 |
SEOUL, South Korea, April 30, 2025 /PRNewswire/ — LG unveiled the Dream Factory, a hub for the production of FC-BGAs (Flip Chip Ball Grid Arrays), the company’s next-generation growth engine, to the media for the first time and announced it on the 30th April.
A process of Line Quality Control (LQC), which checks whether the product meets the specifications (thickness, size, etc.) required by the customer. The inspection result data is immediately transmitted to the customer and cannot be manipulated. This quality transparency is one of the most important factors for LG Innotek’s global clients.
In 2022, LG Innotek announced its plans to launch a business producing FC-BGAs, high-value semiconductor substrates. To build the Dream Factory, the company acquired LG Electronics’ Gumi 4 Factory and began full-scale mass production in February 2024.
The Dream Factory, spanning a total area of 26,000 square meters, is regarded as the industry’s most advanced "smart" factory, integrating the latest IT technologies, including artificial intelligence, deep learning, robotics, and digital twin technologies. By applying automation, information, and intelligence technologies to the entire process, it has established a cutting-edge FC-BGA production infrastructure that eliminates the four major factors known to undermine production competitiveness: human error (Man), failure cost (F-cost), breakdown maintenance (BM) loss, and accidents.
- "Elimination of defects caused by human contact" through automation of all processes and logistics using robots
For semiconductor substrate products such as FC-BGAs, which require a highly demanding ultra-fine process, even the smallest foreign objects (eyelash, saliva, etc.) can cause quality issues. Therefore, it is crucial to minimize human contact with products during production.
To this end, LG Innotek has introduced a completely automated logistics system at the Dream Factory, where coming across a person is a rare event. Apart from essential personnel, such as equipment maintenance and repair workers, all 10 steps of the FC-BGA production and logistics processes are unmanned.
Dozens of autonomous mobile robots (AMRs) move materials around the production line autonomously. When a production order is placed, which is done automatically, based on the customer’s delivery timeframe entered in the RTS (real-time schedule), the AMRs transport the raw materials to the process facility. Once the barcode on the raw material is detected, also done automatically, by the machine, the process recipe is automatically set on the equipment according to the product specifications through the recipe management system (RMS), after which the product processing begins. The AMRs are also responsible for loading the finished product back into the stocker.
In addition, the process of peeling off the protective film from the panel (film detach) is also replaced by a robot. This enables the early prevention of fine scratches and defects caused by foreign objects such as dust particles and foreign substances. The construction of non-touch production facilities, involving such equipment as collaborative robots, in the entire process has significantly reduced mishandling by workers.
- Unmanned AI-based FC-BGA quality inspection, enhancing customer confidence by ensuring quality "transparency"
The Dream Factory generates more than 200,000 files and 100GB of data related to FC-BGA production every day. LG Innotek collects this data throughout the production process through sensors installed in all of its facilities. By applying AI, which continuously learns from this big data, to the defect prediction and inspection system, the company has significantly reduced the lead time caused by defects.
In addition, LG Innotek has applied an AI deep learning vision inspection system to the Automated Optical Inspection (AOI) process, which is the most important step in determining whether a product is of good quality. Tirelessly, the robot moves the finished FC-BGA board products to the vision screening inspection table. Next, the AI, which has been trained on tens of thousands of data points on defective and good quality FC-BGAs, detects micro-level defects that are challenging to identify with the naked eye, and it does so in only 30 seconds.
LG Innotek is operating a further advanced AOI process. In the room next to the AOI equipment for faulty circuits, there is a much larger robot and inspection system. Called Line Quality Control (LQC), it can automatically check whether the various specifications (thickness, size, etc.) of the product requested by the customer have been fulfilled. The inspection data is immediately sent to the customer, ensuring product quality transparency, which leads to higher customer confidence. This industry-leading sophistication of LG Innotek’s AOI equipment has been cited as one of the most impressive aspects of the factory by the global customers who have visited it.
AI can identify defective products, and since every product has a barcode that tracks its process history, products that are deemed defective are automatically filtered out without the need for human intervention, reducing F-costs by more than 50%.
