sensors Article

Power MOSFET Linearizer of a High-Voltage Power Amplifier for High-Frequency Pulse-Echo Instrumentation Hojong Choi 1 , Park Chul Woo 1 , Jung-Yeol Yeom 2, * and Changhan Yoon 3, * 1 2 3

*

Department of Medical IT Convergence Engineering, Kumoh National Institute of Technology, Gumi 39253, Korea; [email protected] (H.C.); [email protected] (P.C.W.) School of Biomedical Engineering, Korea University, Seoul 02841, Korea Department of Biomedical Engineering, Inje University, Gimhae 50834, Korea Correspondence: [email protected] (J.-Y.Y.); [email protected] (C.Y.); Tel.: +82-2-3290-5662 (J.-Y.Y.); +82-55-320-3301 (C.Y.)

Academic Editors: Xiaoning Jiang and Chao Zhang Received: 28 February 2017; Accepted: 31 March 2017; Published: 4 April 2017

Abstract: A power MOSFET linearizer is proposed for a high-voltage power amplifier (HVPA) used in high-frequency pulse-echo instrumentation. The power MOSFET linearizer is composed of a DC bias-controlled series power MOSFET shunt with parallel inductors and capacitors. The proposed scheme is designed to improve the gain deviation characteristics of the HVPA at higher input powers. By controlling the MOSFET bias voltage in the linearizer, the gain reduction into the HVPA was compensated, thereby reducing the echo harmonic distortion components generated by the ultrasonic transducers. In order to verify the performance improvement of the HVPA implementing the power MOSFET linearizer, we measured and found that the gain deviation of the power MOSFET linearizer integrated with HVPA under 10 V DC bias voltage was reduced (−1.8 and −0.96 dB, respectively) compared to that of the HVPA without the power MOSFET linearizer (−2.95 and −3.0 dB, respectively) when 70 and 80 MHz, three-cycle, and 26 dBm input pulse waveforms are applied, respectively. The input 1-dB compression point (an index of linearity) of the HVPA with power MOSFET linearizer (24.17 and 26.19 dBm at 70 and 80 MHz, respectively) at 10 V DC bias voltage was increased compared to that of HVPA without the power MOSFET linearizer (22.03 and 22.13 dBm at 70 and 80 MHz, respectively). To further verify the reduction of the echo harmonic distortion components generated by the ultrasonic transducers, the pulse-echo responses in the pulse-echo instrumentation were compared when using HVPA with and without the power MOSFET linearizer. When three-cycle 26 dBm input power was applied, the second, third, fourth, and fifth harmonic distortion components of a 75 MHz transducer driven by the HVPA with power MOSFET linearizer (−48.34, −44.21, −48.34, and −46.56 dB, respectively) were lower than that of the HVPA without the power MOSFET linearizer (−45.61, −41.57, −45.01, and −45.51 dB, respectively). When five-cycle 20 dBm input power was applied, the second, third, fourth, and fifth harmonic distortions of the HVPA with the power MOSFET linearizer (−41.54, −41.80, −48.86, and −46.27 dB, respectively) were also lower than that of the HVPA without the power MOSFET linearizer (−25.85, −43.56, −49.04, and −49.24 dB, respectively). Therefore, we conclude that the power MOSFET linearizer could reduce gain deviation of the HVPA, thus reducing the echo signal harmonic distortions generated by the high-frequency ultrasonic transducers in pulse-echo instrumentation. Keywords: high-voltage power amplifier; ultrasonic transducer; power MOSFET linearizer

Sensors 2017, 17, 764; doi:10.3390/s17040764

www.mdpi.com/journal/sensors

improvements of the HVPA, such as gain, bandwidth, and linearity have been one of the critical technical issues. Especially, linearity is an important parameter for HVPA used in the ultrasound transmitters because the non-linear HVPA leads to high harmonic distortions, thus deteriorating the echo signal quality of the ultrasonic transducers [5]. For high-frequency ultrasonic transducers compared to low-frequency ultrasonic transducers, Sensors 2017, 17, 764 2 of 12 the size of the piezoelectric materials (the main material of the ultrasonic transducers) is small, thus reducing the capacitance values of the piezoelectric materials as the operating frequency of the 1. Introduction ultrasonic transducers are increased [6,7]. Therefore, the parasitic capacitances of the coaxial cables can affect the performances of the high-frequency transducers. addition, the maximum A pulse-echo instrumentation typically ultrasonic consists of ultrasonicIn transmitters, receivers, excitation voltages into the high-frequency ultrasonic transducers are, accordingly, limited(HVPA) due to and transducers [1–3]. In the ultrasonic transmitters, the high voltage power amplifier the size reduction of the components piezoelectricsince materials [8]. triggers These external interventions affect the is one of the main electronic the HVPA the ultrasonic transducers through performances of high-frequency pulse-echo instrumentation. Therefore, linearizing techniques, such the expander devices, which consist of cross-coupled diodes [2,4]. Therefore, the performance as compensating active and passive components improvements of for the several HVPA, non-linear such as gain, bandwidth, and electronic linearity have been oneinofthe theHVPA, critical could help improve HVPA linearity, thus improving the performance of high-frequency pulse-echo technical issues. Especially, linearity is an important parameter for HVPA used in the ultrasound instrumentations. Thethe basic operating principle linearizer and HVPA is described in Figurethe 1. transmitters because non-linear HVPA leadsof tothe high harmonic distortions, thus deteriorating As depicted in the figure, the gain of the HVPA typically decreases as the input power increases echo signal quality of the ultrasonic transducers [5]. due to thehigh-frequency sensitive parasitic impedances [9–12]. compared to low-frequency ultrasonic transducers, For ultrasonic transducers Linearizer techniques have been widely researched applications [11]. the size of the piezoelectric materials (the main materialinofRF thecommunication ultrasonic transducers) is small, However, the linearizer techniques only been recently introduced the ultrasound research thus reducing the capacitance valueshave of the piezoelectric materials as thein operating frequency of the [13,14]. The systematic approaches use the analog-to-digital converter, digital-to-analog converter, ultrasonic transducers are increased [6,7]. Therefore, the parasitic capacitances of the coaxial cables and which lead to high-frequency increased cost, larger size,transducers. and higher complexity can memory affect thespaces, performances of the ultrasonic In addition,[13]. the Therefore, maximum itexcitation is not preferable for implementation into ultrasound array transducer-based systems. voltages into the high-frequency ultrasonic transducers are, accordingly, limited due toThe the pre-linearizer technique provides low cost, small external size, and low complexity compared to the size reduction of the piezoelectric materials [8]. These interventions affect the performances of systematic approaches [11].instrumentation. However, the reported pre-linearizer requires complex mathematical high-frequency pulse-echo Therefore, linearizing techniques, such as compensating model analysis to compensate non-linearity the HVPAinand it is not suitable for the for several non-linear active andthe passive electronic in components the HVPA, could help improve ultrasound HVPA utilizing input and output DC coupling capacitors [14]. Therefore, we develop a HVPA linearity, thus improving the performance of high-frequency pulse-echo instrumentations. simple linearizer circuit to beof used for HVPAand ultrasound As shown Figure 1, The basic operating principle the linearizer HVPA isapplications. described in Figure 1. Asindepicted in the the linearizers need to be designed to have opposite characteristics with the HVPA. Therefore, the figure, the gain of the HVPA typically decreases as the input power increases due to the sensitive linearizer gain deviation should be positively increased to reduce the gain deviation of the HVPA. parasitic impedances [9–12].

