SonMin-Hyeong1
UmJi-Yong1
-
(Department of Medical IT Convergence Engineering, Kumoh National Institute of Technology)
Copyright © The Institute of Electronics and Information Engineers(IEIE)
Index Terms
Analog front-end amplifier, programmable gain amplifier, seamless time-gain compensation, operational amplifier, ultrasound scanner
I. INTRODUCTION
Ultrasound scanners have found widespread utility across diverse medical applications
due to their compact form factor and application-specific functionalities [1-5]. This paper introduces an analog front-end (AFE) amplifier for ultrasound application-specific-integrated-circuits
(ASICs). The proposed AFE amplifier consists of a preamplifier with a single passband
gain and a programmable gain amplifier (PGA) for a time-gain compensation (TGC).
Ultrasound scanning systems rely on TGC as a fundamental operation [6]. Ultrasound signals experience exponential attenuation in a medium, determined by
factors such as a propagation distance and an acoustic signal frequency, as shown
in Fig. 1(a). To ensure an accurate reconstruction of the instantaneous ultrasound scattering
characteristics, it becomes imperative to compensate for the propagation-dependent
attenuation [6].
There are relevant prior works to perform TGC operations in ultrasound ASICs [7-10]. While a continuous TGC operation closely approximates an ideal scenario [7,8], existing approaches have exhibited drawbacks such as intricate circuit topologies
or the need for analog buffers due to a high output impedance.
Considering a complexity of circuit topology and a driving ability of TGC amplifier,
a capacitive-feedback PGA can be an appropriate topology [9-11]. A voltage gain of PGA needs to be adjusted in a real time to accomplish a TGC operation
along a scanning depth as shown in Fig. 1(b). One of primary concerns of real-time TGC operation is to suppress switching artifacts
[2, 7, 10] during a gain transition. The primary objective of this work is to attain
a seamless gain transition during a real-time TGC operation by using an operating
point initialization method.
In this paper, we propose a preset operation for initializing the operating points
of AFE amplifier. This preset operation effectively aligns the operating points of
switching nodes, ensuring a seamless discrete TGC. In addition, to further enhance
a transient response of AFE amplifier, this work proposes a current-reuse operational
amplifier (op amp) employing dynamic push-pull current sources. The proposed op amps
are deployed in each stage of AFE amplifier.
This paper is organized as follows. Section II addresses the issue of gain transition
in the discrete TGC and presents a concept of proposed preset operation. Section III
presents the circuit implementations of proposed AFE amplifier. Section IV shows the
measurement results, and Section V concludes this work.
Fig. 1. Dynamic range of AFE amplifier output: (a) without TGC; (b) with TGC operation.
II. CONCEPT OF SEAMLESS DISCRETE TGC
Fig. 2 illustrates prior works of discrete TGC with PGA [9,10]. Both PGAs of Fig. 2 incorporate a reconfigurable series capacitor array to facilitate the adjustment
of voltage gain through the manipulation of switches within the array. However, it
is worth noting that these methods can introduce a common-mode fluctuation in the
PGA output due to voltage discrepancies at critical nodes within the series capacitors.
In case of PGA of Fig. 2(a) [10], a left node of series capacitor is basically connected to a common-mode node of
V$_{\mathrm{CM}}$. Meanwhile, the operating point of preamplifier output is typically
determined by its feedback configuration, resulting in a slight difference between
the preamplifier output and the voltage of V$_{\mathrm{CM}}$. During the gain switching
moment, the left node of series capacitor is switched from the V$_{\mathrm{CM}}$ to
the preamplifier output. The discrepancy in operating point between the preamplifier
output and the level of V$_{\mathrm{CM}}$ leads to an input offset of PGA. To mitigate
the fluctuation of output common-mode level of PGA, the work of [10] employs an analog buffer. The analog buffer temporarily shortens the relaxation time
of feedback configuration, facilitating the rapid recovery of the targeted operating
point. Measurement results in [10] demonstrate a transient time of less than 100 ns during gain transitions. Considering
a typical center frequency of medical ultrasound, a further reduction of transient
interval is preferable.
