电容式微机械超声换能器
超声波传感器
专用集成电路
波束赋形
计算机科学
传感器
电子工程
计算机硬件
电气工程
工程类
声学
物理
作者
Kailiang Chen,Hae-Seung Lee,C.G. Sodini
出处
期刊:IEEE Journal of Solid-state Circuits
[Institute of Electrical and Electronics Engineers]
日期:2015-12-30
卷期号:51 (3): 738-751
被引量:71
标识
DOI:10.1109/jssc.2015.2505714
摘要
This paper presents a scalable column-row-parallel ASIC architecture for 3-D portable medical ultrasound. Through its programmable row-by-row or column-by-column operations for both transmit and receive beam-formation, linear scaling in interconnection, data acquisition complexity, power dissipation, and programming time is achieved. In addition, its per-element controllers can activate fine granularity aperture definition when more functionality is favored over the linear-scaling power and speed efficiency. This front-end architecture is backward compatible to implement existing widely used array aperture patterns, while supporting new imaging apertures and algorithms. It lends itself very well for the combination with integrated or external digital beamforming circuits. A 16 × 16 proof-of-concept ASIC is fabricated and flip-chip bonded to a 16 × 16 capacitive micromachined ultrasonic transducer (CMUT). Each three-level pulsing transmitter (Tx) is 46% more power efficient than a traditional two-level version, with high-voltage (HV) multiplexers (MUXs) designed for flexible Tx parallelization. Each low-noise receiver (Rx) consumes 1.4 mW active power and 54 μW sleep power, with optimized source follower stages to combine analog outputs for improved SNR. The transceivers are also fault-tolerant to inevitable defects in transducers, greatly enhancing assembly yield. The system demonstrates 3-D plane-wave generation to implement the coherent compounding algorithm for fast volume rate (62.5 volume/s), high-quality 3-D ultrasonic imaging. An interleaved checker board pattern with Iand Q excitations is also demonstrated for ultrasonic harmonic imaging, which reduces transmitted second harmonic distortion (HD2) by over 20 dB.
科研通智能强力驱动
Strongly Powered by AbleSci AI