Linearizer Gain deviation

HVPA + Linearizer HVPA Input power

Figure1.1.Gain Gaindeviation deviationof ofHVPA HVPAwith withand andwithout without the the linearizer linearizer as as aa function function of of input input power. power. Figure

Linearizer been widely researched in RF communication applications [11]. The HVPA techniques in ultrasonichave transmitters usually utilizes large-value coupling capacitors in order However, linearizer techniques have onlythe been introduced the ultrasound to block DCthe voltages which in turn deteriorate echorecently signal quality of thein performances of research [13,14].pulse-echo The systematic approacheslike usehigh the analog-to-digital converter, digital-to-analog high-frequency instrumentations harmonic distortion components of the echo converter, and memory spaces, which lead to increased cost, and input higherpowers complexity [13]. signals because the coupling capacitors reduce the gain of thelarger HVPAsize, at high [2,15,16]. Therefore, is not preferable implementation ultrasound array transducer-based systems. In a typicalit HVPA topology, for large-value couplinginto capacitors in the input and output ports are The pre-linearizer technique low cost,the small and low complexity compared to the required to block DC voltages provides as they can affect echosize, signal quality of the ultrasonic transducers. systematic approaches [11]. However, pre-linearizer requirescomponents complex mathematical Unfortunately, these coupling capacitorsthe arereported inherently non-linear passive such that it model analysisunwanted to compensate the non-linearity in the HVPA it is notdistortions suitable forgenerated the ultrasound also generates harmonic distortions. Therefore, theand harmonic from HVPA utilizing input and output DC coupling capacitors [14]. Therefore, we develop a simple linearizer circuit to be used for HVPA ultrasound applications. As shown in Figure 1, the linearizers need to be designed to have opposite characteristics with the HVPA. Therefore, the linearizer gain deviation should be positively increased to reduce the gain deviation of the HVPA. The HVPA in ultrasonic transmitters usually utilizes large-value coupling capacitors in order to block DC voltages which in turn deteriorate the echo signal quality of the performances of high-frequency pulse-echo instrumentations like high harmonic distortion components of the echo signals because the coupling capacitors reduce the gain of the HVPA at high input powers [2,15,16]. In a typical HVPA topology, large-value coupling capacitors in the input and output ports are required to block DC voltages as they can affect the echo signal quality of the ultrasonic transducers.

Sensors 2017, 17, 764

3 of 12

Sensors 2017, 17, 764 3 ofit 12 Unfortunately, these coupling capacitors are inherently non-linear passive components such that also generates unwanted harmonic distortions. Therefore, the harmonic distortions generated from HVPAdirectly directlyaffect affectthe theperformance performanceofofultrasonic ultrasonictransducers. transducers.InInorder ordertotoreduce reducethe theharmonic harmonic HVPA distortion components of the echo signal generated by the HVPA, we propose employing simple distortion components of the echo signal generated by the HVPA, we propose employing simple linearizertechniques techniquesutilizing utilizingvariable variablepower powerMOSFET MOSFETdevices deviceswith withshunt shuntinductors inductorsand andcapacitors. capacitors. linearizer The detailed architecture, operating mechanism, and implementation of the linearizer an The detailed architecture, operating mechanism, and implementation of the linearizer with anwith HVPA HVPA are described in 2. Section 2. The capability to the reduce the harmonic distortions of the echo are described in Section The capability to reduce harmonic distortions of the echo signals signals of the ultrasonic transducers are verified with the developed linearizer with an HVPA in of the ultrasonic transducers are verified with the developed linearizer with an HVPA in Section 3. Section 3. The conclusion of the paper is given in Section 4. The conclusion of the paper is given in Section 4.

Materialsand andMethods Methods 2.2.Materials Figure2a,b 2a,bshow showschematic schematicdiagrams diagramsof ofthe theHVPA HVPAand andthe thepower powerMOSFET MOSFETlinearizer. linearizer.In Inthe the Figure HVPAarchitecture architecture(Figure (Figure2a), 2a),the theHVPA HVPAisisaaClass ClassA-type A-typepower poweramplifier amplifierused usedininthe theultrasound ultrasound HVPA applications.The Thelarge large value pF coupling DC coupling capacitors (C1C2and 2) are DC applications. value 100100 pF DC capacitors (C1 and ) areCused toused blockto DCblock voltage. voltage. The two variable resistors (R1 and ) are to of setthethe bias voltage of the The two variable 2 kΩ resistors2(RkΩ R2 ) are used to setR2the biasused voltage high-voltage transistor 1 and high-voltage transistor (T 2 ). The 3.3 μH RF choke inductor (L 1 ) is used to maximize the voltage (T ). The 3.3 µH RF choke inductor (L ) is used to maximize the voltage amplification of HVPA. 2 1 amplification of the The 2 μH inductor (Lb1) is used to prevent bias voltage A 28 The 2 µH inductor (Lb1HVPA. ) is used to prevent bias voltage reduction. A 28 V DC voltage (Vreduction. DC ) is supplied V DCa voltage (VDC) is In supplied from a power supply. In a typical HVPA topology, large value to RF from power supply. a typical HVPA topology, a large value RF choke inductora is required choke the inductor is required to reduce the DC voltage the power supply (VDC), reduce DC voltage drop from the power supply (VDCdrop ), thusfrom allowing the output voltage of thus the allowing theamplified output voltage the HVPA to be amplified without someoutput suppression the HVPA to be withoutofsome suppression because the maximum voltagebecause is limited maximum outputsupply voltage[12]. is limited by the DCresistor power issupply [12]. If aoflarge valuevalue resistor used by the DC power If a large value used instead the large RF is choke instead of largesupply value DC RF choke theaffect power DC voltage drop could affect the inductor, thethe power voltageinductor, drop could thesupply maximum output voltage amplification. maximum output amplification. Bias inductors with values are also set the Bias inductors withvoltage large values are also needed to set the DClarge bias voltage after theneeded resistortodivider DC bias the resistor divider circuit (R1 bias and voltage R2). Therefore, the maximum DC gate bias circuit (R1voltage and R2 ).after Therefore, the maximum DC gate is generated by the resistor divider voltage is generated by the resistor divider circuits. If we use a resistor instead of the inductor, there circuits. If we use a resistor instead of the inductor, there would be an additional DC voltage drop and would be limit an additional DC voltage drop and that could limit the output voltage amplification [12]. that could the output voltage amplification [12].

(a)

(b)

Figure2.2.Schematic Schematicdiagrams diagramsof of(a) (a)HVPA HVPAand and(b) (b)power powerMOSFET MOSFETlinearizer. linearizer. Figure

As shown in Figure 2b, we designed a power MOSFET linearizer which consists of shunt As shown in Figure 2b, we designed a power MOSFET linearizer which consists of shunt capacitors and inductors with power MOSFET devices working with variable DC bias voltages. The capacitors and inductors with power MOSFET devices working with variable DC bias voltages. 3.3 μH inductor value (LL3) is used to minimize the voltage drop from the power supply (VL1). The The 3.3 µH inductor value (LL3 ) is used to minimize the voltage drop from the power supply (VL1 ). 100 pF capacitor (CL3) is used to apply stable DC bias voltage to the power MOSFET device (IRF620, The 100 pF capacitor (CL3 ) is used to apply stable DC bias voltage to the power MOSFET device Vishay Siliconix, Malvern, PA, USA) because DC power noise signals could go to ground. Two (IRF620, Vishay Siliconix, Malvern, PA, USA) because DC power noise signals could go to ground. capacitors (CL1 and CL2) are used to short possible unwanted AC noise from the power supply (VL1). Two capacitors (CL1 and CL2 ) are used to short possible unwanted AC noise from the power supply The 50-Ω coaxial cable is used to minimize the signal fluctuation between the power MOSFET (VL1 ). The 50-Ω coaxial cable is used to minimize the signal fluctuation between the power MOSFET linearizer and the HVPA. linearizer and the HVPA. Figure 3 shows the equivalent circuit model of the power MOSFET linearizer. As shown in Figure 3, depending on the gate bias voltage, the MOSFET in the linearizer works as a variable resistor (TR1) to provide variable performance when used in conjunction with two fixed 20 pF and 100 nH values of capacitors (CL1 and CL2) and inductors (LL1 and LL2). These capacitors are bypass capacitors to block possible noise at the input port generated by the function generator and the

