In these days of severe semiconductor chip shortage, the electronic industry is more prone to buying fake/counterfeit semiconductor ICs, and use them in some life-saving equipments and other critical infrastructure electronics. Using of fake devices in automotives; in today's driving automated systems poses huge dangers. So there is absolute need to check originality of semiconductor ICs for safety critical electronics systems. Testing each and every component for electrical performance is nearly impossible due to the time to market pressure and also the costs involved in that.
The best method to protect from buying fake parts is to buy from authentic approved sources. Modern semiconductor chips are integrating some unique device ID to prevent from faking them. More and more semicoductor vendors should add a unique signature to the chip.
Using artificial intelligence (AI), machine vision and X-Ray scanning of chips, the X Ray scanned copper interconnections and other patterns of original chip can be compared with counterfeit. This method is non-intrusive and doesn't involve thorough test-jig like electrical testing of the chip.
A new start-up called Chiplytics has come out with a product called "Chiplytics One" which tests semiconductor chips for there authenticity using available datasets for original chip.
Chiplytics says Counterfeit semiconductors pose an estimated $200 billion risk per year to the global electronics supply chain. Chiplytics combines optical and electrical testing techniques to capture the holistic state of electrical components and detect anomalies using machine learning.
The "Chiplytics One" is claimed as the first semiconductor inspection platform that combines machine vision and electrical testing to enable rapid, data-driven quality assurance. Testing is performed non-invasively within seconds without needing to power on the chip. Simply upload the part's datasheet to the web app to prepare the system for testing.