* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Class H Power Amplifier for Power Saving in Fluxgate Current
Electrical ballast wikipedia , lookup
Power over Ethernet wikipedia , lookup
Electrical substation wikipedia , lookup
Three-phase electric power wikipedia , lookup
Mercury-arc valve wikipedia , lookup
Electric power system wikipedia , lookup
Electrification wikipedia , lookup
Power inverter wikipedia , lookup
Stray voltage wikipedia , lookup
Current source wikipedia , lookup
Variable-frequency drive wikipedia , lookup
Voltage regulator wikipedia , lookup
History of electric power transmission wikipedia , lookup
Power engineering wikipedia , lookup
Surge protector wikipedia , lookup
Pulse-width modulation wikipedia , lookup
Voltage optimisation wikipedia , lookup
Resistive opto-isolator wikipedia , lookup
Audio power wikipedia , lookup
Mains electricity wikipedia , lookup
Power electronics wikipedia , lookup
Buck converter wikipedia , lookup
Alternating current wikipedia , lookup
> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < 1 Class H Power Amplifier for Power Saving in Fluxgate Current Transducers Guillermo Velasco-Quesada, Member, IEEE, Manuel Román-Lumbreras, Member, IEEE, Raúl Pérez-Delgado, and Alfonso Conesa-Roca Index Terms— AC and DC current measurement, fluxgate transducer, energy savings, class H amplifier. electrical power due to its inherent very high efficiency [6]. This work evaluates the energy efficiency improvement reached in a fluxgate based current transducer when the output PA is replaced by a class D or by a class H amplifiers. In Section II, the basic operation of a current transducer based on fluxgate effect will be introduced. In Section III, the use of different output PA architectures is evaluated and the potential energy savings will be analyzed. A practical implementation of a class H amplifier for this application is described in section IV. The practical results obtained, in terms of functionality and energy savings, are presented in section V, when the features of the designed transducer are tested and different architectures of output PA are compared. I. INTRODUCTION II. FLUXGATE CURRENT TRANSDUCER OPERATION HE ever-rise energy cost, due to the growing demand for energy and the emerging shortage of fossil resources, and the need for greenhouse gas emissions reduction, lead to increased global awareness of the importance of energy savings and energy efficiency. Addressing the energy and climate challenges requires an interdisciplinary approach involving, among others, technological solutions. The power electronics engineers are constantly working on increasing the efficiency of power electronics and power systems through the use of new and faster switching devices, new topologies and advanced controls [1]. However, the energy efficiency is an emerging watchword on issues related to industrial electronics design [2-3], where the structure of power amplifiers (PA) is one of the most promising topics, since they may be the biggest energy consumers in these systems. The architecture of PA is usually based on a class AB output stage, which is characterized by features such as low total harmonic distortion (THD) and low energy efficiency. There are currently two technologies, referred to as “Green Amplifiers” [4], which allows energy savings when they are compared with class AB based designs. Class G and H amplifiers are based on class AB amplifier with adjustable power supply voltages, where these values are adapted in accordance to the amplifier output voltage level [5]. On the other hand, class D amplifiers, also called “switching amplifiers”, use a different technology because they incorporate a switched mode amplifier based in pulse width modulation (PWM) technique. This operating principle saves considerable The most suitable method for measuring electrical current depends on both, the current characteristics and the considered application. In this regard, current transducers based on fluxgate effect are used for the measurement of DC and AC currents of high values with a very high accuracy. Furthermore, in references [7-9] some reviews devoted to the description of electric current sensors technologies can be found. Although the basic operation of fluxgate current transducer is briefly described in this section, more details about this technology [10-12] and technical advances [13-19] can be found in the specialized literature, being also possible find this type of devices in catalogs of sector leading companies [20-21]. The current measurement process in a fluxgate device is based on the principle of magnetic flux compensation: The flux (ΦP) produced in a magnetic core by the main current to be measured (named primary current: IP) is compensated by an opposite flux (ΦS), created by a current flowing in the compensation winding, which is