Sensors 2017, 17, 764

4 of 12

Figure 3 shows the equivalent circuit model of the power MOSFET linearizer. As shown in Figure 3, depending on the gate bias voltage, the MOSFET in the linearizer works as a variable resistor (TR1 ) to provide variable performance when used in conjunction with two fixed 20 pF and 100 nH values of capacitors (CL1 and CL2 ) and inductors (LL1 and LL2 ). These capacitors are bypass capacitors to block possible noise at the input port generated by the function generator and the power Sensors 2017, 17, 764 4 of 12 supply [12,17]. Low frequency ultrasound applications may not require bypass capacitors, however, high-frequency ultrasound are very sensitive to the performance of thebypass HVPAcapacitors, due to the power supply [12,17]. Lowapplications frequency ultrasound applications may not require limited maximum output power of the ultrasonic transducers However, as large capacitance however, high-frequency ultrasound applications are very [8]. sensitive to the performance of can the affect the performance of the circuit, and the MOSFET (T ) inherently has large parasitic capacitance, 1 HVPA due to the limited maximum output power of the ultrasonic transducers [8]. However, as smaller value bypass areperformance preferred here. theand 20 pF value selected tohas be large capacitance cancapacitor affect the of Therefore, the circuit, thecapacitor MOSFET (T1) isinherently much less thancapacitance, the parasiticsmaller capacitances the MOSFET and output capacitances 23020and large parasitic value of bypass capacitor(input are preferred here. Therefore, =the pF 70 pF). The total output of the power MOSFET linearizer can be expressed by: capacitor value is selected to be much less than the parasitic capacitances of the MOSFET (input and  power MOSFET linearizer can be  output capacitances = 230 and 70 pF). The total output of the expressed by:   1     1 1 + jwCds Vin1 1 jwLL2   (1) ≈  + Vin−1 ≈  (1)    1 1  +   1    1 11 + jwCgd + 1 TR1 1 jwL L2 + jwCds + +

where w w is is the the operating operating frequency. frequency. where

LL3

CL3 TR1

In-1

LL1 Cgs +CL1

Cgd

Vin1 CL2 +Cds

LL2

Figure 3. The model of of the the power power MOSFET MOSFET linearizer. linearizer. Figure 3. The equivalent equivalent circuit circuit model

As shown in Equation (1), the performance of the power MOSFET linearizer is determined by As shown in Equation (1), the performance of the power MOSFET linearizer is determined the variable resistor of the MOSFET (TR1) and inductor (LL2) because the MOSFET device has larger by the variable resistor of the MOSFET (TR1 ) and inductor (LL2 ) because the MOSFET device has pre-determined parasitic capacitances (Cgd and Cds) compared to the two capacitors (CL1 and CL2). larger pre-determined parasitic capacitances (Cgd and Cds ) compared to the two capacitors (CL1 and Therefore, the performance of the power MOSFET linearizer with different inductor values (LL2) CL2 ). Therefore, the performance of the power MOSFET linearizer with different inductor values were measured in order to find the appropriate operating frequencies of the power MOSFET (LL2 ) were measured in order to find the appropriate operating frequencies of the power MOSFET linearizer because the power MOSFET linearizer could reduce the gain of the HVPA at certain linearizer because the power MOSFET linearizer could reduce the gain of the HVPA at certain operating operating frequencies. The 100, 33, 22, and 12 nH inductors were selected to have minimum loss of frequencies. The 100, 33, 22, and 12 nH inductors were selected to have minimum loss of the MOSFET the MOSFET linearizer at 50, 70, 80, and 90 MHz operating frequencies. We selected a 33 nH linearizer at 50, 70, 80, and 90 MHz operating frequencies. We selected a 33 nH inductance value to inductance value to minimize the signal loss (lower than −3.8 dB) measured at 70 and 80 MHz. With minimize the signal loss (lower than −3.8 dB) measured at 70 and 80 MHz. With the selected inductor the selected inductor values, the performances of the MOSFET linearizer could essentially be values, the performances of the MOSFET linearizer could essentially be controlled depending on controlled depending on the MOSFET resistance (TR1). In Section 3, the power MOSFET linearizer is the MOSFET resistance (TR1 ). In Section 3, the power MOSFET linearizer is characterized because characterized because the variable resistances of the MOSFET can be varied under different DC the variable resistances of the MOSFET can be varied under different DC values (VL1 ) as shown in values (VL1) as shown in Figure 2b. Figure 2b. 3. Results and Discussion 3.1. Linearizer Performance Measurement The HVPA and power MOSFET linearizer were implemented on a printed circuit board with two layers. In the HVPA, the heat-sink was attached on top of the high-voltage transistors as shown

Sensors 2017, 17, 764

5 of 12

3. Results and Discussion 3.1. Linearizer Performance Measurement The HVPA and power MOSFET linearizer were implemented on a printed circuit board with two layers. In the HVPA, the heat-sink was attached on top of the high-voltage transistors as shown in Figure Sensors 17, 55 of Sensors 2017, 2017, 17, 764 764 4. Our target ultrasonic transducer frequency for the pulse-echo instrumentations of 12 12 is over 50 MHz, as it is known that in studies with the eye, intravascular, acoustic stimulation, resolutions compared to low-frequency than 15 MHz)applications ultrasonic applications [18–21]. and cavitation, high-frequency (over 50(less MHz) ultrasonic have reported betterFigure spatial5 shows the experimental setup to measure the gain deviation at 1-dB compression point of the resolutions compared to low-frequency (less than 15 MHz) ultrasonic applications [18–21]. Figure 5 HVPA the with and withoutsetup the power MOSFET linearizer in order to acquire the proper bias shows experimental to measure the gain deviation at 1-dB compression pointworking of the HVPA voltages the power MOSFET linearizer. The gate-threshold voltage of the MOSFET the power with and of without the power MOSFET linearizer in order to acquire the proper working in bias voltages MOSFET linearizer is typically 2 V and 4 V,voltage depending the drain-source voltage such of the power MOSFET linearizer.between The gate-threshold of theon MOSFET in the power MOSFET that the power MOSFET linearizer performs as a variable over the gate-threshold voltages. linearizer is typically between 2 V and 4 V, depending on theresistor drain-source voltage such that the power The three-cycle 70 or 80 MHzas input pulsesresistor generated an arbitrary signal generator (AFG3252, MOSFET linearizer performs a variable overfrom the gate-threshold voltages. The three-cycle Tektronix Inc.,input Beaverton, OR, USA) from wereanapplied to signal the HVPA without and with the power 70 or 80 MHz pulses generated arbitrary generator (AFG3252, Tektronix Inc., MOSFET linearizer preset DC biasing a high-voltage (E3631A, Beaverton, OR, USA)under were applied to the HVPAwith without and with thepower power supply MOSFET linearizerAgilent under Technologies, Santa CA, USA). Thesupply output was then fed Technologies, through twoSanta 20 dB power preset DC biasing withClara, a high-voltage power (E3631A, Agilent Clara, CA, attenuators (BW-S20W20+, Brooklyn, NY, USA) and acquired with anMini-circuits, oscilloscope USA). The output was then Mini-circuits, fed through two 20 dB power attenuators (BW-S20W20+, (MSOX4154A, Keysight Technology, Santa Clara, CA, USA) to show the Technology, gain deviations the Brooklyn, NY, USA) and acquired with an oscilloscope (MSOX4154A, Keysight Santaof Clara, HVPA with without thedeviations power MOSFET linearizer. CA, USA) toand show the gain of the HVPA with and without the power MOSFET linearizer.

Figure board. Figure 4. 4. The The implemented implemented power power MOSFET MOSFET linearizer linearizer and and HVPA HVPA on on aa printed printed circuit circuit board. board.

