Medical Ultrasound Imaging is widely used clinically because of its relatively low cost, portability, lack of
ionizing radiation, and real-time nature. However, even with these advantages ultrasound has failed to
permeate the broad array of clinical applications where its use could be of value. A prime example of this untapped potential is the routine use of ultrasound to guide intravenous access. In this particular application existing systems lack the required portability, low cost, and ease-of-use required for widespread acceptance.
Our team has been working for a number of years to develop an extremely low-cost, pocket-sized, and
intuitive ultrasound imaging system that we refer to as the "Sonic Window." We have previously described
the first generation Sonic Window prototype that was a bench-top device using a 1024 element, fully
populated array operating at a center frequency of 3.3 MHz. Through a high degree of custom front-end
integration combined with multiplexing down to a 2 channel PC based digitizer this system acquired a full
set of RF data over a course of 512 transmit events. While initial results were encouraging, this system
exhibited limitations resulting from low SNR, relatively coarse array sampling, and relatively slow data acquisition.
We have recently begun assembling a second-generation Sonic Window system. This system uses a 3600 element fully sampled array operating at 5.0 MHz with a 300 micron element pitch. This system extends the
integration of the first generation system to include front-end protection, pre-amplification, a programmable
bandpass filter, four sample and holds, and four A/D converters for all 3600 channels in a set of custom
integrated circuits with a combined area smaller than the 1.8 x 1.8 cm footprint of the transducer array. We
present initial results from this front-end and present benchmark results from a software beamformer
implemented on the Analog Devices BF-561 DSP. We discuss our immediate plans for further integration
and testing. This second prototype represents a major reduction in size and forms the foundation of a fully
functional, fully integrated, pocket sized prototype.
Preliminary results relating to the design, fabrication, and characterization of a 3600 (60 x 60) element, fully sampled, 5 MHz two dimensional (2D) array are presented. The viable element yield of the new array was estimated at 98.3%. Single-element pulse-echo experiments indicate that the center frequency is 4.7 MHz - 7.8% below the resonant frequency determined by Finite Element Analysis (FEA) simulation. Pulse-echo signal fractional bandwidth was measured to be 60.3% at the -6 dB level. Ringdown was longer than anticipated in experimental pulse echo voltage waveforms, which we attribute in part to a lack of matching layer and a low-loss backing material. Based on plane-wave pulse-echo experiments in a water-tank, single element signal-to-noise ratio (SNR) was calculated to be 6.0 when using a plane-wave transmit (all elements excited). Experimental angular beam patterns were more directional than predicted with the standard soft-baffle equation, but in good agreement with FEA simulations that take account of finite acoustic crosstalk.