Search

KR-102962396-B1 - ARTIFICIAL INTELLIGENCE BASED LASER AUTO RECIPE SUPPORTING MICRO BALL REPAIR DEVICE AND METHOD THEREOF

KR102962396B1KR 102962396 B1KR102962396 B1KR 102962396B1KR-102962396-B1

Abstract

A microball repair device supporting an artificial intelligence-based laser auto recipe according to an embodiment of the present invention comprises: a working stage on which a PCB is mounted and a microball repair process is performed; a solderball placement unit for picking up solderballs and attaching them to a location where the repair process is to be performed; a laser optical module for forming microballs by irradiating beams onto the attached solderballs using a plurality of coaxial planar light source lasers having different wavelengths and different irradiation areas; and an artificial intelligence-based laser recipe control unit for extracting feature values regarding pads and conductive pattern regions from an ODB file and determining a laser recipe using an artificial intelligence model trained with the feature values as input.

Inventors

  • 박준성
  • 김동식
  • 이상원
  • 서승현

Assignees

  • 레이저쎌 주식회사

Dates

Publication Date
20260508
Application Date
20240530

Claims (12)

  1. In a microball repair device supporting an artificial intelligence-based laser auto recipe, A working stage where a PCB is mounted and a micro-ball repair process is performed; A solder ball placement unit that picks up a solder ball and attaches it to a location where the repair process is to be performed; A laser optical module that forms a microball by irradiating a beam onto the attached solder ball using a plurality of coaxial planar light source lasers having different wavelengths and different irradiation areas; and An AI-based laser recipe control unit that extracts feature values regarding pads and challenge pattern regions from an ODB file and determines a laser recipe using an AI model trained with the feature values as input. Includes, The feature value of the above ODB file is the isolation status between the pad and the conductive pattern area based on the pad of the PCB board and the ratio of the conductive pattern connected to the pad, and The feature values of the above ODB file are, A first feature value for identifying whether the pad is isolated from or connected to the challenge pattern; A second characteristic value, which is the ratio of conductive copper connected to the pad within a specific area based on the pad; and A third characteristic value, which is the ratio excluding the ratio of the conductive copper connected to the pad within a specific area based on the pad. A microball repair device including
  2. In paragraph 1, The above artificial intelligence-based laser recipe control unit is, An ODB file analysis unit that analyzes the above ODB file to derive microball repair data; A repair area extraction unit that extracts a repair area of a microball using the microball repair data derived from the above ODB file analysis unit; A feature value extraction unit that selects and extracts feature values of the ODB file based on the microball repair data derived from the ODB file analysis unit; An artificial intelligence model unit that receives feature values extracted from the above feature value extraction unit; and A laser recipe determining unit that determines an optimal laser recipe based on the output of the above artificial intelligence model unit. A microball repair device including
  3. delete
  4. delete
  5. In paragraph 1, A micro ball repair device in which each of the above first feature value, second feature value, and third feature value includes a layer value of a multilayer PCB.
  6. In paragraph 2, The above artificial intelligence model unit is, A micro ball repair device learned based on the above-mentioned first feature value, second feature value, and third feature value and parameters of an artificial intelligence algorithm.
  7. In paragraph 6, The above artificial intelligence model unit outputs laser output and irradiation time, and A microball repair device verified as microball image data.
  8. In a microball repair method that supports an artificial intelligence-based laser auto recipe, Step of loading the PCB board; A step of detecting the repair location of a microball on the PCB substrate; A step of attaching a solder ball to the repair location above; A step of irradiating a beam onto the attached solder ball using a plurality of coaxial planar light source lasers having different wavelengths and different irradiation areas; A step of analyzing images captured by a coaxial vision camera and a side vision camera; A step of extracting feature values regarding pads and challenge pattern regions from an ODB file, and controlling a laser recipe using an artificial intelligence model trained with said feature values as input; and After laser irradiation according to the above controlled laser recipe is executed, the step of unloading the PCB substrate Includes, The feature values of the above ODB file are, A first feature value for identifying whether the pad is isolated from or connected to the challenge pattern; A second characteristic value, which is the ratio of conductive copper connected to the pad within a specific area based on the pad; and A third characteristic value, which is the ratio excluding the ratio of the conductive copper connected to the pad within a specific area based on the pad. Microball repair method including
  9. In paragraph 8, The step of controlling the laser recipe using the above artificial intelligence model is, A step of deriving microball repair data by analyzing the above ODB file; A step of extracting a repair area of a microball using the microball repair data derived above; A step of selecting and extracting feature values of the ODB file based on the microball repair data derived above; and Step of determining the optimal laser recipe based on the data learned by the above artificial intelligence model Microball repair method including
  10. In Paragraph 9, The above artificial intelligence model is, A step in which feature values of the ODB file are input into the artificial intelligence model; Step of setting parameters of the above artificial intelligence model; A step of artificial intelligence learning the data based on the input feature values and the set parameters; A step of outputting a laser recipe classification based on the above-mentioned learned data; and A step to verify the appropriateness of laser output and irradiation time based on video analysis data. A microball repair method learned by a learning method including
  11. delete
  12. In paragraph 8, A microball repair method in which each of the above first feature value, second feature value, and third feature value includes a layer value of a multilayer PCB.