Figure 5. The measurement setup for with and without the power MOSFET linearizer. Figure for HVPA HVPA with with and and without without the the power power MOSFET MOSFET linearizer. linearizer. Figure 5. 5. The The measurement measurement setup setup for HVPA

In order to assess the power MOSFET linearizer characteristics, we measured the voltage gain In order to assess power MOSFET linearizer welinearizer measured voltage gain deviation graphs of thethe HVPA with and without the characteristics, power MOSFET at the 70 and 80 MHz deviation graphs of the HVPA with and without the power MOSFET linearizer at 70 and 80 MHz under 5 V DC bias. The gain deviation graphs of (1) the power MOSFET linearizer; (2) the HVPA under 5 V(3) DCthe bias. The with gain the deviation of (1) the power linearizer; (2) the HVPA only, and HVPA power graphs MOSFET linearizer at 70MOSFET MHz input power are plotted in only, HVPA with power MOSFET linearizer at 70linearizer MHz input in Figureand 6a. (3) Thethe measured gainthe deviations of the power MOSFET andpower HVPAare at plotted 20 and 26 Figure 6a. The measured gain deviations of the power MOSFET linearizer and HVPA at 20deviations and 26 dBof m dBmm input power were 0.06 and 0.35 dB, and −0.265 and −2.95 dB, respectively. The gain input power were 0.06 and 0.35 dB, and − 0.265 and − 2.95 dB, respectively. The gain deviations of the the HVPA with the power MOSFET linearizer reduced to −0.261 and −2.09 dB, respectively. The gain HVPA with the power MOSFETlinearizer, linearizer HVPA, reducedand to − 0.261 with and − 2.09 dB, respectively. The gain deviation graph of the MOSFET HVPA the power MOSFET linearizer at 70 MHz input power is plotted in Figure 6b. The measured gain deviations of the linearizer at 20 and 26 dBmm input power were 0.25 and 0.64 dB, respectively, and those of the HVPA at 20 dBmm and 26 dBmm input power were −0.33 and −3.00 dB, respectively. The gain deviations of the HVPA with a power MOSFET linearizer improved to −0.16 and −1.30 dB, respectively. From these results, we can conclude that the gain deviation of the HVPA with the MOSFET linearizer was suppressed as the

Sensors 2017, 17, 764

6 of 12

deviation graph of the MOSFET linearizer, HVPA, and HVPA with the power MOSFET linearizer at 70 MHz input power is plotted in Figure 6b. The measured gain deviations of the linearizer at 20 and 26 dBm input power were 0.25 and 0.64 dB, respectively, and those of the HVPA at 20 dBm and 26 dBm input power were −0.33 and −3.00 dB, respectively. The gain deviations of the HVPA with a power MOSFET linearizer improved to −0.16 and −1.30 dB, respectively. From these results, we can conclude that the gain deviation of the HVPA with the MOSFET linearizer was suppressed as the input power increased. The difference in results when 70 and 80 MHz input power were applied to the power MOSFET linearizer is due to the non-linear frequency-dependent power MOSFET device parameters (total parasitic capacitances and inductance in the equivalent circuit model of the power MOSFET Sensors 2017, 17, 764 6 of 12 devices) shown in Figure 3. As derived in Equation (1), the frequency-dependent gain characteristic of thecharacteristic power MOSFET is dependent on the inductanceon (LL2 ) and two parasitic capacitances of thelinearizer power MOSFET linearizer is dependent the inductance (LL2) and two (Cds and C ). parasitic gd capacitances (Cds and Cgd). Power MOSFET linearizer HVPA 0.06 dB at 20 dBm 0.35 dB HVPA 1 with power MOSFET linearizer at 26 dBm

-1

-0.265 dB at 20 dBm

-0.261 dB at 20 dBm

-2 -3 -4

-2.95 dB at 26 dBm

-10

0 10 20 Input power (dBm)

(a)

-2.09 dB at 26 dBm

30

Gain deviation (dB)

Gain deviation (dB)

0

power MOSFET linearizer HVPA 0.25 dB at 20 dBm 0.64 dB HVPA at 26 dBm 1 with power MOSFET linearizer 2

2

0 -1

-0.33 dB at 20 dBm

-2 -3 -4

-3.00 dB at 26 dBm

-10

0 10 20 Input power (dBm)

-0.16 dB at 20 dBm -1.30 dB at 26 dBm

30

(b)

Figure 6. The gain deviation graphs of the power MOSFET linearizer, HVPA, and HVPA with

Figure 6. The gain deviation graphs of the power MOSFET linearizer, HVPA, and HVPA with power power MOSFET linearizer when (a) 70 and (b) 80 MHz input power were applied. MOSFET linearizer when (a) 70 and (b) 80 MHz input power were applied.

Figure 7 shows the gain deviation graphs of the HVPA with and without the MOSFET linearizer several DC deviation bias voltages at 70ofand MHz,with respectively. In Figure 7a, whenlinearizer 70 Figure 7under shows the gain graphs the 80 HVPA and without the MOSFET MHz and 26 dB m input powers were applied to the power MOSFET linearizer, the gain deviation under several DC bias voltages at 70 and 80 MHz, respectively. In Figure 7a, when 70 MHz and 26 dBm values underwere 6, 9, applied and 10 V conditions (−2.01, −1.85, the and gain −1.80deviation dB, respectively) were 6, 9, input powers to DC the bias power MOSFET linearizer, values under reduced compared to 1 and 3 V DC bias conditions (−2.25 and −2.16 dB, respectively). In Figure 7b, to 1 and 10 V DC bias conditions (−2.01, −1.85, and −1.80 dB, respectively) were reduced compared when 70 MHz and 26 dB m input powers were applied to the power MOSFET linearizer, the gain and 3 V DC bias conditions (−2.25 and −2.16 dB, respectively). In Figure 7b, when 70 MHz and 26 dBm deviation under 11, 12, 13, 14, and 15 V DC bias conditions (−1.80, −1.82, −1.80, −1.83, and −1.82 dB, input powers were applied to the power MOSFET linearizer, the gain deviation under 11, 12, 13, 14, respectively) were still lower compared to that of the HVPA (−2.95 dB). In Figure 7a,b, when 70 and 15 V DC bias conditions (−1.80, −1.82, −1.80, −1.83, and −1.82 dB, respectively) were still lower MHz and 26 dBm input power were applied to the HVPA with and without the power MOSFET compared tothe thatgain of the HVPA values (−2.95of dB). Figurewith 7a,b, when 70 MOSFET MHz andlinearizer 26 dBm input linearizer, deviation theInHVPA the power were power higher were applied to the HVPA with and without the power MOSFET linearizer, the gain deviation values than that of the HVPA without the MOSFET linearizer (−2.95 dB). In Figure 7b, when 80 MHz andof the HVPA the powers power MOSFET linearizer higher thanlinearizer, that of thethe HVPA MOSFET 26 dBwith m input were applied to the were power MOSFET gain without deviationthe values linearizer Figure when 80(−1.11, MHz and 26and dBm−0.96 inputdB, powers were applied to the power under 6,(− 9,2.95 anddB). 10 VInDC bias7b, conditions −1.01, respectively) were reduced MOSFET linearizer, gain deviation under 6, 9,and and−1.31 10 VdB, DCrespectively). bias conditions (−1.11, when compared tothe 1 and 3V DC bias values conditions (−1.37 In Figure 7d,−1.01, when 80 dB, MHzrespectively) and 26 dBm input applied to the to power the gain(−1.37 and −0.96 were powers reducedwere when compared 1 andMOSFET 3 V DClinearizer, bias conditions deviation under 11, 12, 13, 14,Inand 15 V 7d, DC when bias conditions −1.00, −1.00, −1.07,were and applied −1.01 and −1.31 dB, respectively). Figure 80 MHz were and 26 dBm−1.06, input powers to dB, respectively. In Figure 7c,d, when 80 MHz and 26 dB m input power were applied to the HVPA the power MOSFET linearizer, the gain deviation under 11, 12, 13, 14, and 15 V DC bias conditions with and without power MOSFET linearizer, the gain deviation values the HVPA with were −1.00, −1.06, −the 1.00, −1.07, and −1.01 dB, respectively. In Figure 7c,d,ofwhen 80 MHz andthe 26 dBm power MOSFET linearizer were higher than that of the HVPA without the power MOSFET input power were applied to the HVPA with and without the power MOSFET linearizer, the gain linearizer (−3.00 dB). Figure 7e,f show the normalized gain deviation of the HVPA with and without deviation values of the HVPA with the power MOSFET linearizer were higher than that of the HVPA the MOSFET linearizer vs. the output power under 1, 3, 6, 9, and 10 V DC bias conditions, when 70 without the power MOSFET linearizer (−3.00 dB). Figure 7e,f show the normalized gain deviation of and 80 MHz input were applied to the power MOSFET linearizer respectively. As shown in Figure the7,HVPA with and without the MOSFET vs. the output power 3, 6, 9,MOSFET and 10 V DC the gain deviation performances of thelinearizer HVPA improved with the helpunder of the 1,power bias conditions, when 70 and 80 MHz input were applied to the power MOSFET linearizer linearizer over high input power ranges. These experimental results illustrate that therespectively. gain Asdeviation shown inofFigure 7, the gain performances the with the the HVPA with thedeviation power MOSFET linearizerof can beHVPA alteredimproved by controlling the DChelp biasof the feed voltages of the MOSFET. This fact is also meaningful because high-frequency ultrasonic transducers requires a relatively higher input power from the HVPA to achieve reasonable echo sensitivity compared to low-frequency ultrasonic transducers [1,4,8].