The PGA of Fig. 2(b) adjusts the voltage gain by switching the right node of series capacitor [9]. During the reset phase, the right nodes of series capacitors are connected to the
node of V$_{\mathrm{CM}}$, and the summing node of op amp is set to its operating
point, which is slightly different from the level of V$_{\mathrm{CM}}$. During the
gain switching moment, the right node of series capacitor is switched from the node
of V$_{\mathrm{CM}}$ to the summing node of op amp. The discrepancy of operating point
of right node of series capacitor results in an equivalent input referred offset in
PGA, generating a deviation of common-mode level in PGA output.
Fig. 3 shows the proposed preset operation to alleviate the voltage discrepancy of gain-switching
node. A single-ended circuit topology is illustrated for simplicity. Similar to the
work of [9], the voltage gain of PGA is adjusted by switching the right node of series capacitor.
To initialize the right node of series capacitor and the summing node of op amp with
the same operating point, the right node of series capacitor is also preset to the
operating point of op amp during the reset phase. The voltage deviation attributable
to charge injection of switches can be negligible as a result of the fully differential
topology and sufficient large capacitance of series capacitors compared to a relatively
small size of switch. After the reset phase, series capacitors are disconnected from
the summing node of op amp. During the gain switching moment, the right node of the
corresponding series capacitor is re-connected to the summing node of op amp. Since
the time duration between the preset operation and the gain switching moment is less
than 40 ${\mu}$s considering the maximum scanning depth, the charge loss due to the
leakage current on the floated right node of series capacitor is negligible. Furthermore,
this preset operation achieves the seamless gain transition with a relatively simple
reconfigurable series capacitor array.
Fig. 2. Discrete TGC in prior works: (a) variable feedback-resistance based PGA; (b) summing-node-switching based PGA.
Fig. 3. Proposed preset operation for seamless discrete TGC: (a) circuit diagram; (b) timing diagram.
III. CIRCUIT IMPLEMENTATION
1. Overall Circuit Diagram of AFE Amplifier
Fig. 4(a) shows an overall circuit diagram of AFE amplifier. The AFE amplifier consists of
a preamplifier and a PGA. The gain of both stages is determined by a capacitance ratio
of series capacitor and feedback capacitor. The preamplifier maintains a fixed gain
of 20 dB, while the PGA offers four selectable gain values from 16 dB to 33 dB. The
PGA supports the proposed preset operation with the dedicated timing diagram.
The proposed AFE has two operational phases: a reset phase and an RX phase, denoted
in Fig. 4(b). The reset phase corresponds to a typical ultrasound firing phase, so called as a
TX phase, initializing each of nodes in the amplifier for the following echo amplification
operation [6]. The proposed preset operation is concurrently executed during the reset phase, configuring
the operating points of both nodes of series capacitors in PGA as the preamplifier
output and the PGA output. After the reset phase, the gain-control switches of gain[2:0]
are disconnected to make the gain of PGA the minimum value. As a scanning depth increases,
each of gain-control switches is sequentially connected to the summing node of op
amp with the pre-determined time interval ${\Delta}$t to perform the periodic discrete
TGC operations.
Fig. 4. Proposed AFE amplifier: (a) circuit diagram; (b) timing diagram.
2. Op Amp with Dynamic Push-pull Current Sources
Each stage of AFE amplifier incorporates a fully differential op amp, as shown in
Fig. 4(a). In this work, the employed op amp is based on a current-reuse topology [11-13]. Fig. 5(a) shows a schematic of a conventional current-reuse op amp. This amplifier reuses a
constant bias current of I$_{\mathrm{Q}}$ for both NMOS and PMOS input differential
pairs, leading to an increment of transconductance, a reduction of noise power, and
an enhancement of power efficiency [11-13].