Description

Artificial Intelligence-Based Microball Repair Device Supporting Laser Auto Recipe and Method Thereof The present invention relates to a microball repair device and method that supports an artificial intelligence-based laser auto recipe, and more specifically, to a microball repair device and method that supports an artificial intelligence-based laser auto recipe that can prevent the manufacturing of defective microballs and form good quality microballs by analyzing an ODB file to extract feature values and determining a laser recipe by training an artificial intelligence with the same, thereby controlling the recipe (laser output and irradiation time) of the laser emitted onto the microball. The reflow process is one of the core processes of Surface Mount Technology (SMT), and it is a process that electrically bonds a semiconductor and a substrate by applying thermal energy. For example, a method of attaching a semiconductor device to a substrate by applying heat to a semiconductor device mounted on solder balls on the substrate can be used. To do this, a prerequisite must be satisfied that the solder balls are placed in a good condition on a substrate with a pre-designed pattern. However, due to defects in the substrate manufacturing process, missing or defective solder balls are inevitable, and a process to repair these missing or defective solder balls is required. Conventionally, to address this problem, the missing ball was attached to the substrate and bonded using spot laser irradiation. However, spot lasers according to conventional technology have a Gaussian energy distribution in which the energy intensity increases toward the center of the irradiation area. Consequently, energy higher than the solder ball melting temperature is concentrated, causing defects due to thermal shock, or it is difficult to supply sufficient thermal energy as one moves away from the irradiation area. As one of the methods to solve these problems, it is possible to perform reflow by irradiating an area laser corresponding to the repair area. Since an area laser can irradiate a homogenized laser over a certain area, it is possible to transfer thermal energy uniformly. Meanwhile, due to the recent development of high-specification memory, AI semiconductors, and high-performance AP chips, the density of semiconductors is reaching its limit; in conjunction with this, substrates are also being equipped with micro-sized solder balls featuring micro-level fine pitches. Furthermore, as the pattern density and complexity of substrates have increased, multilayer substrates with multiple layers are being utilized. To repair microballs on such fine-pitch and multilayer substrates, precise control of the irradiation position and laser recipe (laser output and irradiation time) is required. Therefore, in order to repair missing or defective microballs in fine and complex substrate structures, a microball repair device capable of ultra-precision control and laser recipe control is required. FIG. 1 is a block diagram illustrating the configuration of a microball repair device according to an embodiment of the present invention. FIG. 2 is a perspective view illustrating a microball repair device implemented according to an embodiment of the present invention. FIG. 3 is a drawing illustrating a cross-section of a laser optical module according to an embodiment of the present invention. FIG. 4 is a drawing illustrating the detailed configuration and specifications of an optical unit according to an embodiment of the present invention. FIG. 5 is a drawing illustrating a laser optical module additionally equipped with a side vision camera according to an embodiment of the present invention. FIG. 6 is a block diagram illustrating the detailed configuration of an artificial intelligence-based laser recipe control unit according to an embodiment of the present invention. FIG. 7 is a drawing illustrating an example of a feature value according to an embodiment of the present invention. FIG. 8 is a flowchart illustrating an artificial intelligence learning method of an artificial intelligence model unit according to an embodiment of the present invention. FIG. 9 is a diagram illustrating the formation of microballs by artificial intelligence-based laser recipe control according to an embodiment of the present invention. FIG. 10 is a flowchart illustrating the operation method of a microball repair device according to an embodiment of the present invention. FIG. 11 is a flowchart illustrating a detailed method for controlling an artificial intelligence-based laser recipe of a microball repair device according to an embodiment of the present invention. The present invention will be described below with reference to the attached drawings. However, the present invention may be implemented in various different forms and is therefore not limited to the embodiments described herein. Furthermore, in order to clearly explain the present invention in the