(click to enlarge)Pin-compatible ICs, such as the AD9272/73, enable designs to be leveraged across multiple ultrasound systems.

Hospitals, medical clinics, and mobile emergency units are increasingly relying on large and small ultra-sound equipment for a broad spectrum of patient care. In addition, ultrasound applications that were virtually unimaginable in the past—such as the use of portable devices in remote locations—are becoming realities.

These new and varied demands are fueling modern ultrasound development. Equipment designers are asked to deliver high image quality and reliable performance without compromising power efficiency—in devices that are sometimes scarcely larger than a human hand. To do so, designers are harnessing improvements in underlying technologies, particularly in semiconductor component integration. IC ultrasound innovations are easing designers' tasks and also enabling significant advances in the performance, size, and power of ultrasound equipment.

The Impact of Integrated Receivers

The receive chain in an ultrasound system consists of four basic elements: a low-noise amplifier (LNA), a variable-gain amplifier (VGA), an antialiasing filter (AAF), and an analog-to-digital convertor (ADC) (see Figure 1). These components, which comprise a signal processing channel, are replicated many times in ultrasound's common digital beamforming (DBF) architecture.

Portable ultrasound systems generally use from 16 to 64 channels. The greater the number of channels, the greater the dynamic range (as long as the channel noise is uncorrelated). When the number of channels is doubled, the noise is halved and the dynamic range increases by 3 dB. So, for example, a 64-channel system offers a dynamic range that is 6 dB greater than that of a 16-channel system. This increased channel density results in higher image quality, which enables doctors to diagnose medical conditions more accurately.

(click to enlarge)Figure 1. A simplified block diagram for a typical ultrasound system.

To meet the need for high channel counts and other portability demands, IC vendors are integrating the key elements of the receive chain on a single chip. These integrated octal (eight-channel) receivers enable a reduced form factor, which results in smaller printed circuit boards (PCBs). Therefore, ultrasound equipment designers can add more channels in less space, increasing channel density to ensure high-grade images on even the most compact systems.

Equally important, a small form factor allows ultrasound equipment to be brought directly to patients, whether in a hospital bed, doctor's office, or field hospital. This miniaturization is significantly expanding ultrasound's potential, creating a mobile point-of-care device that is driving visionary applications. For example, it may soon be commonplace to find portable ultrasound units on sports fields and in locker rooms. Or they may be used by EMTs in ambulances to scan patients on the way to the hospital to save critical time in life-threatening situations. Furthermore, as ultrasound scanning becomes easier and more accessible, it will likely be used for a wider variety of clinical and diagnostic applications.

Less Power + Lighter Weight = Greater Versatility

Another advantage of octal ultrasound receivers is energy efficiency. Integrated components consume much less power than discrete components, without eroding performance. For example, the AD9273 (Analog Devices Inc.; ADI) octal ultrasound receiver with crosspoint switch requires about 50% less power than a discrete component. Because power demands are reduced, ultrasound devices can be used for longer periods of time without recharging. This is a considerable advantage in applications where electricity may be unreliable, scarce, or even nonexistent.

Decreased power demands also enable the use of smaller, lighter batteries, leading to the development of new, lightweight ultrasound models. This reduction in weight eases the physical strain on the user, which is critical for devices that are carried to distant patients or held for extended periods of time. (Most people who have carted around a laptop will agree that a one- or two-pound weight reduction has an enormous effect on user comfort.)

Design Advantages beyond Size and Power

As higher-level integration schemes become more predominant and components shrink in size, the advantages of integrated components will only become stronger. However, there is more to ultrasound integration advances than size and power alone. Performance is the ultimate driver of today's developments, and application-specific standard product manufacturers are moving beyond size to optimize the functions that are critical to ultrasound's effectiveness. Performance challenges are being met by components that allow the performance versus power level to be set digitally within the integrated circuit. This enables ultrasound equipment designers to produce a system that is configurable and scalable depending on the imaging application.

Adapting to Diverse Probe Frequencies

Ultrasound transducer probes come in a variety of shapes and sizes, with multiple probe types regularly serving a single system. As ultrasound systems become smaller, the need for probe variety is likely to expand—i.e., if a handheld system can be used virtually anywhere, it may be called upon to perform more and varied imaging tasks with different probe types.