This work proposes a current-reuse op amp with dynamic push-pull current sources,
as shown in Fig. 5(b). To illustrate the transient operation for instant increment and decrement of inputs
of op amp, the increment or decrement of voltage and current are denoted as red color
in Fig. 5(b). The push-pull current source corresponds to the flipped voltage follower (FVF),
which is based on a feedback loop [14]. The drain-node voltage of input transistor in FVF operates oppositely to the changes
in input voltage of FVF, contributing to an instant increase or decrease in pull-up
or pull-down currents during the transient period. In case of Fig. 5(b), the input V$_{\mathrm{INP}}$ is increased and the input V$_{\mathrm{INN}}$ is decreased
instantly during the transient period. Then, the pull-up current of M1 and the pull-down
current of M4 instantly increase as a result of feedback loop in each FVF. Subsequent
to this transient interval, the dynamic current sources maintain a quiescent current
I$_{\mathrm{Q}}$. In a comparative analysis, the proposed op amp with dynamic current
sources shows a 40% reduction in settling time, when both op amps with the same quiescent
current are subjected to an abrupt input change in SPICE simulations. Hence, the proposed
op amp facilitates a seamless discrete TGC operation by further reducing a transient
interval.
Fig. 5. Current-reuse op amp with (a) constant tail current sources; (b) dynamic push-pull current sources.
IV. MEASUREMENT RESULTS
The proposed AFE amplifier has been implemented in a 180-nm BCD process with an active
area of 0.19 mm$^{2}$ including a bias circuit, as shown in Fig. 6. Operating at a 1.8-V supply voltage, a total power consumption of AFE is 2.3 mW.
Considering several primary specifications such as the center frequency of 4 MHz,
the maximum scanning depth of 4 cm, and the attenuation coefficient of 0.7 dB/(MHz${\cdot}$cm),
the required TGC amount for the maximum scanning depth corresponds to 22.4 dB. We
set a number of discrete TGC moments as three, as illustrated in Fig. 1. Correspondingly, we determined the specifications of PGA to cover the gain range
of 16.8 dB with four gain steps.
Fig. 7 shows the measured frequency response of AFE. It shows four discrete gain steps from
35.7 dB to 53 dB at a low frequency band. The bandwidth of AFE is 4.5 MHz at the maximum
gain setting, effectively supporting the bandwidth of the target transducer. At the
target center frequency of 4 MHz, the gain of PGA spans from 34.3 dB to 50.5 dB, featuring
an average gain step of 5.4~dB.
Fig. 8 demonstrates the real-time TGC operations in the measured time-domain waveforms.
We applied a sinusoid signal with a frequency of 4 MHz as the input for the implemented
AFE amplifier, as shown in Fig. 8(a). During each of gain switching moments, the output of AFE amplifier was further amplified,
showing a seamless gain transition with a transient interval less than 18 ns. In Fig. 8(a), the measured DC-level shift is less than 7 mV, which corresponds to the input-referred
offset of 21 ${\mu}$V. Considering the typical ultrasound signal processing such as
a band-pass filtering and an envelope detection, it is estimated that this DC-level
shift rarely causes the switching artifact in an ultrasound image. Due to the limitation
of measurement setup, we applied the emulated echo signals as the input (Fig. 8(b)), rather than real ultrasonic echoes. The output waveforms show the consistent peak-to-peak
amplitude, demonstrating the uniform gain steps in the real-time TGC operations.
Fig. 9 shows the simulated input-referred voltage noise density of the implemented AFE amplifier.
The input referred noise density at the center frequency of ultrasound is 6.3 nV/${\sqrt{}}$Hz.
As a result of utilization of the current-reuse op amp and an optimization of transconductance
of transistors in op amp [12], it shows a sufficient noise performance as the AFE amplifier.
Table 1 compares the proposed AFE amplifier with the prior works of [7-11]. The proposed AFE amplifier performs a discrete TGC operation for the voltage input
signal. The real-time discrete TGC operation of proposed AFE amplifier was verified,
demonstrating a transient interval less than 18 ns. Considering a typical center frequency
of medical ultrasound applications, it is a sufficiently small value, estimated not
to generate the imaging artifacts associated with the discrete TGC operation.
Fig. 6. Chip micrograph and layout.
Fig. 7. Measured frequency response of AFE amplifier.
Fig. 8. TGC operations in measurements: (a) a case of applying a single-frequency sinusoid signal; (b) a case of applying exponentially attenuated echo signals.