wired on the same core (named secondary current: IS), so that the resulting flux is zero (Fig. 1). Abstract—This paper presents a new improvement in the design of a fluxgate-based current transducers in order to reduce the power consumption of control electronics. The proposed improvement involves the replacement of the output linear amplifier of the transducer by a class H amplifier. The output amplifier is devoted to the magnetic flux compensation and generates the transducer output current, which is proportional to the current to be measured. In this way, it is possible to reduce significantly the power consumption of these current transducers without affecting their performance in terms of linearity, accuracy and bandwidth. T Fig. 1. Closed loop operation principle of the fluxgate transducer. > REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < The fluxgate effect and the suitable control electronics are used to detect the condition of zero resulting flux. In the case of zero flux, the secondary current (IS) is proportional to the primary current (IP), according to the number of turns ratio (NP and NS) of the windings through which these two currents are flowing. Eq. (1) shows this relationship when NP = 1. I P NS I S (1) 2 A. Class AB amplifier The method commonly used to generate the secondary current (IS) is based on the utilization of a class AB linear amplifier (Fig. 4). This solution perfectly meets the high performance required by fluxgate transducers, but has the disadvantage of poor features associated with its low efficiency. In this case, the power consumed by electronics (PI), increases linearly with the value of the secondary current. III. GENERATION OF CURRENT COMPENSATION The zero flux condition in the magnetic circuit, in which the primary and secondary currents are involved, is obtained when the average value of the fluxgate excitation current generated by an auxiliary winding (NA) is zero (Fig. 2). Fig. 4. Structure of class AB amplifier. Fig. 2. Fluxgate excitation current IA(t) with no flux due to the primary current. The voltage across the excitation winding is VA(t). When the primary current is zero (IP = 0), the average value of the excitation current is zero (<IA> = 0), and the waveform of this current is as shown Fig. 2. However, when the primary current is nonzero, or a residual flux exists in the magnetic circuit, the average value of the fluxgate excitation current is also nonzero (<IA> ≠ 0), as shown Fig. 3. The circuit associated with the compensating winding of the transducer and the measuring resistor is a LR circuit, where R is the series association of the measuring (or burden) resistor (RB) and the compensation winding intrinsic resistance (RS), and L is the inductance of the compensation winding (LS). In this way, the amplifier output power (PO) matches the power consumed in this resistance (R). Input and output power and the efficiency (η) of the PA are formulated in Eqs. (2). PI VCC I S PO R I S When the average value of the fluxgate excitation current is zero with nonzero primary current, means that the flux due to the primary current (IP) is equal and opposite of the flux due to the secondary current (IS). In this case, IS is proportional to IP according to the windings turns ratio (Eq. (1)). The secondary current (IS) is generated by the control system, and it is supplied by an output PA with the condition of zero average value of the excitation current as operation set point. Given the wide range of operation of these transducers (large full-scale), and the high accuracy that they can achieve (up to 0.001%), it is mandatory the use of an accurate, linear and free of interferences output PA. As a consequence, the architecture used for the implementation of these PA is usually based on a traditional class AB linear amplifier. In this context, this work explores the use of high efficiency architectures for the implementation of this output PA, specifically the two “green” architectures introduced in the previous section. The obtained results, in terms of power consumption, are compared with the use of the class AB PA. PO R I S PI VCC (2) If the R value is selected such that R·IN = VCC, where IN is the maximum value that the current transducer can measure, and assuming an ideal class AB amplifier, that can deliver on its output up to the supply voltage VCC (rail-to-rail operation), then the PA efficiency can be expressed by Eq. (3). Fig. 3. Fluxgate excitation current IA(t) with non-zero flow due to a non-zero primary current. The voltage across the excitation winding is VA(t). 