Sensors 2017, 17, 764

7 of 12

0.5

0.5

0.0

0.0

-0.5 HVPA only 1V 3V 6V - 2.16(3V) - 2.25 (1V) 9V 10 V

-1.0 -1.5 -2.0 -2.5 -3.0

-10

0 10 20 Input power (dBm)

- 1.8 (10V) - 1.85(9V) - 2.01(6V)

- 2.95

Gain deviation (dB)

Gain deviation (dB)

power MOSFET linearizer over high input power ranges. These experimental results illustrate that the gain deviation of the HVPA with the power MOSFET linearizer can be altered by controlling the DC bias feed voltages of the MOSFET. This fact is also meaningful because high-frequency ultrasonic transducers requires a relatively higher input power from the HVPA to achieve reasonable echo sensitivity to low-frequency ultrasonic transducers [1,4,8]. Sensors 2017,compared 17, 764 7 of 12

-0.5 HVPA only 11 V 12 V - 1.80 (11V) 13 V - 1.82(12V) 14 V 15 V

-1.0 -1.5 -2.0 -2.5 -3.0

30

-10

(a)

-0.5

- 0.96 (10V)

HVPA only 1V 3V 6V 9V 10 V

-1.0 -1.5 -2.0 -2.5 -10

- 1.31 (3V) - 1.37 (1V)

- 1.01(9V) - 1.11(6V)

- 3.00

0 10 20 Input power (dBm)

Gain deviation (dB)

Gain deviation (dB)

30

(b)

0.0

-3.0

0.0 -0.5

- 1.00(13V)

HVPA only 11 V 12 V 13 V 14 V 15 V

-1.0 -1.5 -2.0 -2.5 -3.0

30

-10

(c)

- 1.00 (11V) - 1.06(12V)

0 10 20 Input power (dBm)

- 1.01(15V) - 1.07(14V)

- 3.00

30

(d) 0.5

0.0 -0.5 1V HVPA only 3V 6V 9V 10 V

-1.0 -1.5 -2.0 -2.5 0

- 2.16(3V) - 2.25 (1V)

10 20 30 Output power (dBm)

(e)

- 1.8 (10V) - 1.85(9V) - 2.01(6V)

- 2.95

40

Gain deviation (dB)

0.5 Gain deviation (dB)

- 2.95

0.5

0.5

-3.0

0 10 20 Input power (dBm)

- 1.80 (13V) -1.82(15V) - 1.83(14V)

0.0 -0.5

- 0.96 (10V)

-1.0

HVPA only - 1.31 (3V) 1V - 1.37 (1V) 3V 6V 9V 10 V

-1.5 -2.0 -2.5 -3.0

0

10 20 30 Output power (dBm)

- 1.01(9V) - 1.11(6V)

- 3.00

40

(f)

Figure 7. Normalized gain deviation of the HVPA with and without a power MOSFET linearizer vs. Figure 7. Normalized gain deviation of the HVPA with and without a power MOSFET linearizer vs. input power at (a) 70 MHz under DC bias voltages (1, 3, 6, 9, and 10 V DC, respectively), (b) under input power at (a) 70 MHz under DC bias voltages (1, 3, 6, 9, and 10 V DC, respectively), (b) under DC DC bias voltages (11, 12, 13, 14, and 15 V DC, respectively), at (c) 80 MHz under DC bias voltages (1, bias voltages (11, 12, 13, 14, and 15 V DC, respectively), at (c) 80 MHz under DC bias voltages (1, 3, 3, 6, 9, and 10 V DC, respectively), at (d) 80 MHz under DC bias voltages (11, 12, 13, 14, and 15 V DC, 6, 9, and 10 V DC, respectively), at (d) 80 MHz under DC bias voltages (11, 12, 13, 14, and 15 V DC, respectively). Normalized gain deviation of the HVPA with and without a power MOSFET respectively). Normalized gain deviation of the HVPA with and without a power MOSFET linearizer linearizer vs. output power at (e) 70 MHz under DC bias voltages (1, 3, 6, 9, and 10 V DC, vs. output power at (e) 70 MHz under DC bias voltages (1, 3, 6, 9, and 10 V DC, respectively), and (f) at respectively), and (f) at 80 MHz under DC bias voltages (1, 3, 6, 9, and 10 V DC, respectively). 80 MHz under DC bias voltages (1, 3, 6, 9, and 10 V DC, respectively).

The input 1-dB compression point (IP1dB) provides an index of the linearity of the power amplifier components [22]. In fact, HVPA higheran IPindex 1dB hasof lower gain suppression than that The input 1-dB compression point (IP1dBwith ) provides the linearity of the power amplifier with lower IP 1dB. In Figure summarizes the measured IP1dB of the HVPA with and without power components [22]. fact,8HVPA with higher IP1dB has lower gain suppression than thatthe with lower linearizer at 70 and 80 MHz, respectively. As shown in Figure 8a, when 70 MHz, 26 dB m IPMOSFET . Figure 8 summarizes the measured IP of the HVPA with and without the power MOSFET 1dB 1dB and 3-cycle pulsed powers were sent to the HVPA with and without the power MOSFET linearizer, linearizer at 70 and 80 MHz, respectively. As shown in Figure 8a, when 70 MHz, 26 dBm and 3-cycle the measured IP1dB of thetoHVPA without the power MOSFET linearizer (22.03linearizer, dBm) was lower than pulsed powers were sent the HVPA with and without the power MOSFET the measured that with the power MOSFET linearizer under 10V DC bias voltage (24.17 dBm). In Figure 8b, when the 80 MHz, 26 dBm, and three-cycle pulsed powers were sent to the HVPA with and without the power MOSFET linearizer, the measured IP1dB of the HVPA without the power MOSFET linearizer (22.13 dBm) was lower than that with power MOSFET linearizer under 10 V DC bias voltage (26.19 dBm). Therefore, we confirm that the power MOSFET linearizer increased the IP1dB of the HVPA.

Sensors 2017, 17, 764

8 of 12

IP1dB of the HVPA without the power MOSFET linearizer (22.03 dBm ) was lower than that with the power MOSFET linearizer under 10 V DC bias voltage (24.17 dBm ). In Figure 8b, when the 80 MHz, 26 dBm , and three-cycle pulsed powers were sent to the HVPA with and without the power MOSFET linearizer, the measured IP1dB of the HVPA without the power MOSFET linearizer (22.13 dBm ) was lower than that with power MOSFET linearizer under 10 V DC bias voltage (26.19 dBm ). Therefore, we confirm that the power MOSFET linearizer increased the IP1dB of the HVPA. Sensors 2017, 17, 764 8 of 12 Input 1-dB compression points

Input 1-dB compression points

0 23.88(6V)

-1

22.03(HVPA Only) 23.43 (1V) HVPA Only 23.56 (3V)

1V 3V 6V 9V 10 V

-2

-3

0

10 20 Input power (dBm)

23.97 (9V)

24.17(10V)

30

Gain deviation (dB)

Gain deviation (dB)

0

25.89(6V)

-1

-2

-3

22.13(HVPA Only)

HVPA Only 24.43 (1V) 24.85 (3V) 1V 3V 6V 9V 10 V

0

10 20 Input power (dBm)

(a)

26.16 (9V)

26.19(10V)

30

(b)

Figure 8. Measured 1-dB compression point graphs of the HVPA with and without power MOSFET

Figure 8. Measured 1-dB compression point graphs of the HVPA with and without power MOSFET linearizer at (a) 70 MHz and (b) 80 MHz under DC bias voltages (1, 3, 6, 9, and 10 V DC, linearizer at (a) 70 MHz and (b) 80 MHz under DC bias voltages (1, 3, 6, 9, and 10 V DC, respectively). respectively).