Additionally, AI has been applied to the digital simulation system, which prevents product defects and equipment failures. Previously, the process of workers manually checking products for defects and identifying which machines were faulty and how to repair them in response to defects required a lot of time. This can now be significantly improved.
By 2026, LG Innotek plans to introduce an intelligent Quality Management System (i-QMS) that detects and analyzes quality irregularities during production in real time and automatically corrects them. The company plans to automate the entire FC-BGA production process, especially by developing a platform that uses digital twin technology to share information on all processes, from product development to production, with customers in real time to enhance customer responsiveness.
- Optimized FC-BGA process equipment using digital twin technology, "halving the ramp-up period"
Since even the smallest variables can lead to poor performance in FC-BGAs, equipment optimized for mass production and process recipes and production environments set to perfect values are fundamental to achieving high yields.
The FC-BGA process equipment installed in the Dream Factory is set to optimal conditions through digital twin technology. In the past, identifying optimal conditions for the FC-BGA process required a lot of time and money, along with hundreds of tests. Before building the facility, LG Innotek managed to identify problems with the initial setup of the FC-BGA process facility in advance by conducting a "factory simulation" using 3D modelling in virtual space. This enabled the facility to be carefully set for optimal conditions, such as liquid, heat, and air flow, which were difficult to measure inside the actual facility. As a result, the ramp-up period (increase in production capacity by improving initial production yields) was shortened by nearly half compared to the previous operation period.
In addition, digital twin technology is applied to the Line Monitoring System (LMS), which monitors production status in real time. The real-time monitoring system enables users to monitor the production line currently in operation, product movement, inventory status, equipment irregularities, production performance, and product quality status at a glance on the large screen of the integrated control room where the LMS is installed. This makes it possible to respond immediately in the event of any irregularity.
- Internalization of glass core technology, step-by-step entry into the high-end FC-BGA market: "Fostering the business into 700 million USD business by 2030"
Over the past 50 years, LG Innotek has accumulated core technologies for high-value semiconductor substrates, such as ultra-fine microcircuits and high-density, multi-layer substrate matching technology (stacking multiple substrate layers accurately and evenly), through its substrate material components business.
Based on this know-how, the company began full-scale mass production of FC-BGAs for personal computers (PCs) for North American big-tech customers at the end of last year and recently succeeded in securing additional global big-tech clients. This year, LG Innotek aims to enter the FC-BGA market for PC central processing units (CPUs). Its strategy is to enter the high-end FC-BGA market in phases, including entering the server FC-BGA market as early as 2026. In preparation for this, LG Innotek has already acquired facilities that are essential in the manufacturing of FC-BGA products for servers, such as "edge coating" that blocks the generation of dust particles.
In line with this goal, LG Innotek will also accelerate the development of next-generation substrate technologies in collaboration with global big-tech clients. By 2027, the company plans to internalize technologies such as re-distribution layer (RDL) technology, which engraves microcircuit patterns directly onto the substrate; device embedding technology, which minimizes power loss by embedding devices into the substrate; and multi-layer core (MLC) and glass core (glass substrate) technologies, which prevent warping when implementing large-area substrates. In particular, LG Innotek has been promoting glass substrates by strengthening its collaboration with global customers.
Minseok Kang, vice president and head of LG Innotek’s Substrate & Material Business Unit, said, "LG Innotek will continue to expand the production of FC-BGAs that provide exceptional customer value based on its state-of-the-art Dream Factory and develop the FC-BGA business into 700 million USD business by 2030."
According to the Fuji Chimera Research Institute, the size of the global FC-BGA market is expected to more than double from USD 8 billion in 2022 to USD 16.4 billion in 2030.
[Glossary]
FC-BGA (Flip Chip Ball Grid Array): This is a semiconductor substrate that is widely applied to electronic devices equipped with semiconductor chips (CPU, GPU, AI chips, etc.) that perform various computing functions. Demand for high-performance semiconductor substrates is growing rapidly due to increased data throughput, higher semiconductor processing speeds, and the need for low-power semiconductors. For these reasons, FC-BGAs have a larger area and more layers than conventional semiconductor substrates. They are also known to have high barriers to market entry because they require top-level facilities and technologies to produce.