2 IS IN (3) As is known, and as shows Eq. (3), the class AB amplifier performance is strongly dependent on the supplied current that in this case is the secondary current (IS). The difference between the input and the output power (PA) is dissipated as heat in the PA and this power can be calculated through Eq. (4). PA PI PO VCC I S R I S I S VCC R I S 2 (4) The maximum value of the PA dissipated power is reached accordingly to Eq. (5). dPA 0 dI S IS VCC I N 2 R 2 50 % (5) This low power efficiency leads to a high power dissipation and usually requires the use of heat-sinks for an effective heat management. B. Class D amplifier A class D amplifier is a pulse width modulation (PWM) > REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < switched mode amplifier. The analog output voltage of class D PA is related to the average value of the PWM signal used as control input signal. Fig. 5 shows the structure of a class D amplifier based on a half-bridge converter. It is formed by two N channel MOSFETs (M1 and M2) and one driver circuit. These components are usually integrated into a single commercial chip when a low and medium power applications (up to 300 W) are considered, as is the case of the integrated circuits TAS5624A and TAS5630B commercialized by Texas Instruments (www.ti.com). Fig. 5. Structure of the class D amplifier. As in the case of class AB amplifier, the output magnitude of the class D PA is the secondary current (IS), and its average value can be calculated using Eq. (6). VO VCC 2d 1 (6) R R VO is the output voltage of the class D PA and d is the duty cycle of the half-bridge inverter operation. It is defined as the ratio of the M1 transistor conduction time (TON) to the duration of the switching period (TS). If the conduction time of the transistor M2 is defined as TOFF, then the switching period can be expressed as: TS = TON + TOFF, and the duty cycle of the class D amplifier as: d = TON / TS. The great advantage of this type of amplifiers is their high efficiency, which is theoretically equal to 100 % if the losses in semiconductors M1 and M2 are neglected. In this regard, input and output powers of class D amplifiers have practically the same value and can be calculated through Eq. (7). IS PI PO R I S 2 (7) The small energy consumption of fluxgate current sensors based on class D amplifiers is highlighted in [13, 22]. But, given that the magnitude of interest is the PA output current (IS), it is necessary to considerer that there is a ripple in this current due to the switching process, whose value can be estimated using Eq. (8). I S 2 VCC TS d 1 d LS (8) The current ripple ΔIS, which does not exist in a class AB PA, is added to the value of the secondary current and introduces an error in the measurement of this current (IS + ΔIS). The maximum value of this error, which can be inadmissible in high accuracy applications, occurs when d = 0.5 and can be valued using Eq. (9). I S max VCC TS 2 LS (9) In a commercial fluxgate transducer the value of LS may be 3 of some Henry, and it is determined by the current measuring range, the characteristics of the magnetic core and the number of turns of the compensation winding (NS). Therefore, this value is generally not eligible and it is dependent on the electrical and physical characteristics of the current transducer. The value of the switching period (TS) or switching frequency (fS = 1 / TS) can be set, within certain limits, to control the value of secondary current ripple (ΔIS), but there are two different factors to consider: - The switching losses in the two transistors (M1 and M2) increase with frequency. - The switching noise transmission to the secondary current (IS) due to stray capacitance between windings also increases with frequency [23]. As a result, the use of class D amplifiers achieves a considerable energy saving with respect to a linear class AB amplifiers solution. This saving can be up to 50 % at medium current values (IP = IN / 2). But, due to the existence of a ripple in the secondary current, the value of the measurement accuracy is reduced by one or two orders of magnitude in relation to that obtained with the use of class AB amplifiers. In this regard, it is possible to find references to class D amplifiers with low distortion [24] and distortion-free baseband [25], but these designs involve the use of DSP (Digital Signal Processor) due to the high computational complexity required of their control. Therefore, in applications where high accuracy in current measurement is required, the utilization of class D amplifier as output PA of fluxgate current transducers is not applicable in practice. C. Class H amplifier The adjustable power supply amplifiers are based on a class AB linear amplifier (shown in Fig. 4) but with a not constant power supply voltage (variable values of ±VCC). The value of the power supply voltage is adapted in accordance to the value of the amplifier output voltage (VO). These amplifiers are widely used in different fields of industrial electronics (audio frequency amplifiers [26] and arc welding systems [27]) being in the area of radio frequency (RF) communications [28-30] and mobile communications [31, 32] where the largest number of applications are documented. The value of the power supply voltage can be changed in two different modes: stepped mode (multi-stage) and continuous mode (envelope-tracking). But, in any case, the power supply voltage must be always higher than the output voltage of the class AB amplifier. In the first case (multi-stage), and depending on the value of the amplifier output voltage, the amplifier power supply voltage switches between two or three different levels, but always higher than the instantaneous value of the amplifier output voltage. These amplifiers are referred as class G amplifiers. Thus, the difference between the supply and output voltage is reduced in steps, in the same way as power losses in the amplifier, which are proportional to this voltage difference and to the secondary current (IS). In the second case (envelope-tracking), the power supply voltage of the amplifier follows the output voltage waveform, > REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < increased by a value (ΔU), as low as possible without causing distortion in the output signal. These amplifiers are referred as class H amplifiers or envelope tracking (ET) amplifiers. The power losses in a class H amplifier are, as in the class AB amplifier case, directly proportional to the output current (IS) and to the difference between the power supply voltage and the amplifier output voltage. But with the advantage that the voltage difference value is considerably lower in the case of class H amplifier than in a conventional class AB amplifier. The input power of the class H amplifier can be evaluated by Eq. (10). PI VO U I S R I S U I S 4 consumption, when the same operating conditions of the current transducers are considered. (10) And its efficiency can be calculated using Eq. (11). PO VO I S 1 (11) PI VO U I S 1 U VO In this way, the lower the value ΔU the better the efficiency and, in the ideal case with ΔU = 0, the efficiency is the unity. Therefore, the class H amplifier supplying the fluxgate secondary current has an accuracy similar to the class AB amplifier, but with similar energy consumption than a class D amplifier. Consequently, the use of a class H amplifier to supply the secondary current in a fluxgate current transducer is proposed in this work. In this way, the following advantages could be obtained: - High precision, as in the conventional fluxgate current transducers based on a linear class AB amplifier. - Large decreasing of electrical power requirements for the transducer operation. This power can be reduced up to 50% when is compared with a conventional fluxgate transducer. A drawback, related to the use of a class H amplifier, is the transient response against fast changes in the primary current (IP). This is due to the lower bandwidth of class H amplifier in relation to the conventional linear class AB amplifiers. However, this is not an important disadvantage in this application since, in a current transducer based on the fluxgate effect, the transient feature is assumed by the conventional current transformer effect and not by the fluxgate effect [13]. D. Energy shaving comparison The input power of the three compared PA is evaluated via Eqs. (2), (7) and (10) respectively. The parameters involved in these expressions are the amplifier supply voltage (VCC), the resistance of the measurement circuit (R), the secondary current (IS) and the incremental voltage (ΔU) used in the class H amplifier. The values of input power of the three compared PA versus the primary current (IP) are shown in Fig. 6, assuming the utilization of transducers with the same maximum current IN = 1000 A, power supply voltage ±VCC = ±15 V and input power PI(0) = 3 W in not input current conditions (IP = IS = 0). Also is assumed ΔU = 1 V and NS = 1000. As discussed above, and as Fig 6. shown, class D and class H amplifiers show similar features in terms of input power, while the use of class AB amplifiers lead to a higher power Fig. 6. Theoretical input power of compared amplifiers IV. IMPLEMENTATION OF CLASS H AMPLIFIER A. Structure of power supply The power supply voltage of a class H amplifier is related to its own output voltage, and it is determined using the value of the incremental voltage ΔU. The positive (+VAB) and negative (-VAB) values of the power supply voltage are can be calculated accordingly to Eq. (12). VAB VO U (12) Naturally, the power supply system that provides the power supply voltages to the class AB amplifier should be high performance system, as is the case of switched power sources. The utilized switched source must provide positive and negative voltages and currents, so it must be able to work in all four quadrants V/I. The usual power converter structure that can meet these demands is a full-bridge converter. But, other