Table 1 summarizesthe theperformances performances ofofthe with andand without the linearizer at 70 and Table 1 summarizes theHVPA HVPA with without the linearizer at 70 and 80 MHz under different DC bias voltages. 80 MHz under different DC bias voltages. Table 1. Summary of the performances of the HVPA with and without the power MOSFET

Table 1. Summary of the performances of the HVPA with and without the power MOSFET linearizer linearizer under various DC bias voltages. under various DC bias voltages. HVPA with the Power MOSFET Linearizer HVPA without (DC Bias Voltages) Power MOSFET HVPA with the Power MOSFET Linearizer HVPA without 1 V Linearizer 3V 6 V 9V 10 V (DC Bias Voltages) Power MOSFET−2.25 Gain deviation at 70 MHz (dB) −2.95 −2.16 −2.01 −1.85 −1.80 Linearizer −1.37 1 V −1.31 3 V −1.11 6 V −1.01 9 V −0.9610 V Gain deviation at 80 MHz (dB) −3.00 1dB at 70 MHz m) 22.03−2.95 23.43 −2.2523.56 −2.1623.88 −2.0123.97 −1.85 24.17−1.80 GainIPdeviation at 70(dB MHz (dB) 1dB at 80 MHz m) 22.13−3.00 24.43 −1.3724.85 −1.3125.89 −1.1126.16 −1.01 26.19−0.96 GainIPdeviation at 80(dB MHz (dB)

IP1dB at 70 MHz (dBm ) 22.03 at 80 MHz (dB ) 22.13 3.2.IPHigh-Frequency Pulse-Echo Instrumentation m 1dB

23.43 24.43

23.56 24.85

23.88 25.89

23.97 26.16

24.17 26.19

Figure 9 shows the measurement setup for high-frequency pulse-echo instrumentation using

3.2. the High-Frequency Pulse-Echo Instrumentation HVPA with the power MOSFET linearizer. The DC voltages generated by a high-voltage power supply (E3631A, Agilent Technologies, San Jose, CA, USA) were applied to the power MOSFET Figure 9 shows the measurement setup for high-frequency pulse-echo instrumentation using linearizer and HVPA and the pulse signals were generated by an arbitrary signal generator the (AFG3252C, HVPA with Tektronix the powerInc., MOSFET linearizer. voltages generated by a The high-voltage Beaverton, OR, The USA)DC were fed to the HVPA. amplifiedpower supply (E3631A, Agilent Technologies, San Jose, CA, USA) were applied to the power MOSFET high-voltage signals from the HVPA are transmitted through an expander into the ultrasonic linearizer and HVPA and the pulse signals were generated by an arbitrary signal generator (AFG3252C, transducer. The acoustic echo signals reflected from a quartz target were converted to the electrical Tektronix Inc., Beaverton, OR, USA) were fedMeanwhile, to the HVPA. The amplified high-voltage signals echo signals by the ultrasonic transducers. the discharged high-voltage signals sent from HVPA are blocked by the limiter before the into electrical echo signalstransducer. pass throughThe the acoustic limiter. echo the from HVPA are transmitted through an expander the ultrasonic The echo signals were amplified by a were preamplifier (AU-1114, Technologies Newby York, NY, signals reflected from a quartz target converted to theL3 electrical echo Inc., signals the ultrasonic USA) with a DC power supply (E3630A, Agilent Technologies) and displayed the echo signals andby the transducers. Meanwhile, the discharged high-voltage signals sent from HVPA are blocked their spectra on the oscilloscope (MSOX4154A, Keysight Technology, Santa Clara, CA, USA). limiter before the electrical echo signals pass through the limiter. The echo signals were amplified Figure 10 shows the measured pulse-echo responses of the HVPA with and without the power by a preamplifier (AU-1114, L3 Technologies Inc., New York, NY, USA) with a DC power supply MOSFET linearizer in the high-frequency pulse-echo instrumentation using a 75 MHz (E3630A, Agilent Technologies) and displayed theOlympus echo signals and theirTokyo, spectra on the oscilloscope immersion-focused ultrasonic transducer (V3349, Inc., Shinjuku, Japan) when 75 (MSOX4154A, Keysight Technology, Santa Clara, CA, USA). MHz three-cycle 26 dBm input power was applied to the function generated. In Figure 10a,b, the echo Figure 10 thewith measured pulse-echo responses of the with the amplitude of shows the HVPA the power MOSFET linearizer (13.92 mV)HVPA is higher thanand that without of the HVPA without the power MOSFET linearizer (12.35 mV). In Figure instrumentation 10c,d, the second, third, power MOSFET linearizer in the high-frequency pulse-echo usingfourth, a 75 MHz and fifth harmonic distortion components of a 75 MHz ultrasonic transducer (V3349) applied by the HVPA with the power MOSFET linearizer (−48.34, −44.21, −48.34, and −46.56 dB, respectively) are lower than those of the HVPA (−45.61, −41.57, −45.01, and −45.51 dB, respectively). The −6 dB

Sensors 2017, 17, 764

9 of 12

immersion-focused ultrasonic transducer (V3349, Olympus Inc., Shinjuku, Tokyo, Japan) when 75 MHz three-cycle 26 dBm input power was applied to the function generated. In Figure 10a,b, the echo amplitude of the HVPA with the power MOSFET linearizer (13.92 mV) is higher than that of the HVPA without the power MOSFET linearizer (12.35 mV). In Figure 10c,d, the second, third, fourth, and fifth harmonic distortion components of a 75 MHz ultrasonic transducer (V3349) applied by the HVPA with theSensors power MOSFET linearizer (−48.34, −44.21, −48.34, and −46.56 dB, respectively) are lower 2017, 17, 99of 12 Sensors 2017, 17,764 764 ofthan 12 those of the HVPA (−45.61, −41.57, −45.01, and −45.51 dB, respectively). The −6 dB bandwidth of the signal using HVPA with power MOSFET is bandwidth ofusing the echo echo signalwith using the HVPA with the the power (36.48%) MOSFETislinearizer linearizer (36.48%) is to thebandwidth echo signalof the HVPA thethe power MOSFET linearizer increased(36.48%) compared increased compared to that using the HVPA without the power MOSFET linearizer (33.46%). increased compared to that using the HVPA without the power MOSFET linearizer (33.46%). that using the HVPA without the power MOSFET linearizer (33.46%).

Figure The measurement setup Figure 9.The Themeasurement measurementsetup setup for for the the high frequency pulse-echo instrumentation. Figure 9.9. the highfrequency frequencypulse-echo pulse-echoinstrumentation. instrumentation.

