Integrated 1024 Channel Ultrasound Beamformer For Ultrasound Research

Volumetric ultrasound imaging is of most importance not only for real time cardiac imaging and quantization of flow but also for niche applications like motion tracking and robust 3D microbubble tracing. In ultrasound research, the use of fully connected matrix arrays is the most flexible solution to develop scan strategies or micro beamforming techniques that will later be transferred into more cost-effective hardware solutions. We developed a fully integrated 1024 channel ultrasound system providing freely programmable transmit parameters and raw received data access for each channel and element. It integrates four ultrasound research scanners of our latest "DiPhAS" generation with 256 channels each and an internal synchronization logic for parallel operation into a single 19-inch rack with a height of 800 mm. Compared to other systems in the literature providing a high channel count by linking and synchronizing multiple electronics systems we present a solution that controls all 1024 channels integrated with only one PC without the need to transfer data from synchronized computers. Each ultrasound electronics connects with PCIexpress and a 25 Gbit/s link providing 100 Gbit/s overall bandwidth to transfer raw single element channel data of all the 1024 transducer elements. Data from all devices is directly transferred to DMA memory in the PC for further processing using the dual Intel Xeon processors and four additional Nvidia graphics adapters for GPU processing implementing real-time software-based volumetric beamforming. Received signal bandwidth ranges from 300 kHz up to 20 MHz with a sampling rate of 80 MSPS and a resolution of 12 bit. Real-time volumetric imaging using commercial and custom matrix arrays is implemented using volumetric plane wave compounding. We will investigate the usage of custom matrix array transducers with large footprints to realize deeper penetration compared to the specific cardiac imaging transducers in the future improving applications for real-time motion and microbubble tracking in tissue.