Improving Portable Continuous Wave Doppler Processing Capabilities with SHARC 2147x Processors

Floating-point digital signal processing has become a constant requirement for precision technology, often in applications requiring high accuracy in areas such as aviation, industrial machinery, and healthcare. Medical ultrasound equipment is one of the most complex signal processing machines in use today, and is increasingly moving into the portable space. The challenge is to achieve this intensive signal processing without sacrificing system performance. With the introduction of the low-power SHARC 2147x processors, Analog Devices has been able to fully address the challenge of providing precision processing while reducing the power budget to enable applications such as portable ultrasound. This article discusses the use of portable ultrasound equipment, the processing technology used, and how the SHARC 2147x family of processors provides the necessary functionality at the lowest power levels.

Portability without sacrificing performance

Key care technologies like ultrasound systems require adequate reliability and consistent quality for both in-clinic and remote use. While advances in low-power technology have enabled the development of portable devices, there are still many essential components in medical ultrasound system design that are required to bring full hospital capabilities to disaster areas where ultrasound was previously inaccessible. It is incumbent on medical device developers to provide products that provide the highest image integrity performance within a variety of environmental, commercial, and technical constraints. For portable ultrasound devices in particular, system performance means being able to interpret images with the same clarity and accuracy as some specialized systems—only now it carries the burden of meeting specific category constraints such as weight, size, battery life, and cost. These design constraints require components with real-time computing capabilities, low power consumption, and low cost and compactness for product design considerations. With the rise of portable ultrasound devices, the challenge of simultaneously meeting low power consumption and maintaining the same performance level becomes increasingly daunting.

Continuous wave Doppler imaging

Ultrasound imaging is based on Johann Christian's Doppler principle, which states that moving objects emit detectable frequencies - "Doppler shifts" or sounds. For example, an ultrasound image of blood density is created by directing a beam into a blood vessel and then detecting (the "sound" of) the blood flow. There are two main modes of Doppler ultrasound imaging, pulsed wave (PW) Doppler and continuous wave (CW) Doppler. Pulsed wave Doppler transmits ultrasound pulses along a scan line and uses the relative time between received signals to calculate the Doppler frequency - thus using the pulse characteristics of the transmitter to obtain information about the location of blood flow.

Continuous wave Doppler ultrasound

This article focuses on the second type, continuous wave Doppler ultrasound, which can detect and measure the velocity of moving tissues in the body. Because it produces continuous waves, continuous wave Doppler has higher sensitivity and lower bandwidth requirements, usually less than 100kHz, so it is particularly effective for estimating higher blood flow velocities. High-velocity detection with continuous wave Doppler can be used to diagnose congenital or valvular heart disease, because high blood flow velocity configuration tracking is the basis for detecting these diseases.

As the name implies, when using continuous wave Doppler ultrasound technology, the transmitting transducer (piezoelectric crystal) transmits a continuous single frequency tone while the receiving transducer records the acoustic echo ultrasound signal. Because the interpretation of the beat frequency (Doppler shift) determines the speed and direction of blood flow in the cardiovascular system, high performance signal processing in the continuous wave path is a key element of measurement accuracy. The dynamic range of the continuous wave Doppler signal is the largest of all signals in an ultrasound system, partly due to leakage from the transmitted signal traversing the receive path (caused by the half-duplex nature of signal transmission) and reflections from fixed body parts close to the body surface. Detecting blood flow in deeper vessels in the body will produce very weak Doppler signals, so a wide dynamic range is required for the entire continuous wave signal chain. High quality ultrasound system performance is directly related to whether good signal chain integration is achieved.

Dynamic range of floating point processing

The exponentiation inherent in floating-point arithmetic ensures that a much larger dynamic range is available—the largest and smallest values ​​that can occur—which is particularly important when processing extremely large data sets or data sets whose ranges may be unpredictable. As a result, floating-point processors are ideal for computationally intensive applications such as Doppler ultrasound. This dynamic range processing enables portable ultrasound systems that use continuous-wave Doppler technology to detect these very low signals. The functions of the digital signal processing unit in the continuous-wave path are to implement at least wall filtering, envelope detection, and fast Fourier transforms (FFTs).

Analog Devices’ Full Signal Chain Integration

Analog Devices' SHARC 2147x family of DSPs and analog front-end (AFE) components can process ultrasound signals throughout the entire signal chain. As with any complex technology, highly integrated components improve overall system efficiency and performance. For signal processing intensive applications like portable ultrasound, the speed and efficiency of the entire signal chain will directly affect the maintenance of quality, albeit in a portable form factor. To achieve precision analysis, it is critical to maintain strong signal integrity from reception, to front-end analog signal processing components, to digital signal processing and back-end.

