Low-noise circuitry is critical to image quality of medical ultrasound; an indispensable diagnostic tool.
Despite the conceptual simplicity of the underlying principle, a complete, high-end ultrasound imaging system is a complicated device (Figure 1). The ultimate performance of the system is largely determined by the transducer and analog front end (AFE), while post-processing of the digitized reflected signal allows algorithms to enhance the situation.
There is a transmit/receive (T/R) switch between the piezoelectric transducer array and the active electronics. The role of this switch is to prevent the high-voltage transmit signals driving the transducer from reaching and damaging the low-voltage receive-side analog front end. After the reflected received is amplified, and conditioned, it is passed to the analog/digital converter (ADC) of the AFE, digitized, and undergoes software-based image processing and enhancement.
Each of the available imaging modes of an ultrasound system has different requirements for the dynamic range and thus SNR or noise requirements:
For black-and-white image mode – which provides the highest resolution – 70-dB dynamic range is required. The noise floor is important as it impacts the maximum depth at which the smallest ultrasound echo can be seen in the far field.
For pulse-wave-doppler (PWD) mode, which uses Doppler principles and false color images to show the flow of fluid or motion of organs, 130 dB is required.
For continuous-wave-doppler (CWD) mode which is a blend of the previous modes with some enhancements, 160 dB is needed.
Noise: a major factor in system performance
Not surprisingly, system noise of various types is one of the limiting factors in image quality and performance, again analogous to the consideration of bit error rate (BER) versus signal-to-noise ratio (SNR) in digital communication systems.
This noise has multiple electronic sources as well as sources in the target (the tissue). For the latter, the dominant noise is “speckle-noise” due to its on-screen appearance (Figure 2). This noise is due to the unavoidable fact that living tissue within an organ is not homogeneous and uniform but instead has small irregularities. These cause micro-reflections of the impinging ultrasound wave.
Speckle noise is not due to anything in the electronics and so cannot be improved by better design of the circuitry. Some ultrasound systems, however, use sophisticated algorithms to try to minimize it.
These requirements are not easy to meet. As the ultrasound transducer frequency is typically from 1 MHz to 15 MHz, it will be affected by any switching frequency noise within this range. If there are intermodulation frequencies within the PWD and CWD spectrums (from 100 Hz to 200 kHz), the obvious noise spectrums will appear in the Doppler images, which is unacceptable in the ultrasound system. For maximum system performance and image quality (clarity, dynamic range, lack of image speckling, and other figures of merit), it’s important to look at sources that cause loss of signal quality and degradation of SNR.
The first one is obvious: due to attenuation versus distance, the returns from tissues and organs deeper in the body (such as kidneys) are far weaker than those from those close to the transducer. Therefore, the reflected signal is “gained up” by the AFE to boost the signal-to-noise ratio (SNR), so the signal occupies as much of the AFE’s input range as possible.
To do this, an automatic gain control (AGC) function is used, similar to the one in wireless systems where the AGC assesses wireless RF received signal strength (RSS) and dynamically compensates for its random, unpredictable changes over a span of tens of decibels.
The AGC situation, however, is different in the ultrasound application than it is for a wireless link, where the received signal strength (RSS) varies continuously as a function of propagation and path-link conditions relative and absolute movement, the position of system and source, and many other factors; in many cases, RSS is random or random-like and is hard to predict or bound.
Instead, in an ultrasound system, the path attention is known with a fair degree of accuracy, as it is the velocity of acoustic-energy propagation (1540 m/sec in soft tissue — about five times faster than propagation in air at about 330 m/sec), and therefore the attenuation rate is also known.
Based on this a priori knowledge, the AFE uses a variable-gain amplifier (VGA) which is arranged as time-gain compensation (TGC) amplifier. The gain of this VGA is linear-in-dB and is configured so a linear-versus-time ramping control voltage increases the gain versus time to compensate to a large extent for the attenuation (Figure 3). This maximizes SNR and the use of the dynamic range at the AFE.
The final part of this article looks at the many critical noise issues associated with ultrasound imaging system performance.
EE World References
Signal generator board produces complex waveforms for ultrasound transducers, medical apps
High-voltage analog switch ICs handle medical ultrasound imaging applications
The Doppler effect: From highly ridiculed to absolutely indispensable, Part 1
External references
National Library of Medicine/IEEE, “Despeckling of Medical Ultrasound Images”
Elsevier/Science Direct, “Performance Enhancement and Analysis of Filters in Ultrasound Image Denoising”
Radiologic Clinics of North America, “The Essentials of Extracranial Carotid Ultrasonographic Imaging”
Maxim/Analog Devices, “Overview of Ultrasound Imaging Systems and the Electrical Components Required for Main Subfunctions”
Analog Devices, “Ultrasound Analog Electronics Primer”
ST Microelectronics, “STMicroelectronics introduces highly integrated 32-channel ultrasound transmitter optimized for handheld scanners”
ST Microelectronics, “Products and solutions for Medical Ultrasound”
