This paper presents electromagnetic analysis (EMA) experiments on three different CMOS chips, each executing cryptographic algorithms (DES, COMP128, RSA) and successfully retrieving the complete key material. The study demonstrates that EM signals can be used to extract secret information from tamper-resistant devices, even when no software countermeasures are present. The experiments were conducted using custom-made probes and advanced measurement techniques, and the results were compared to power analysis (PA) results. The paper discusses the principles of EMA, including the use of inductive sensors and the relationship between EM fields and power consumption. It also describes the design of the probes, the electrical behavior of the chips, and the spatial positioning of the probes to capture data-dependent signals. The results show that EM signals, while noisier than power signals, can provide more accurate and distinct data signatures, particularly when compared to power analysis. The study includes three main experiments: one on an alleged COMP128 algorithm, one on DES, and one on RSA. In each case, the EM analysis successfully retrieved the secret key, demonstrating the effectiveness of EMA. The results show that EM analysis can be more efficient than PA in certain scenarios, particularly when the correct guess is identified with fewer acquisitions and higher signal-to-noise ratio. The paper also discusses the differences between EM and PA, noting that while both can be used to extract secret information, EM analysis has the advantage of exploiting local information, allowing for more precise identification of data-dependent areas. However, PA is simpler to implement and is often used in conjunction with EM analysis when necessary. The study concludes that EM analysis is a viable method for extracting secret information from cryptographic devices, and that further research is needed to develop more effective countermeasures against such attacks. The paper also highlights the importance of hardware and software countermeasures in protecting against side-channel attacks.This paper presents electromagnetic analysis (EMA) experiments on three different CMOS chips, each executing cryptographic algorithms (DES, COMP128, RSA) and successfully retrieving the complete key material. The study demonstrates that EM signals can be used to extract secret information from tamper-resistant devices, even when no software countermeasures are present. The experiments were conducted using custom-made probes and advanced measurement techniques, and the results were compared to power analysis (PA) results. The paper discusses the principles of EMA, including the use of inductive sensors and the relationship between EM fields and power consumption. It also describes the design of the probes, the electrical behavior of the chips, and the spatial positioning of the probes to capture data-dependent signals. The results show that EM signals, while noisier than power signals, can provide more accurate and distinct data signatures, particularly when compared to power analysis. The study includes three main experiments: one on an alleged COMP128 algorithm, one on DES, and one on RSA. In each case, the EM analysis successfully retrieved the secret key, demonstrating the effectiveness of EMA. The results show that EM analysis can be more efficient than PA in certain scenarios, particularly when the correct guess is identified with fewer acquisitions and higher signal-to-noise ratio. The paper also discusses the differences between EM and PA, noting that while both can be used to extract secret information, EM analysis has the advantage of exploiting local information, allowing for more precise identification of data-dependent areas. However, PA is simpler to implement and is often used in conjunction with EM analysis when necessary. The study concludes that EM analysis is a viable method for extracting secret information from cryptographic devices, and that further research is needed to develop more effective countermeasures against such attacks. The paper also highlights the importance of hardware and software countermeasures in protecting against side-channel attacks.