The new SHARC 2147x series processors from Analog Devices are mainly used for computationally intensive floating-point applications. The SHARC 2147x series processors integrate a 32-bit floating-point unit with 40-bit extended precision capability, support wide dynamic range and very high precision calculations, and are designed to operate at high frequencies with very low power consumption. These processors use low-power process technology to reduce overall power consumption, and also use other power reduction technologies, so that power consumption is very low when idle. This combination of low active power consumption and very low idle power consumption can extend battery life. Low power consumption also means that no forced cooling technology is required, allowing the processor to be used in places where space is quite tight. The SHARC 2147x series processors have a very small form factor, so they can achieve high space efficiency - all of these features are extremely suitable for portable ultrasound applications.

The SHARC 2147x processor family has 5Mb of on-chip memory, which minimizes the need for external memory and improves overall system performance. In some compact embedded application implementations, on-chip memory alone is sufficient, eliminating the need for external memory, thereby reducing bill of materials (BOM) costs. With 5Mb of on-chip memory, ultrasound system development can achieve the lowest BOM cost and maximum portability. To maximize chip performance, the SHARC 2147x processor family also integrates a dedicated hardware accelerator with independent operation units and DMA memory mapping capabilities, which can be used to implement parallel FFT processing functions for various velocity components in decoding and analyzing the returned Doppler signal. In addition, offloading the FFT operation to this parallel engine can also reduce the power consumption of the FFT processing cycle.

Analog front-end (AFE) components can be used to optimize analog signal chain performance while limiting the number of board components and minimizing power consumption. The AD9276 octal receiver from Analog Devices not only includes processing capabilities for B-mode and pulsed wave Doppler mode imaging, but also includes an integrated I/Q demodulator, so continuous wave Doppler processing can be achieved in a very small form factor and at ultra-low power. A time gain control (TGC) channel with 8-channel variable gain amplifier (VGA) with low noise preamplifier (LNA), anti-aliasing filter (AAF), and 12-bit 10MSPS to 80MSPS analog-to-digital converter (ADC) can provide a high-quality imaging system for high-end ultrasound systems. The built-in I/Q demodulator with programmable phase delay per channel allows the system to process continuous wave Doppler signals with a particularly large dynamic range.

Ultrasound equipment implementation using Analog Devices SHARC 2147x as the core floating-point processor and AD9276 as the analog front end.

Analog Devices' SHARC DSPs, such as the SHARC 2147x family, and analog front-end products can help medical product designers transform complex, highly reliable, computationally intensive technologies such as continuous wave Doppler ultrasound into portable designs. The SHARC 2147x family of DSPs achieves system developers' design goals by keeping costs low, reducing system complexity through integration, and shortening development cycles, all without sacrificing the key design goal: ultrasound functionality and reliability in field situations that can deliver diagnostic capabilities no less than systems used in hospitals.

Analog Devices SHARC Floating-Point Digital Signal Processor

Analog Devices' 32-bit floating-point SHARC digital signal processors feature a Super Harvard architecture that balances excellent core and memory performance with outstanding I/O throughput capabilities. This Super Harvard architecture extends the traditional concept of separate program and data memory buses by adding an I/O processor with an associated dedicated bus. In addition to meeting the requirements of the most computationally intensive real-time signal processing applications, the SHARC processors integrate large memory arrays and dedicated peripherals to simplify product development and reduce time to market.

Complete development and support ecosystem

Analog Devices' hardware and software development tools are designed to provide engineers with easier and more robust system development and optimization methods, simplifying the product development process and shortening time to market. The SHARC processor family uses well-known development tools, including the VisualDSP++ integrated development and debugging environment (IDDE) and the EZ-Kit Lite* evaluation and application prototyping platform. A rich third-party software support network can further help developers design smarter and more efficient solutions for a wide range of applications.

Analog Devices Healthcare Solutions

ADI provides healthcare customers with a comprehensive portfolio of linear, mixed-signal, MEMS and digital signal processing technologies for medical imaging, patient monitoring, medical instruments, and consumer/home healthcare. Backed by leading design tools, application support and system expertise, ADI's products and technologies play a pivotal role in medical design, helping to influence and determine the future of diagnostic and monitoring equipment, as well as health and wellness devices.