AI Applications for Failure Analysis - Adaptive Offset Cancellation in Bit-Line Sense Amplifiers: A Machine Learning Approach
AI Applications for Failure Analysis - Adaptive Offset Cancellation in Bit-Line Sense Amplifiers: A Machine Learning Approach
Wednesday, November 19, 2025
Summary:
When reading the cell's data in DRAM, it is amplified through a Bit-Line Sense Amplifier. Due to discrepancies in the threshold voltages of the transistors in the amplifier, the data signal could be sensed incorrectly. To adjust this, an Offset Cancellation process is proceeded, which compensates for the strength of the transmitted signal to ensure normal sensing occurs. However, if the Offset Cancellation timing is not optimized, it may not achieve the intended sensing results. In this paper, we introduce an adaptive Machine Learning-based strategy to address this issue. Our experimental results verify that our approach is effective in optimizing the timing of Offset Cancellation, leading to improved sensing performance in Bit-Line Sense Amplifiers.
When reading the cell's data in DRAM, it is amplified through a Bit-Line Sense Amplifier. Due to discrepancies in the threshold voltages of the transistors in the amplifier, the data signal could be sensed incorrectly. To adjust this, an Offset Cancellation process is proceeded, which compensates for the strength of the transmitted signal to ensure normal sensing occurs. However, if the Offset Cancellation timing is not optimized, it may not achieve the intended sensing results. In this paper, we introduce an adaptive Machine Learning-based strategy to address this issue. Our experimental results verify that our approach is effective in optimizing the timing of Offset Cancellation, leading to improved sensing performance in Bit-Line Sense Amplifiers.