options based on multiple levels buck converter are documented [33, 34]. In this development, two half-bridge converters are used, because it is necessary that the output signal is referred to a common point with the electronics devoted to power control. The structure of the converter used is shown in Fig. 7, where one half-bridge generates the positive voltage (+VAB) and the second converter generates the negative voltage (-VAB). Fig. 7. Structure of class H power amplifier > REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < B. Control of power supply The control system of the class AB amplifier power supply must control the drivers of the two half-bridge converters. The voltage reference used as set-point of these two half bridges converters is equal to the output voltage of the class AB amplifier (VO) plus the constant value of the used incremental voltage (ΔU), as shown Eq. (12). A closed loop control is applied to only one of these two converters (the half-bridge devoted to the positive supply voltage generation). The duty cycle of this converter (d+) is obtained by using as reference voltage (V+*) the positive value of Eq. (12), as shown Eq. (13). V * VO U (13) The duty cycle of the second half-bridge converter (d-) will be the complementary, as stated in Eq. (14). TON TOFF d 1 d (14) As a consequence, the output voltage of both converters has the same value but opposite in polarity. Fig. 8 shows a simplified block diagram of the principle applied to the control of the converter devoted to the positive supply voltage generation. 5 +VAB VO ̵ VAB Fig. 9. Power supply voltages ±VAB (magenta and green traces) and output voltage VO (blue trace) in class H amplifier. V. EXPERIMENTAL RESULTS Fig. 10 presents the final industrial prototype of the designed transducer, where the main components of the system can be identified: measuring transformer, printed circuit board (PCB), and box to contain all transducer elements. Fig. 10. Designed transducer based on class H output PA. Fig. 8. Block diagram of inverter control class H C. Behavior of class H amplifier The design of the proposed PA (shown in Figs. 7 and 8) was simulated under different conditions in order to verify its proper operation before using this amplifier as output stage of a fluxgate current sensor. Once verified by simulations the functionality of the designed PA, a laboratory prototype was built and tested. Fig. 9 shows the waveforms measured for the output voltage and power supply voltages of the class H amplifier when 5 V and 200 Hz sinusoidal voltage is applied as amplifier input voltage. The voltage tracking difference (ΔU) is set to 1V, 1 Ω resistor is used as RB and ±VCC are set to ±15 V. In order to evaluate the performance of this device, some laboratory tests were conducted. This section presents the results obtained in these experimental tests. The first test evaluates the linearity and measurement errors in full-scale DC current. The second one evaluates the device response for highlevel current transients. Finally, the power consumption was compared with the consumption of fluxgate transducers based on class AB and class D PA as output stage, in order to validate the energy saving achieved. As is well known, the frequency response (or bandwidth) of fluxgate current transducers is based on the combination of two (or even three) different techniques [16]: the fluxgate principle to operate from DC to low frequencies (few Hz), and the current transformer principle to operate up to few hundred of kHz. As a consequence, the upper cutoff frequency of the transducer is related to the materials and construction parameters of to the used magnetic components (as are the leakage inductance and the stray capacitance of the measuring transformer [35]).The upper cutoff frequency also depends on the value of output resistor RB but it has no dependence with the electronics devoted to the fluxgate operation. Because the fluxgate transducer presented in this paper uses the same magnetic components presented in [13], it can be assumed that it also has the same bandwidth (about 170 kHz). > REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < A. DC current measurement For this test, the primary current (IP) was generated by a QPX1200 PowerFlex DC power supply and it was adjusted by a programmable DC active load. IP was measured by a no inductive shunt (80 A/150 mV), and this value was used as a reference. Furthermore, this current was coupled to the fluxgate transducer through 14 turns. Finally, the output resistor value utilized for this test was RB = 1 Ω. Fig. 11 shows the input-to-output current characteristic measured on the device under test, and Eq. (15) shows the expression of the linear regression of this characteristic. I Measured 1.0008 I P 0.0031 (15) This result indicates good system linearity (1.0008) and