Today's diverse probes use a broad range of operating frequencies, typically between 1 and 15 MHz. Continual improvements in probe quality are leading to higher signal outputs from the probes of greater than 500 mV p-p. Therefore, receive channels are now being engineered to adjust to a wider input range, providing much needed latitude for today's ultrasound designers. Octal ultrasound receivers simplify the design task by making it easier to match the signal level and frequency settings required for this myriad of expanding probe technologies and frequencies.

One engineering approach used with receiver IC products with a limited input range of 250 mV p-p is to add an attenuator at the input of the device to accommodate higher signal outputs from today's more-sophisticated probes. The inclusion of these extra components adds additional noise, reducing the system's signal-to-noise ratio.

(click to enlarge)Figure 2. A single-channel block diagram of the AD9272/73, both pin-compatible ICs.

Another approach provides an adjustable input range. Some companies offer receivers with three different values: 773, 550, and 367 mV p-p (see Figure 2). This enables a system designer to accommodate low-, mid-, and high-range signals, supporting multiple, diverse probe technologies without supplemental components.

With the ADI receive channel, a proprietary ultra-low-noise LNA resides at the input of the signal chain, minimizing the noise contribution in the VGA that follows. The LNA is programmable to optimize input dynamic range. It supports differential output voltages as high as 4.4 V p-p with positive and negative excursions of +1.1 V from a common-mode voltage of 1.5 V. The LNA differential gain sets the maximum input signal before saturation. One of three gains is set through the serial port interface (SPI). The corresponding full-scale input for the gain settings of 15.6, 17.9, and 21.3 dB is, as noted above, 733, 550, and 367 mV p-p, respectively. Overload protection ensures quick recovery time from large input voltages. Because the inputs are capacitively coupled to a bias voltage near midsupply, very large inputs can be handled without interacting with the electrostatic discharge protection.

Impedance Matching with Minimal Noise

Today's extensive array of transducers also presents a broad range of termination impedance levels, typically between 50 and 300 Ω, with some probes having no termination at all. Most ultrasound receiver manufacturers prefer to match the impedance of the probe with the impedance of the processing unit's front end to minimize reflections and improve the system's transient response.

In the past, ultrasound equipment designers addressed this issue with multiple discrete components that consumed significant power and board space. This began to change in 2007 with the introduction of an integrated analog front-end receiver. For the first time, a single integrated chip actively matched to two different termination values, via an SPI port control, to accommodate multiple ultrasound probes. This approach simplified design and freed up PCB space.

(click to enlarge)Figure 3. Common input termination schemes.

Today, octal receiver vendors offer two distinct termination methods–resistive and active (see Figure 3). Resistive termination involves placing an external shunt resistor across the input of the LNA to ground. Active termination uses an external resistor in the feedback path of the front-end LNA to actively match the source to the load.

In both approaches, the input-referred noise of the receiver must be minimized. This can prove troublesome with resistive termination, as the input noise figure increases by 3 dB (or more) at normal probe impedances, when compared with active termination. With resistive termination, the input noise increases due to the thermal noise of the matching resistor and the increased contribution of the LNA's input voltage noise generator. With active impedance matching, however, the contributions of both are smaller by a factor of 1/(1 + LNA gain) than they would be for resistive termination.

(click to enlarge)Figure 4. Noise figure versus Rs for various terminations.

Figure 4 provides an example of relative noise performance. In this graph, the input impedance was swept with Rs to preserve the match at each point. The noise figures for a source impedance of 50 Ω are 4.1, 3.3, and 2.6 dB for the resistive termination, active termination, and unterminated configurations, respectively. The noise figures for 200 Ω are 2, 1.5, and 1.2 dB, respectively.

Figure 5 (on p. 11) shows the noise figure as it relates to Rs for various values of Rin, which is helpful for design purposes. The penalty paid for shunt termination is made clear in these examples and can be a serious issue in ultrasound system performance.

(click to enlarge)Figure 5. Noise figure versus Rs for various fixed resistor values of Rin, with active termination.

These termination methods offer designers the ability to provide user-adjustable input impedance termination. Consequently, doctors, sonographers, and clinicians can employ multiple probes with diverse termination requirements on a single, compact unit and, in the case of active termination, do so with lower noise.

