Está en la página 1de 4

Paper accepted for presentation at 2003 IEEE Bologna Power Tech Conference, June 23th-26th, Bologna, Italy

Output voltage ripple reduction of a high power factor mode operated isolated charge-pump AC/DC converter
M. Sippola
major drawback of the need for a relatively large output capacitor to reduce low frequency output voltage ripple [6]. Further, in [7] an AC/DC converter with the input current shaper inductance integrated into the isolation transformer (leakage inductance) and operated with secondary side charge pump capacitor was presented. Like resonant converter this has inherently high power factor but also a considerable low frequency output voltage ripple. In [7] only a simple, constant switching frequency control was considered in order to minimize the complexity of the circuitry. In this paper a method to reduce output voltage ripple is presented. The method is based on the switching frequency variation by instantaneous mains voltage feed forward. Among the output voltage ripple reduction the possible effects on line current harmonics and conducted electromagnetic interference (EMI) are considered. II. ISOLATED CHARGE PUMP AC/DC CONVERTER The proposed AC/DC converter topology is sketched into Figure 1. Mains voltage Vac is rectified by a diode bridge and chopped across the transformer primary by a half bridge M1, M2. During the first half of switching period (M1 conducting) the charge pump capacitor Cp is charged towards instantaneous value of rectified line voltage ( Vac) divided by transformer turns ratio (N2/N1). However, because the increase of charging current magnitude is limited by transformer leakage inductance and because of the large capacitance value of the charge pump capacitor Cp the voltage across it stays approximately constant during the charging period. During the other half of switching period (M2 conducting) some charge from pump capacitor Cp is discharged to output capacitor Co through the output inductor Lo.

Abstract In this paper output voltage ripple reduction technique for a new AC/DC-converter topology based on a halfbridge driven transformer isolated charge pump is presented and studied experimentally. The main principle is that the double line frequency output voltage ripple related to high power factor mode operated resonant AC/DC converters is reduced by the switching frequency variation controlled by instantaneous line voltage feed forward. Among output voltage ripple the line current waveform and conducted electromagnetic interference (EMI) are considered as well. The measurements with a prototype proved output voltage ripple reduction from 1.4 (rms) to 0.3V (rms) by using the principle suggested in the paper. Other interesting observations were that this also shaped the input current waveform from EN61000-3-2 Class D to A and also reduced the low frequency differential mode average emissions by approximately 15dB. Index Terms AC/DC converter, Charge pump converter, Harmonic currents, Power factor correction, Rectifier, Switched mode power supplies

I. INTRODUCTION he regulatory limits on line current harmonics (EN610003-2) have motivated work on power factor corrected AC/DC converters. In particular single-stage power factor corrected AC/DC converters have been studied in order to reduce cost and complexity of two stage solutions (boost prerequlator + isolated DC/DC stage). [1],[2],[3],[4]. Often these are based on the power factor correction implemented with an inductor connected between line voltage and further power processing stages such as switches and transformers and on the use of line frequency ripple energy storage capacitor. Usually these circuits use the input inductor in discontinuous mode (DCM) to achieve power factor correction without the need to measure instantaneous line voltage and inductor current. However it has been noted in practice that the topologies which try to do everything in a single stage may become quite complex and inefficient and their advantages over two-stage approach may become marginal when cost per watt is considered [5]. An alternative, cost per watt driven solution might be based on resonant AC/DC converters with inherently sinusoidal input current characteristics but with the
Mika Sippola is with Schaffner-electroFerrum, PO BOX 10, FIN-08151 Lohja, Finland (email: msippola@schaffner.com).

Fig. 1. Converter schematic

0-7803-7967-5/03/$17.00 2003 IEEE

The operation was analysed in more detail in [7]. Here it is noted that the derivative of input current iVin versus input voltage Vin can be shown to be:

course, by doing this the input current shape is also affected as will be seen from the measurements.

IVin Vin
where

( Ls + Lo ) + Ls 8 f s ( Ls + Lo ) Ls
Ls is isolation transformer leakage inductance Lo is output inductance Fs is switching frequency

(1)

From (1) it can be concluded: 1) When switching frequency fs is constant during the line period Eq. (1) is also constant and the input current shape follows the shape of input voltage, 2) Input current magnitude can be controlled by switching frequency fs. This can be utilised for controlling the output voltage with the switching frequency fs. It should be noted that the increasing switching frequency reduces output voltage and vice versa. On the other hand it should also be noted that the charge pump rectifier can deliver energy to output capacitor only during a limited period during line cycle i.e. the line current waveform is cross-over distorted (the conduction angle a < 360) (Figure 2). The limited conduction angle not only causes cross over distortion to input current but also a mains frequency ripple to output voltage as and relatively large Co is needed to keep this at acceptable level.