Fig. 9. Simulated the input-referred voltage noise density of AFE amplifier.
Table 1. Performance comparison of TGC amplifiers
|
This work
|
JSSC’18 [9]
|
JSSC’20 [7]
|
JSSC’18 [11]
|
TCAS-II’22 [8]
|
JSTS’14 [10]
|
Process
|
180 nm
|
180 nm
|
180 nm
|
180 nm
|
180 nm
|
350 nm
|
Gain control
|
Discrete
|
Discrete
|
Continuous
|
Discrete
|
Continuous
|
Discrete
|
Type
|
Voltage amp
|
Voltage amp
|
TIA
|
TIA
|
Voltage amp
|
Voltage amp
|
Real-time TGC verification
|
O
|
N/A
|
O
|
N/A
|
O
|
O
|
Transient interval
|
≤ 18 ns
|
N/A
|
N/A
|
N/A
|
N/A
|
≤ 100 ns
|
Gain range (gain step)
|
16.2 dB (5.4 dB)
|
46 dB (4.5 dB)
|
33 dBW
|
12 dBW (6 dBW)
|
37 dB
|
37 dB (2.5 dB)
|
Bandwidth
|
4.5 MHz
|
11.9 MHz
|
7 MHz
|
16 MHz
|
3.1 MHz
|
8 MHz
|
Input referred noise
|
6.3 nV/√Hz
@ 4 MHz
|
6.3 nV/√Hz
@ 5 MHz
|
1.7 pA/√Hz
@ 5 MHz
|
3 pA/√Hz (1)
@ 16 MHz
|
8.6 nV/√Hz
@ 2 MHz
|
N/A
|
P1dB
|
8.5 dBm
|
N/A
|
N/A
|
N/A
|
5.1 dBm
|
9.5 dBm
|
Power consumption
|
2.3 mW(2)
|
0.6 mW(2)
|
5.2 mW(3)
|
3.3 mW(2)
|
0.96 mW(3)
|
1.9 mW(3)
|
(1) Estimated from the measurement results
(2) Power consumption of analog front-end circuits including a preamplifier and a
variable/programmable gain amplifier
(3) Power consumption of a variable/programmable gain amplifier
V. CONCLUSIONS
This paper presents an AFE amplifier with a seamless discrete TGC operation for an
ultrasound scanner. The proposed AFE amplifier consists of a preamplifier and a PGA.
We analyze an issue of fluctuations of operating points of amplifier during a gain
switching moment of discrete TGC operation. To mitigate this issue, we propose a preset
operation during a reset phase, minimizing a discrepancy of voltage levels in critical
floating nodes. Additionally, we propose a current-reuse op amp with dynamic push-pull
current sources. We employ an FVF to replace the constant tail current sources, further
improving a transient response for a discrete TGC operation. The proposed AFE amplifier
has been implemented in a 180-nm BCD process, occupying an active area of 0.19 mm$^{2}$
with a power consumption of 2.3 mW. In measurements, a gain range of AFE amplifier
is 16.2 dB with a gain step of 5.4 dB. A real-time discrete TGC operation is also
demonstrated, showing a transient interval less than 18 ns at a gain switching moment.
Considering a target center frequency of ultrasound, the measured transient interval
value is estimated not to result in the imaging artifacts associated with the gain
switching.
ACKNOWLEDGMENTS
This work was supported in part by the National Research Foundation (NRF) grants
funded by the Korea Governments (2019R1I1A3A01060591, RS-2023-00244635), the Commercialization
Promotion Agency for R&D Outcomes grant funded by the Korea Government (MSIT, 2021I100),
and Korea Institute for Advancement of Technology grant funded by the Korea Government
(MOTIE, P0017011). The EDA tools were supported by IDEC, Korea.
References
W. Du et al., “Conformable ultrasound breast patch for deep tissue scanning and imaging,”
Sci. Adv., vol. 9, no. 30, eadh5325, Jul. 2023.