Filtering Out Signal Distortions at Higher Frequencies

Depending on the imaging application, an ultrasound unit may operate at a different range of frequencies. Therefore, filters must be added to the receiver signal chain to reduce signal distortions (aliasing). As mentioned earlier, system designers had to build their own passive filters using multiple discrete components or had to create active filters using an op amp with passive components and switches. Today, AAFs are integrated directly into chips, eliminating the need for multiple components and reducing costs and space requirements.

(click to enlarge)Figure 6. A simplified schematic of a programmable antialias filter.

AAFs are available in two formats—fixed-range and programmable. With fixed-range filters, system designers select the filter that best represents their system's frequency range, for example a fixed range of 8–12 MHz. Alternatively, a programmable AAF lets a designer adapt the filter to meet specific frequency needs, adjusting the filter over a wide range. With the AD9273 receiver, an AAF between the TGC path and the ADC is programmable through an SPI, with an adjustable range of 8–18 MHz (see Figure 6 on p. 12).

Using the AD receiver's eight SPI-programmable settings, for example, a system designer can vary the high-pass filter cutoff frequency as a function of the low-pass cutoff frequency. As the ratio decreases, the amount of rejection on the low-end frequencies increases. By making the entire AAF frequency pass band narrow, the designer can reduce low-frequency noise or maximize dynamic range for harmonic processing. So in the end, the designer has a greater number of choices and the end-user can employ more probes—with better noise immunity than a fixed-range system.

Ensuring High Image Quality

Ultimately, these advances in signal filtering and noise reduction lead to higher-quality ultrasound images. The lower the noise floor, the greater the dynamic range, which means the better the ultrasound image. And higher-quality images allow more-accurate diagnoses.

In fact, with innovations such as the ability to scale LNA power, designers can balance power with performance to improve dynamic range as necessary. This could enable ultrasound manufacturers to design systems that allow doctors to scan with a lower-quality image to conserve battery life or to turn up the power to improve image quality when a diagnosis requires it.

Dynamic range is also critical for signal penetration depth. For example, the signal from a stomach scan is considerably weaker than that of a scan just below skin level. Or a portion of bone might interrupt a signal, sending back a strong reflection that requires a gain decrease to avoid image distortion.

For example, with the AD9273, signals penetrating through the body are attenuated by about 1 dB/cm/MHz. If an 8-MHz probe with a depth penetration of 4 cm is used, the signal amplitude variation from reflections near the surface is 64 dB (or 4 × 2 × 8). If 50 dB of imaging resolution is required, and loss from bones, cables, and other mismatches is taken into account, the desired dynamic range would approach 114 dB.

Accelerating Time to Market

Innovations in semiconductor component integration are providing ultrasound designers with new flexibility. Depending on imaging modality or probe type, designers can system-scale their designs in real-time, offering maximum performance at minimum power. In addition, as discussed earlier, integrated components offer countless advantages in size, power, and performance.

These innovations also provide time efficiencies. One example can be found in octal receivers that allow user controls to be programmed from the SPI, such as the scalable LNA. Designers can easily customize a broad range of systems, saving time by quickly changing SPI registers to ensure the lowest possible noise, the greatest dynamic range, and the longest battery life.

Another example is pin compatibility. Pin-compatible ICs allow designs to be leveraged across multiple ultrasound systems. For example, the AD receivers mentioned earlier are both pin-compatible, with the AD9272 emphasizing optimal performance and the AD9273 providing optimal power. Designers can use a common layout for both receivers, populating the PCB with the appropriate part based on the system requirements with minimal changes. This not only saves man-hours and resources, but can also have a positive effect on R&D budgets. In turn, reduction in development time means savings in time to market.

Bringing Ultrasound Wherever It's Needed

As ultrasound technologies continue to evolve, advances in integrated multichannel devices are pushing system flexibility further than ever. Tomorrow's portable ultrasound units are strongly positioned to deliver the same performance and image quality as today's high-end cart systems—only in smaller, lighter units with increased energy efficiency and patient convenience.

As a result, medical practitioners worldwide will soon be able to make accurate, noninvasive diagnoses more quickly than ever before, in places never thought possible. Although tomorrow's ultrasound devices will be smaller, their influence is certain to be greater.