Dan Tang, President and CEO
Dan Tang, founder, president, and CEO of Global Electronic Testing Services (Global ETS), has been involved in the electronic testing industry for the past 25 years. He has a thorough understanding of counterfeiting issues impacting the semiconductor industry. He has assisted original equipment manufacturers (OEMs) and independent distributors in mitigating counterfeit issues by providing comprehensive testing solutions. In 2012, he founded Global ETS, capitalizing on his decades of testing experience.
Global ETS is an internationally recognized independent component testing and authentication laboratory. The company employs various testing methods to ensure the authenticity and compliance of electronic products, per manufacturer’s standards.
The company’s testing methods include Global ETS patented AI Automated Visual System, Remarking / Resurfacing, Decapsulation, XRF Testing, Automated X-ray Inspection using Global ETS patented software, Automated Extended Electrical Testing with full AC & DC parametric at specific temperatures, Functional Testing, and Up Screening. Other services include Failure Analysis with extensive Scanning Electron Microscopy (SEM), energy dispersive X-ray spectroscopy (EDS), Hermeticity Testing, Material Analysis, automated Hi-Rel Re-tinning, and Re-balling.
Automated of AI Visual Inspection
With GETS system, the texture features of the IC package are extracted and analyzed. The image’s parameterized texture features are used to find evidence of counterfeit IC
Global ETS realizes that AI vision is becoming one of the essential parts of the automated testing process. The problem has been the visual inspection process has relies on the experts manually inspecting the product.
As there is no industry standard for individuals while inspecting, inspectors need to find evidence of counterfeit, such as ghost marking, sanding marks, or the signs of secondary coating. A wrong decision by an inspector can cause error in "authentication process." Also, the continuous high-focused work environment can lead to fatigue, impacting an inspector's ability to find subtle evidence of the counterfeit.
“Global ETS’s automatic AI vision system is one of the solutions to these problems. With the system, the texture features of the IC package are extracted and analyzed. The image’s parameterized texture features are used to find evidence of counterfeit IC,” says Tang.
The system inspects the images taken from different angles giving a 360 degrees view of the IC. By analyzing these images and comparing them to each other, it provides an inspection result whether the IC is authentic or a counterfeit. The system reduces the inspection time from hours to minutes and provides a comprehensive inspection instead of random sampling.
It can reduce biased judgments between inspectors, making accurate and consistent decisions.
AI vision is also used in the quality control of the tinning process. After the automated tinning process is finished by Global ETS’ robot system, the AI-based lead inspection system will examine the quality of every lead on the IC to ensure there is no bent lead, no blistering on the lead, and no bridging between leads and lead coplanarity. The system also detects and estimates the coverage area of the soldering.
AI vision also is enable Global ETS to build a fully automatic testing line to reduce human involvement in non-destructive testing (external visual inspection, X-Ray, and electrical testing).
Global ETS will soon be releasing its patented Universal FPGA testing system that will allow testing FPGA products up to 3000 BGA. The system identifies the package pin configuration from the chip datasheet and automatically configures the ship interface for each individual pin of the chip, such as different power supply timing and access of different data types and external clock types. Chips using the system interface are correctly configured for automatic chip testing without needing to build additional external circuits. At the same time, the external electrical properties of multiple chips can be remotely configured in real time for a multi-state target test system. The applicable chip types include but are not limited to: functional chips based on digital communication, FPGA, MCU, memory chips, CPU chips, and GPU chips.
Electronic Component Authenticity via AI
Even before the pandemic, customers purchasing non-standard chips on the grey market has been a recurrent concern for the semiconductor industry. Product counterfeiting is common when purchases are made through non authorize distributors as no manufacturer certification is available. This can be mitigated by testing through a certified testing facility such as Global ETS.
Furthermore, the recent struggling semiconductor chip supply chain has forced many customers to use outdated but functional boards. Global ETS assists them by salvaging obsolete chips from the board and reworking them. These chips are then electrically tested 100 percent to determine the life expectancy, and after the post-processing, they are repackaged for use at the OEM.
“GETS system employs AI software and technology, allowing us to increase the speed and accuracy of testing while maintaining full traceability,” says Tang.
The system uses electrical signal measurement and artificial intelligence with deep learning to identify the authenticity of electronic components. It employs a quick, signature-based method for part authentication by scanning pins and comparing the signatures with GETS database. The user-friendly system can be operated by individuals without an electrical engineering degree. Additionally, it eliminates the need for complex data extraction, saving time. The system also includes a selfimproving deep learning model that learns from previous test cases, improving accuracy over time. Importantly, the method is non-destructive and can be completed instantly, resulting in efficient testing without damaging the chip.
The input data for the model can be divided into knowngood data and the device under test. The data features for the known-good data and the device under test are the same and include the total number of pins, input impedance, and output impedance.
Before inputting them into the model, the data is noise-reduced and then converted into a uniform dimension to be the extracted feature of the model.
As the model is trained, it adjusts the impact weight of each feature using the attention mechanism. This mechanism is widely used in various applications, including computer vision and speech recognition.
Undertaking Research on Robotics and AI Technology
In collaboration with the University of South Florida, Global ETS has developed various testing methodologies to develop AIbased testing technology that has the potential to set the industry standard for component testing. Unlike competitors that take 8 to 10 weeks to generate test results for their components, the company’s AI component testing technology delivers results in less than a week. This collaboration contributes to enhancing not only technology but also human resources and team building. It has enabled Global ETS to recruit highly qualified Electrical Engineering personnel. Additionally, Global ETS has three filed patents for AI component testing approaches, helping it acquire a competitive advantage.
GETS system employs AI software and technology, allowing us to increase the speed and accuracy of testing while maintaining full traceability
Global ETS has successfully assisted numerous clients in testing their electronic components. One of the testimonials of its success concerns the application of its up-screening capability. The customer presented a product sourced through their distribution channel. They needed the semiconductor component of the product to perform at specified temperatures and in a distinct environment under stress. When no other firm could develop a test plan, Global ETS used its up-screening abilities to screen the complex internal interface of the chip in the component and ensure it worked as expected by the customer.
Due to its highest quality and technical competence in component testing, the company has been certified by several independent organizations, such as ISO 17025 and SAE AS6171 accreditation. It has also been certified as an AS9001D and ISO9001 company, implying global acknowledgment of its quality. Along with the certifications, Global ETS is now authorized to test defense equipment due to its inclusion on the Defense Logistics Agency’s coveted Qualified Testing Suppliers List. The aforementioned certifications enhance the company’s global reputation as a reliable testing institution.
Headquartered in Odessa, Florida, Global ETS has facilities in Penang, Malaysia; Shenzhen, China; and Ontario, California. With over 300 employees worldwide and reputable clients such as Raytheon, Honeywell, Lockheed Martin, and Google, its long-term goal is to improve testing accuracy while reducing manual labor through robotic automation. The management plans to open a new Global ETS 2.0 facility in Eindhoven, Netherlands, in the 1st quarter of 2023. The facility consists of a new in-house developed automated testing system from an automated AI visual system to an automatic X-ray system and AI electrical testing system. In parallel, Global ETS will establish an engineering facility in Taiwan in 2023 to support its global location and undertake research on robotics and AI technology.