0.004 0.004

33Cycle Cycle26 26dB dBmmHVPA HVPA

33Cycle Cycle26 26dBm dBmHVPA HVPA++Power PowerMOSFET MOSFETLinearizer Linearizer

0.008 0.008

12.35 12.35mV mV

Amplitude (V) (V) Amplitude

Amplitude (V) (V) Amplitude

0.008 0.008

0.000 0.000

0.000 0.000

36.2 36.2

36.4 36.4 36.6 36.6 Time Time(us) (us)

36.8 36.8

37.0 37.0

-0.008 -0.008 36.0 36.0

33Cycle Cycle26 26dB dBmmHVPA HVPA 00 --66dB dBBW BW==33 33.46 .46% % HD3 HD3== HD2 HD2== --41.57 HD4== 41.57dB dB HD4 -20 -20 --45.61 45.61dB dB --45.01 45.01dB dB

100 200 300 400 100 200 300 400 Frequency Frequency(MHz) (MHz)

(c) (c)

36.4 36.4 36.6 36.6 Time Time(us) (us)

36.8 36.8

37.0 37.0

33Cycle Cycle26 26dBm dBmHVPA HVPA++Power PowerMOSFET MOSFETLinearizer Linearizer

00

-20 -20

HD5 HD5== --45.51 45.51dB dB

-40 -40

36.2 36.2

(b) (b) Normalized Spectrum Spectrum (dB) (dB) Normalized

Normalized Spectrum Spectrum (dB) (dB) Normalized

(a) (a)

-60 -60 00

13.92 13.92mV mV

-0.004 -0.004

-0.004 -0.004 -0.008 -0.008 36.0 36.0

0.004 0.004

500 500

-40 -40

-60 -60 00

--66dB dBBW BW==36.48 36.48% % HD3 HD3== --44.21 44.21dB dB HD2 HD2== HD5 HD5== HD4 HD4== --48.34 dB 48.34 dB - 48.34 dB --46.56 46.56dB dB - 48.34 dB

100 100 200 200 300 300 400 400 Frequency Frequency(MHz) (MHz)

500 500

(d) (d)

Figure 10. The echo amplitude of the HVPA (a) without (b) with the power MOSFET linearizer Figure The echo amplitudeof ofthe theHVPA HVPA(a) (a) without without and and linearizer Figure 10.10. The echo amplitude and (b) (b) with withthe thepower powerMOSFET MOSFET linearizer and and the the spectrum spectrum of of the the HVPA HVPA (c) (c) without without and and (d) (d) with with the the power power MOSFET MOSFET linearizer linearizer when when and the spectrum of the HVPA (c) without and (d) with the power MOSFET linearizer when three-cycle three-cycle three-cycle 26 26dB dBmm input input power power was was applied. applied. 26 dBm input power was applied.

Figure Figure 11 11 shows shows the the measured measured pulse-echo pulse-echo responses responses of of the the HVPA HVPA with with and and without without the the power power MOSFET MOSFET linearizer linearizer using using aa 75 75 MHz MHz immersion-focused immersion-focused ultrasonic ultrasonic transducer transducer (V3349) (V3349) when when 75 75 MHz MHz five-cycle 20 dB m input power was applied. In Figure 11a,b, the echo amplitude of the HVPA five-cycle 20 dBm input power was applied. In Figure 11a,b, the echo amplitude of the HVPA with with the the power power MOSFET MOSFET linearizer linearizer (10.22 (10.22 mV) mV) is is higher higher than than that that of of the the HVPA HVPA (9.95 (9.95 mV). mV). In In Figure Figure 11c,d, 11c,d, the the second, second, third, third, fourth, fourth, and and fifth fifth harmonic harmonic distortions distortions of of the the transducer transducer applied applied by by the the HVPA HVPA with with the the power power MOSFET MOSFET linearizer linearizer (−41.54, (−41.54, −41.80, −41.80, −48.86, −48.86, and and −46.27 −46.27 dB, dB, respectively) respectively) are are lower lower than than

Sensors 2017, 17, 764

10 of 12

Figure 11 shows the measured pulse-echo responses of the HVPA with and without the power MOSFET linearizer using a 75 MHz immersion-focused ultrasonic transducer (V3349) when 75 MHz five-cycle 20 dBm input power was applied. In Figure 11a,b, the echo amplitude of the HVPA with the power MOSFET linearizer (10.22 mV) is higher than that of the HVPA (9.95 mV). In Figure 11c,d, the second, third, fourth, and fifth harmonic distortions of the transducer applied by the HVPA with the power MOSFET linearizer (−41.54, −41.80, −48.86, and −46.27 dB, respectively) are lower Sensors 2017, 17, 764 10than of 12 those of the HVPA without the power MOSFET linearizer (−25.85, −43.56, −49.04, and −49.24 dB, respectively). respectively). The The −6 −6dB dBbandwidth bandwidthofofthe theecho echosignal signalusing usingthe theHVPA HVPA with with the the power power MOSFET linearizer (30.01%) (30.01%)increased increased compared to that the HVPA theMOSFET power MOSFET compared to that usingusing the HVPA withoutwithout the power linearizer linearizer in the measurement of the pulse-echo power (17.35%). (17.35%). As shownAs in shown the measurement results of results the pulse-echo responses,responses, the powerthe MOSFET MOSFET could the harmonic unwanteddistortion harmonic distortion and components andamplitudes increase the linearizer linearizer could reduce the reduce unwanted components increase the of amplitudes of the echo signal generated by 75 MHz ultrasonic transducers. the echo signal generated by 75 MHz ultrasonic transducers. 0.008

5 Cycle 20 dBm HVPA + Power MOSFET Linearizer

5 Cycle 20 dBm HVPA

0.008

Amplitude (V)

Amplitude (V)

0.004 9.95 mV

0.000

0.004 0.000

-0.004

-0.004 -0.008 36.0

10.22 mV

36.2

36.4 36.6 Time (us)

36.8

37.0

-0.008 36.0

36.2

(a)

- 49.04 dB

HD5 = - 49.24 dB

-40

100 200 300 Frequency (MHz)

400

Normalized Spectrum (dB)

Normalized Spectrum (dB)

- 6 dB BW =17.35 % HD2 = - 25.85 dB HD3 = - 43.56 dB HD4 =

0

37.0

(b)

0

-60

36.8

5 Cycle 20 dBm HVPA + Power MOSFET Linearizer

5 Cycle 20 dBm HVPA

-20

36.4 36.6 Time (us)

500

(c)

0 - 6 dB BW =30.01 %

-10

HD3 = HD2 = - 41.80 dB - 41.54 dB HD4 =

-20 -30

HD5 = - 48.86 dB - 46.27 dB

-40 -50 -60

0

100

200 300 400 Frequency (MHz)

500

(d)

Figure 11. The Figure 11. The echo echo signal signal amplitude amplitude when when using using the the HVPA HVPA (a) (a) without without and and (b) (b) with with the the power power MOSFET MOSFET linearizer linearizer and and the the spectrum spectrum of of the the HVPA HVPA (c) (c) without without and and (d) (d) with with the the power power MOSFET MOSFET linearizer linearizer when when five-cycle five-cycle 20 20 dB dBm input input power power is is applied. applied. m

Finally, while circuit topologies and operation frequencies are different, making direct Finally, while circuit topologies and operation frequencies are different, making direct comparison comparison difficult, Table 2 summarize the performances of our linearizer with that of selected difficult, Table 2 summarize the performances of our linearizer with that of selected previously previously reported linearizers. reported linearizers. Table 2. Performance comparison of our linearizer and a previously-reported design, including pulse-echo performances with ultrasound transducers.

Parameters Output power Operating frequency Second harmonic distortion (HD2) Third harmonic distortion (HD3)

[13] 55.1 dBm 5 MHz −50 dB -

[14] 41.59 dBm 140 MHz −24.8 dB -

Our Design 37.6 dBm 80 MHz −41. 54 dB −41.80 dB

4. Conclusions Conventional HVPA used in the high-frequency pulse-echo instrumentation includes DC coupling capacitors to minimize the high-voltage DC effects on the high-frequency ultrasound

Sensors 2017, 17, 764

11 of 12

Table 2. Performance comparison of our linearizer and a previously-reported design, including pulse-echo performances with ultrasound transducers. Parameters

[13]

[14]

Our Design

Output power Operating frequency Second harmonic distortion (HD2) Third harmonic distortion (HD3)