low offset error (around 3 mA). Fig. 12 shows the relative error in the measurement referred to the device full-scale. Accordingly with the obtained data, the relative error is lower than ±0.2 % in the ±700 A range. VO _ H I P 1000 6 (16) In the class AB device NS = 1750 and it uses RB = 2.5 Ω (as is suggested by manufacturer). This implies that the output voltage of this transducer is: VO _ AB I P 700 (17) Finally, both devices use the same power supply voltages (±15 V). In the first test, a slow slope current ramps (±12 A/ms), with initial and final values of IP equal to 0 and ±600 A respectively, are generated. The obtained results for ramps with positive and negative slopes are shown in Fig. 13. Red and magenta traces are the evolution of ±VAB, green trace is the output voltage of the class AB transducer (VO_AB) and blue trace is the output voltage of the device based in class H PA (VO_H). +VAB VO_AB ̵ VAB VO_H +VAB Fig. 11. Experimental input-to-output characteristic of the transducer. VO_AB VO_H ̵ VAB Fig. 13. Response of transducers to current ramps with positive (top) and negative (bottom) slopes. Fig. 12. Measured relative output error of the designed transducer. B. Transient Current Measurement For these experimental tests, a Danfysik Ultrastab 867-700i fluxgate transducer has been used as reference instrument for current measurements. This transducer uses a class AB power amplifier as output stage. The experimental setup used for tests is completed by a DC power supply as generator of the DC primary current (IP) and a 6060B Hewlett Packard DC electronic load. The same IP current flows through the two fluxgate transducers and their output voltages are measured and recorded by a DPO3014 Tektronix digital oscilloscope. In the class H device NS = 1000 and it uses RB = 1 Ω. This implies that the transducer output voltage is: In the second test, a pulsating current for testing the dynamic response of the class H device is generated. The frequency of the current pulse train is 100 Hz with a 50 % duty cycle, the slope of the current pulses is ±1200 A/ms and the initial and final values of IP are 0 and ±600 A respectively. The obtained results for ramps with positive and negative slopes are shown in Fig. 14. Red and magenta traces are the evolution of ±VAB, green trace is VO_AB and blue trace is VO_H. > REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < +VAB VO_AB ̵ VAB VO_H VO_H +VAB VO_AB ̵ VAB Fig. 14. Response of transducers to current ramps with positive (top) and negative (bottom) slopes. Finally, a triangular IP with 150 A of amplitude, 250 A of DC offset and 1 kHz of frequency was generated. Fig. 15 shows the transducers response and the evolution of the power supply in the class H device (same colors are used for traces). 7 In order to appreciate the differences with a linear class AB amplifier transducer, the same class H amplifier was tested without the implementation of the output voltage tracking function, behaving as a simple linear class AB amplifier. This strategy enables testing three different transducers. They uses the same magnetic circuit and the same electronics for implementing the fluxgate measure principle, but using three variants for the output amplifier implementation: class AB, D and H amplifiers. The following figure shows the experimental results obtained in terms of the transducer energy consumption. Various measurements of the current transducers energy consumption have been made on the power supply circuit when ±VCC = ±15 V. Measurements include the self-consumption of the control electronics and the obtained results are depicted in Fig. 16. As it shown, the energy consumption characteristic of the class AB amplifier is proportional to the measured current (IS), with a non-zero power consumption when no current is measured due to the self-consumption of the control electronics. In the case of class D amplifier, as is expected, the evolution of the energy consumption characteristic is proportional to the resistance of the compensation circuit (R) and to the square of the RMS value of the secondary current (IS2). In conditions of zero primary current, non-zero energy consumption exists due to the self-consumption of the control electronics and, in this case, it is slightly higher than in the case of class AB amplifier. +VAB ̵ VAB VO_AB VO_H Fig. 15. Response of transducers to a triangular IP waveform. These results show the concordance between measurements made using the designed transducer and the linear transducer used as a reference instrument (in agreement with Eqs. (16) and (17)). These results also demonstrate the proper operation of the class H PA used in the designed fluxgate transducer. C. Power consumption For these experimental tests, a PREMO fluxgate DCT700A with IN = 1000 A as maximum current (700 A rated current) has been used. This transducer