the gradient amplitude of each position in the first through hole filling shape description element modal auxiliary feature mapping pattern representation vector, performing a mask processing based on the gradient amplitude on the first through hole filling shape description element modal auxiliary feature mapping pattern representation vector to obtain a masked first through hole filling shape description element modal auxiliary feature mapping pattern representation vector, including: Calculate the first through-hole filling shape description element modal auxiliary feature mapping mode representation vector Position and The square of the difference between the eigenvalues ​​at the positions is used to obtain a gradient magnitude representation vector describing the first through-hole filling shape; Calculating the absolute value of the square root of each eigenvalue in the first through hole filling shape description gradient amplitude representation vector to obtain a first through hole filling shape description gradient amplitude modulation representation vector; The first through-hole filling shape description gradient amplitude modulation representation vector is masked, and in response to the eigenvalue of each position in the first through-hole filling shape description gradient amplitude modulation representation vector being greater than a predetermined threshold, the eigenvalue of the position is set to 1, and the rest are set to 0, to obtain the masked first through-hole filling shape description meta-modal auxiliary feature mapping pattern representation vector.

6. The method for preparing a high transmission rate optical module circuit board according to claim 5, characterized in that: Performing preliminary feature selection on the through-hole filling structure feature map based on the masked first through-hole filling shape description element modal auxiliary feature mapping mode representation vector and the hyperbolic tangent function to obtain a through-hole filling structure feature map after preliminary selection, including: Multiplying the feature values ​​of each position in the masked first through-hole filling shape description element modal auxiliary feature mapping pattern representation vector by each feature matrix along the channel dimension of the through-hole filling structure feature map by position to obtain a through-hole filling structure weighted feature map; The hyperbolic tangent function is used to perform feature selection on the weighted feature map of the via filling structure to obtain the initially selected via filling structure feature map.

7. The method for preparing a high transmission rate optical module circuit board according to claim 6, characterized in that: The method further performs feature selection on the initially selected through-hole filling structure feature map based on the masked second through-hole filling shape description element modal auxiliary feature mapping pattern representation vector and the Sigmoid function to obtain a filling shape constrained through-hole filling structure feature map, including: Multiplying the feature values ​​of each position in the masked second through-hole filling shape description element modal auxiliary feature mapping pattern representation vector by each feature matrix along the channel dimension of the through-hole filling structure feature map after preliminary selection by position to obtain a weighted feature map of the through-hole filling structure after preliminary selection; The weighted feature map of the through-hole filling structure after the preliminary selection is processed using a Sigmoid function to obtain the filling shape constrained through-hole filling structure feature map.

8. A high transmission rate optical module circuit board, characterized in that: The high transmission rate optical module circuit board is manufactured by the method for manufacturing a high transmission rate optical module circuit board as described in any one of claims 1-7.

Images