Y. M. Hopf et al., “A pitch-matched transceiver ASIC with shared hybrid beamforming
ADC for high-frame-rate 3-D intracardiac echocardio-graphy,” IEEE J. Solid-State Circuits,
vol. 57, no. 11, pp. 3228-3242, Nov. 2022.
B. Jana et al., “Smarphone-based point-of-care system using continuous-wave portable
Doppler,” IEEE Trans. Instrum. Meas., vol. 69, no. 10, pp. 8352-8361, Oct. 2020.
H.-T. Park, and J.-Y. Um, “Image-free ultrasound blood-flow monitoring circuit system
with automatic range-gate positioning scheme: a pilot study,” Appl. Sci., vol. 11,
no. 22, Nov. 2021.
A. K. Sahani et al., “An imageless ultrasound device to measure local and regional
arterial stiffness,” IEEE Trans. Biomed. Circuit Syst., vol. 10, no. 1, pp. 200-208,
Feb. 2016.
W. R. Hedrick et al., Ultrasound Physics and Instrumentations. Philadephia, PA, USA:
Elsvier, 2005.
E. Kang et al., “A variable-gain low-noise transimpedance amplifier for miniature
ultrasound probes,” IEEE J. Solid-State Circuits, vol. 55, no. 12, pp. 3157-3168,
Dec. 2020.
J.-Y. Um, “A compact variable gain amplifier with continuous time-gain compensation
using systematic predistorted gain control,” IEEE Trans. Circuits Syst. II, Exp. Briefs,
vol. 69, no. 2, pp. 274-278, Feb. 2022.
C. Chen et al., “A pitch-matched front-end ASIC with integrated subarray beamforming
ADC for miniature 3-D ultrasound probes,” IEEE J. Solid-State Circuits, vol. 53, no.
11, pp. 3050-3064, Nov. 2018.
S.-E. Cho et al., “A single-stage 37dB-linear digitally-controlled variable gain amplifier
for ultrasound medical imaging,” IEIE J. Semicond. Technol. Sci., vol. 14, no. 5,
pp. 579-587, Oct. 2014.
M. Tan et al., “A front-end ASIC with high-voltage transmit switching and receive
digitization for 3-D forward-looking intravascular ultrasound imaging,” IEEE J. Solid-State
Circuits, vol. 53, no. 8, pp. 2284-2297, Aug. 2018.
L. Shen, N. Lu, and N. Sun, “A 1-V 0.25-mW inverter stacking amplifier with 1.07 noise
efficiency factor,” IEEE J. Solid-State Circuits, vol. 53, no. 2, pp. 896-905, Mar.
2018.
J. Zhang, H. Zang, Q. Sun, and R. Zhang, “A low-noise, low-power amplifier with current-reused
OTA for ECG recordings,” IEEE Trans. Biomed. Circuit Syst., vol. 12, no. 3, pp. 700-708,
Jun. 2018.
A. J. Lopez-Martin et al., “Low-voltage supper class AB CMOS OTA cells with very high
slew rate and power efficiency,” IEEE J. Solid-State Circuits, vol. 40, no. 5, pp.
1068-1077, May 2005.
Min-Hyeong Son received the B.S. degree in Department of Electronic Engineering
from Hannam University, Daejeon, Korea, in 2022. He is currently pursuing the M.S.
degree in Department of Medical IT Convergence Engineering from Kumoh National Institute
of Technology, Gumi, Korea. His research interests include analog front-end circuits
and analog-to-digital converters for sensor interface.
Ji-Yong Um received the B.S., M.S. and Ph.D. degrees in Department of Electronic
and Electrical Engineering from Pohang University of Science and Technology (POSTECH),
Pohang, Korea, in 2006, 2008, and 2013, respectively. From 2014 to 2016, he was a
Senior Engineer at SK Hynix, Korea. From 2016 to 2021, He was an Assistant Professor
in Department of Electronic Engineering at Hannam University. In 2021, he joined the
faculty of Kumoh National Institute of Technology, where he is currently an Assistant
Professor in Department of Medical IT Convergence Engineering. His research interests
include biomedical circuits and systems for medical ultrasound and physiological signal
instrumentations.