includes a class D amplifier for the generation of the secondary current in the measurement circuit. Using the same magnetic components, but in substitution of class D amplifier, a new control electronics with a class H amplifier has been incorporated. Fig. 16. Power consumption of the class AB, D and H amplifiers. Ordinate axis in watts [W] and abscissa axis in amperes [A]. In relation to energy consumption, the class H amplifier behaves like a class D amplifier, having a similar power consumption characteristic. The energy saving of class D and class H amplifiers, in relation to the class AB amplifier, almost reaches the value of 40 % for some values of measured current (IS) and when the class H amplifiers is used. Differences in energy consumption, when they are compared with the expected values, and the asymmetry observed between positive and negative currents of class H and D amplifiers (shown in Fig. 16) can be attributed to the peculiarities of the control implementation (only the positive supply voltage is regulated) and to the switching and magnetic losses, since these effects have been neglected in the analytical description of current transducers operation. > REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < VI. CONCLUSIONS This work presents a new approach in order to reach a significant improvement, in terms of energy consumption, in current transducers based on the fluxgate effect. Since most of the energy consumption is produced in the output power amplifier that supplies the flux compensation winding, interest has focused on this output stage which is present in all fluxgate transducers. Traditionally, the linear class AB amplifier has been considered as the optimal solution for its accuracy, linearity, low-noise and low-distortion in current measurement. As alternative, and in order to optimize the power consumption of the control electronics, the use of class D and class H amplifiers have been analyzed. Theoretical considerations and experimental results have shown that both types of amplifiers significantly improve the energy consumption by almost 40 % in some ranges of the current transducer operation. However, the class D amplifier has the disadvantage that it adds a current ripple to the output current signal, which can be an inconvenience in cases of high precision applications. The use of class H amplifiers as output stage of compensation and current measurement circuit, is the most advantageous option but with a slightly more complex design. Experimental results also have shown the good performances of the designed transduces when it is compared with a transducer based on a class AB amplifier as output stage. In summary, the use of class H amplifier has the advantages of the low power consumption associated with a class D amplifier, while maintaining all other benefits in terms of linearity, accuracy, bandwidth, low-noise, etc. which are typical of a linear class AB amplifier. REFERENCES [1] [2] [3] [4] [5] [6] [7] [8] [9] Popović-Gerber, J.; Oliver, J.A.; Cordero, N.; Harder, T.; Cobos, J.A.; Hayes, M.; O'Mathuna, S.C.; Prem, E., "Power Electronics Enabling Efficient Energy Usage: Energy Savings Potential and Technological Challenges", IEEE Transactions on Power Electronics, vol.27, no.5, pp.2338-2353, May 2012 Consoli, A.; Scelba, G.; Scarcella, G.; Cacciato, M., "An Effective Energy-Saving Scalar Control for Industrial IPMSM Drives", IEEE Transactions on Industrial Electronics, vol.60, no.9, pp.3658-3669, September 2013 Lu Lu; Bin Yao, "Energy-Saving Adaptive Robust Control of a Hydraulic Manipulator Using Five Cartridge Valves With an Accumulator", IEEE Transactions on Industrial Electronics, vol.61, no.12, pp.7046-7054, December 2014 Frink, M., “Green Amps”, Front of House (Magazine), vol.10, no.5, pp.34-35, February 2012 Guanghai Gong, H. Ertl, J.W. Kolar, "Novel Tracking Power Supply for Linear Power Amplifiers", IEEE Transactions on Industrial Electronics, vol. 55, no. 2, pp. 684-698, February 2008 Chang, Joseph S.; Meng-Tong Tan; Zhihong Cheng; Yit-Chow Tong, "Analysis and design of power efficient class D amplifier output stages", IEEE Transactions on Circuits and Systems I: Fundamental Theory and Applications, vol.47, no.6, pp.897-902, June 2000 Ziegler, S.; Woodward, R.C.; Iu, H.H.