55.1 dBm 5 MHz −50 dB -

41.59 dBm 140 MHz −24.8 dB -

37.6 dBm 80 MHz −41. 54 dB −41.80 dB

4. Conclusions Conventional HVPA used in the high-frequency pulse-echo instrumentation includes DC coupling capacitors to minimize the high-voltage DC effects on the high-frequency ultrasound transducer. However, this structure could reduce the gain of the HVPA with higher input powers. Harmonic distortions of the echo signals can severely limit the performance of high-frequency pulse-echo instrumentation. Therefore, a linear HVPA to trigger the ultrasonic transducer is highly desirable for such high-frequency pulse-echo instrumentation. Here, we proposed a novel power MOSFET linearizer in order to increase the linearity (reducing gain deviations) over a wide range of input power. The gain deviation of the power MOSFET linearizer has shown to have the opposite effect to that of the HVPA, thus increasing gain flatness over a wide input power range. Therefore, the power MOSFET linearizer can decrease the harmonic distortions of the HVPA, thereby reducing the unwanted harmonic distortions of the ultrasonic transducer in high-frequency pulse-echo instrumentation. The structure of the power MOSFET linearizer is simple to implement because, in the power MOSFET linearizer structure, the MOSFET acts as a variable resistor to primarily affect the gain deviation performances. Therefore, the operation of the MOSFET under varying DC bias conditions in the power MOSFET linearizer was characterized for the HVPA. At 70 and 80 MHz, the measured gain deviation values of the HVPA with power MOSFET linearizer under 10 V DC bias voltage (−2.95 dB and −3.00 dB, respectively) were lower than that of the HVPA without power MOSFET linearizer (−1.80 dB and −1.11 dB, respectively). At 70 and 80 MHz, the measured IP1dB point of the HVPA with power MOSFET linearizer under 10 V DC bias voltage (24.17 and 26.19 dBm , respectively) were higher than that of the HVPA without power MOSFET linearizer (22.03 and 22.13 dBm , respectively). In order to verify the capability of the power MOSFET linearizer, the echo signal harmonic distortions of the 75 MHz ultrasonic transducer were measured. Using 75 MHz 5-cycle 20 dBm input power, the echo signal harmonic distortions using HVPA with power MOSFET linearizer (−41.54, −41.80, −48.86, and −46.27 dB, respectively) were measured to be lower than those of HVPA without the power MOSFET linearizer (−25.85, −43.56, −49.04, and −49.24 dB, respectively). We conclude that the power MOSFET linearizer has proven to be a simple, yet effective, technique for the HVPA to reduce unwanted harmonic distortions in the echo signals while providing higher echo signal amplitude and even wider echo signal bandwidth. Acknowledgments: This work has been supported in part by National Research Foundation of Korea (NRF-2016M2B2A9A02945226) and by the MSIP (Ministry of Science, ICT and Future Planning), Korea, under the ITRC (Information Technology Research Center) support program (IITP-2016-R2718-16-0015) supervised by the IITP (National It Industry Promotion Agency). Author Contributions: H.C., J.-Y.Y. and C.Y. conceived the idea. J.-Y.Y. and P.C.W. performed the experiments; H.C., J.-Y.Y. and C.Y. wrote the paper together. Conflicts of Interest: The authors declare no conflict of interest.

References 1.

Shung, K.K. Diagnostic Ultrasound: Imaging and Blood Flow Measurements; Taylor & Francis: Boca Raton, FL, USA, 2015.

Sensors 2017, 17, 764

2. 3.

4. 5. 6.

7.

8. 9. 10. 11. 12. 13.

14. 15.

16. 17. 18. 19. 20.

21.

22.

12 of 12

Qiu, W.; Yu, Y.; Tsang, F.K.; Sun, L. A multifunctional, reconfigurable pulse generator for high-frequency ultrasound imaging. IEEE Trans. Ultrason. Ferroelectr. Freq. Control 2012, 59, 1558–1567. [PubMed] Qiu, W.; Wang, C.; Li, Y.; Zhou, J.; Yang, G.; Xiao, Y.; Feng, G.; Jin, Q.; Mu, P.; Qian, M.; et al. A scanning-mode 2D shear wave imaging (s2D-SWI) system for ultrasound elastography. Ultrasonics 2015, 62, 89–96. [CrossRef] [PubMed] Szabo, T.L. Diagnostic Ultrasound Imaging: Inside Out; Elsevier Academic Press: London, UK, 2004. Song, J.; Kim, S.; Sohn, H.-Y.; Song, T.-K.; Yoo, Y.M. Coded excitation for ultrasound tissue harmonic imaging. Ultrasonics 2010, 50, 613–619. [CrossRef] [PubMed] Zhu, B.; Xu, J.; Li, Y.; Wang, T.; Xiong, K.; Lee, C.; Yang, X.; Shiiba, M.; Takeuchi, S.; Zhou, Q. Micro-particle manipulation by single beam acoustic tweezers based on hydrothermal PZT thick film. AIP Adv. 2016, 6, 035102. [CrossRef] [PubMed] Zhu, X.; Guo, J.; He, C.; Geng, H.; Yu, G.; Li, J.; Zheng, H.; Ji, X.; Yan, F. Ultrasound triggered image-guided drug delivery to inhibit vascular reconstruction via paclitaxel-loaded microbubbles. Sci. Rep. 2016, 6, 21683. [CrossRef] [PubMed] Zhou, Q.; Lam, K.H.; Zheng, H.; Qiu, W.; Shung, K.K. Piezoelectric single crystal ultrasonic transducers for biomedical applications. Prog. Mater. Sci. 2014, 66, 87–111. [CrossRef] [PubMed] Vuolevi, J.; Rahkonen, T. Distortion in RF Power Amplifiers; Artech House: London, UK, 2003. Kazimierczuk, M.K. RF Power Amplifier; John Wiley & Sons: Hoboken, NJ, USA, 2014. Cripps, S.C. Advanced Techniques in RF Power Amplifier Design; Artech House: Norwood, MA, USA, 2002. Lee, T.H. The Design of CMOS Radio-Frequency Integrated Circuits; Cambridge University Press: Cambridge, UK, 2004. Gao, Z.; Gui, P.; Jordanger, R. An integrated high-voltage low-distortion current-feedback linear power amplifier for ultrasound transmitters using digital predistortion and dynamic current biasing techniques. IEEE Trans. Circuits Syst. II Express Briefs 2014, 61, 373–377. [CrossRef] Choi, H.; Jung, H.; Shung, K.K. Power Amplifier Linearizer for High Frequency Medical Ultrasound Applications. J. Med. Biol. Eng. 2015, 35, 226–235. [CrossRef] [PubMed] Park, J.; Hu, C.; Li, X.; Zhou, Q.; Shung, K.K. Wideband linear power amplifier for high-frequency ultrasonic coded excitation imaging. IEEE Trans. Ultrason. Ferroelectr. Freq. Control 2012, 59, 825–832. [CrossRef] [PubMed] Svilainis, L.; Motiejunas, G. Power amplifier for ultrasonic transducer excitation. Ultragarsas 2006, 1, 30–36. Razavi, B. Design of Analog CMOS Integrated Circuits, 1st ed.; McGraw-Hill Science: New York, NY, USA, 2000. Shung, K.K. Diagnostic Ultrasound: Past, Present, and Future. J. Med. Biol. Eng. 2011, 31, 371–374. [CrossRef] Li, X.; Wei, W.; Zhou, Q.; Shung, K.K.; Chen, Z. Intravascular photoacoustic imaging at 35 and 80 MHz. J. Biomed. Opt. 2012, 17, 106005. [CrossRef] [PubMed] Li, X.; Li, J.; Jing, J.; Ma, T.; Liang, S.; Zhang, J.; Mohar, D.; Raney, A.; Mahon, S.; Brenner, M. Integrated IVUS-OCT imaging for atherosclerotic plaque characterization. IEEE J. Sel. Top. Quantum Electron. 2014, 20, 196–203. Qiu, W.; Wang, X.; Chen, Y.; Fu, Q.; Su, M.; Zhang, L.; Xia, J.; Dai, J.; Zhang, Y.; Zheng, H. A Modulated Excitation Imaging System for Intravascular Ultrasound. IEEE Trans. Biomed. Eng. 2016. [CrossRef] [PubMed] Katz, A. Linearization: Reducing distortion in power amplifiers. IEEE Microw. Mag. 2001, 2, 37–49. [CrossRef] © 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).