-C.; Borle, L.J., "Current Sensing Techniques: A Review", IEEE Sensors Journal, vol.9, no.4, pp.354-376, April 2009 Ripka, P., “Electric Current Sensors: A Review”, Measurement Science and Technology, vol.21, no.11, pp.1-23, November 2010 Velasco-Quesada, G. et al. “Technologies for Electric Current Sensors”, in Novel Advances in Microsystems Technologies and Their Applications, L.A. Francis and K. Iniewski Eds. FL: CRC Press, 2014, pp.265-292 8 [10] Ripka, P., “Review of fluxgate sensors”, Sensors and Actuators A: Physical, vol.33, issue 3, pp.129-141, June 1992 [11] Ripka, P.; Janosek, M., "Advances in magnetic sensors," 2008 IEEE Sensors, pp.1-4, 26-29 Oct. 2008 [12] “Isolated Current and Voltage Transducers: Characteristics-ApplicationsCalculations (3rd Edition),” LEM Components, CH 24101 E/US, 2004 [13] Velasco-Quesada, G.; Román-Lumbreras, M.; Conesa-Roca, A.; Jeréz, F., "Design of a Low-Consumption Fluxgate Transducer for High-Current Measurement Applications," IEEE Sensors Journal, vol.11, no.2, pp.280,287, Feb. 2011 [14] Xiaoguang Yang; Yuanyuan Li; Wei Guo; Weidong Zheng; Cunfu Xie; Hongli Yu, "A New Compact Fluxgate Current Sensor for AC and DC Application", IEEE Transactions on Magnetics, vol.50, no.11, pp.1,4, Nov. 2014 [15] Xiaoguang Yang; Yuanyuan Li; Weidong Zheng; Wei Guo; Youhua Wang; Rongge Yan, "Design and Realization of a Novel Compact Fluxgate Current Sensor", IEEE Transactions on Magnetics, vol.51, no.3, pp.1,4, March 2015 [16] Michel Ghilardi; Horst Bezold, “Maintaining Precision in Varying Temperatures”, Bodo’s Power Systems, vol.06-15, pp.16,21, June 2015 [17] Ripka, P.; Draxler, K.; Styblikova, R., "Measurement of DC Currents in the Power Grid by Current Transformer", IEEE Transactions on Magnetics, vol.49, no.1, pp.73-76, Jan. 2013 [18] Ripka, P.; Draxler, K.; Styblikova, R., "AC/DC Current Transformer With Single Winding", IEEE Transactions on Magnetics, vol.50, no.4, pp.1-4, April 2014 [19] Kui Li; Feng Niu; Yi Wu; Yao Wang; Yihua Dai; Lili Wang; Erping Li, "Nonlinear Current Detection Based on Magnetic Modulation Technology”, IEEE Transactions on Magnetics, vol.51, no.11, pp.1-4, Nov. 2015 [20] PREMO Group, Catalog: DCT 700A Transducers (2014). [21] LEM International SA, Catalog: High Precision Current Transducers (2014). [22] Radivoje Đurić; Milan Ponjavić, “Self-Oscillating Fluxgate Current Sensor with Pulse Width Modulated Feedback”, ELECTRONICS, vol.14, no.2, December 2010 [23] W. T. McLyman, “Transformer and Inductor Design Handbook”, 3rd edition. New York: Marcel Dekker, 2004, p. 532 [24] Stefanazzi, L.; Chierchie, F.; Paolini, E.E.; Oliva, A.R., "Low Distortion Switching Amplifier With Discrete-Time Click Modulation", IEEE Transactions on Industrial Electronics, vol.61, no.7, pp.3511-3518, July 2014 [25] Chierchie, F.; Paolini, E.E.; Stefanazzi, L.; Oliva, A.R., "Simple RealTime Digital PWM Implementation for Class-D Amplifiers With Distortion-Free Baseband", IEEE Transactions on Industrial Electronics, vol.61, no.10, pp.5472-5479, October 2014 [26] Jingqi Liu; Allasasmeh, Y.; Gregori, S.; Leesti, B.; Snelgrove, M., "An envelope tracking H-bridged audio amplifier with improved efficiency and thd less than 0.1%", 25th IEEE Canadian Conference on Electrical & Computer Engineering (CCECE), pp. 1-4, April 29-May 2 2012 [27] Anderson, K.; Cook, G.E.; Barnett, R.J.; Eassa, E.H., "A class-H amplifier as a generic power source for arc welding", IEEE Southeastcon '90 Proceedings, vol.3, pp.792-796, 1-4 April 1990 [28] Tafuri, F.F.; Sira, D.; Jensen, O.K.; Larsen, T., "Efficiency enhancement of an envelope tracking power amplifier combining supply shaping and dynamic biasing", European Microwave Integrated Circuits Conference (EuMIC), pp.520-523, 6-8 October 2013 [29] Junghyun Ham, Haeryun Jung, Hyungchul Kim, Wonseob Lim, Deukhyoun Heo; Youngoo Yang. “A CMOS Envelope Tracking Power Amplifier for LTE Mobile Applications”, Journal of Semiconductor Technology and Science, vol.14, no.2, April 2014 [30] Huan Xi; Qian Jin; Xinbo Ruan, "Feed-Forward Scheme Considering Bandwidth Limitation of Operational Amplifiers for Envelope Tracking Power Supply Using Series-Connected Composite Configuration", IEEE Transactions on Industrial Electronics, vol.60, no.9, pp.3915-3926, September 2013 [31] Jonmei, J.Y.; Kimball, D.; Chin Hsia; Hsuan-yu Pan; Ricketts, S.; Ghajari, H., "High-efficiency, high-linearity envelope tracking power amplifier for mobile UHF applications," Military Communications Conference - MILCOM 2010, pp.1572,1576, Oct. 31 2010-Nov. 3 2010 [32] Young, J.P.; Ripley, D.; Lehtola, P., "Envelope Tracking power amplifier optimization for mobile applications", IEEE SOI-3D-Subthreshold Microelectronics Technology Unified Conference (S3S), pp.1-4, 7-10 October 2013 > REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < [33] Rodriguez, M.; Miaja, P. F.; Rodriguez, A.; Sebastian, J.; ”A MultipleInput Digitally-Controlled Buck Converter for Envelope Tracking Applicationsin Radiofrequency Power Amplifiers”, IEEE Transactions on Power Electronics, vol.25, no.2, pp.1078-1089, February 2010 [34] V. Yousefzadeh, E. Alarcon, D. Maksimovic, “Three-level buck converter for envelope tracking applications,” IEEE Transactions on Power Electronics, vol.21, no.2, pp. 549-552, March 2006 [35] N. Kondrath; M.K. Kazimierczuk, "Bandwidth of Current Transformers", IEEE Transactions on Instrumentation and Measurement, vol.58, no.6, pp.2008,2016, June 2009 9