Fig. 3. Controll block diagram IV. PROTOTYPE CONVERTER A Prototype converter with component values in Table 1 was build into 2-sided FR4 PCB boards (Figure 4). The nominal switching frequency of the prototype converter was set to 240kHz and the maximum switching frequency during peak mains voltage to 400kHz. DC output voltage was adjusted to 10V. The switching frequency was generated using GW INSTEK GFG-8215A Function Generator with VCF (frequency control voltage) input connected to rectified mains voltage through voltage divider network.
TABLE 1. PROTOTYPE COMPONENT VALUES Output capacitor Cout Pump capacitor Cpump Output inductor Lout Half-bridge capacitors Input filter capacitor Transformer core Primary winding Secondary winding 10000uF 4.4uF 10uH 470nF 100nF 2 x (Planar E38 + I-plate 3F3) 11 turns 2 turns

Fig. 2. Line voltage and line current III. OUTPUT VOLTAGE RIPPLE REDUCTION Output voltage ripple reduction based on the switching frequency modulation was chosen instead of for example pulse width modulation (PWM) since this also has an interesting potential to affect EMI levels by stirring the emission into a wider frequency range with lower peak values. After some initial trials a very simple concept based on input voltage feed forward was selected (Figure 3). The average converter output voltage level is controlled by a PI-regulator which is controlling the inverter switching frequency with a voltage controlled oscillator (VCO). Instantaneous mains voltage is then summed into PI-output voltage to provide mains voltage feedforward. As a result the converter current throughtput tends to reduce at the peak mains voltage. Similarly, the converter current throughput tends to increase at the lower mains voltage. As a result the converter throughput power tends to be constant as will be output voltage as well thus the line frequency output voltage ripple is reduced. Of

Fig. 4. Prototype converter V. MEASUREMENTS AC/DC converter prototype was fed from mains through a variac in order to provide soft-start and to adjust mains voltage level. The converter was loaded with 20W resistive load. Output voltage and input current waveforms and conducted electromagnetic interference (EMI) were measured using LeCroy LT264 Digital Storage Oscilloscope (300MHz : 1Gs/s) and R&S ESHS 30 EMI test receiver (150kHz 30MHz) with ESH 2-Z5 Artifical mains network,

respectively. Measurements were conducted both with and without output voltage ripple reduction, also with and without switching frequency variation. A. Low frequency output voltage ripple The low frequency output voltage ripple measurement results are shown in Figures 5 and 6 without and with switching frequency variation, respectively. It can be noted that in latter case the rms value of this double line frequency voltage perturbation has been reduced as discussed in Chapter III. The reduction from 1.4V to 0.3V corresponds to reduction to 21% and after reduction the ripple is only 3% of DC voltage level (10V). It is also noted that in both cases there was a significant switching frequency voltage ripple at the converter output that should be further filtered by for example by using a lower effective series resistance (ESR) filter capacitor.

(Fig. 7.) the converter exhibits more or less linear voltage / current characteristics according to Eq.1. and Fig.2. Althought the total harmonic distortion (THD) with variable switching frequency control was determined to be as high as 56% when compared to 26% without frequency variation it should be noted that the current waveform with variable switching frequency control (Fig. 8.) does not fall within EN61000-3-2 Class D envelope thus Class A with often easier to comply (absolute) harmonic limits apply.

Fig. 7. Input current without switching frequency variation

Fig. 5. Output voltage ripple without switching frequency variation (2V/div : 2ms/div)

Fig. 8. Input current with switching frequency variation C. Conducted EMI Fig. 6. Output voltage low frequency ripple with switching frequency variation (2V/div : 2ms /div) B. Input current waveform The input current measurement results are shown in Figures 7 and 8 without and with switching frequency variation, respectively. As required for output voltage ripple reduction the switching frequency variation has increased the input current at low input voltage levels and reduced it at the mains voltage peak values. This corresponds to constant input power as required by constant output power and output voltage as well. On the other hand without switching frequency variation The conducted EMI measurement results of the test converter are shown in Figures 9 and 10 without and with switching frequency variation, respectively. Both peak (higher curve) and average (lower curve) measurements are shown in the plots. 20ms measurement time and 10kHz IF bandwidth were used. General envelope of the emission spectrum has remained the same but some interesting observations can be made. 1) Without switching frequency variation (Fig. 9.) the discrete harmonics of switching frequency are clearly visible 2) At higher frequencies were interference is typically of common mode type the switching frequency variation results approx. 10dB higher peak emission values. This is obviously due to fact that the increase of switching frequency also increases the energy associated with capacitively coupled

displacement (i.e. common mode) currents as well. 3) At the low frequency range where emissions are typically of differential mode type the switching frequency variation has reduced the average switching frequency harmonic components about 15dB. Because it is often just the average low frequency differential emissions, such as the switching frequency and the first harmonics that require most bulky filtering this can be considered as a particularly interesting result.

low frequency differential mode average emissions were reduced by approximately 15dB. The proposed half-bridge driven isolated charge pump rectifier seems feasible for low cost per watt / high power density AC/DC converters. Potential applications for of the converter are for example: low cost on board PFC AC/DC converters of AC-distributed power systems, 12V halogen lamp power supplies, battery chargers and portable AC/DC converters with power levels requiring EN61000-3-2 compliant performance. VII. REFERENCES
[1] Chongming Q., Smedley, K.M., A topology survey of single-stage power factor corrector with a boost type input-current-shaper, Proceedings of 15th IEEE Applied Power Electronics Conference and Exposition 2000 , vol.1, p. 460 467, 2000 Kheraluwala, M.H.; Steigerwald, R.L.; Gurumoorthy, R. A fastresponse high power factor converter with a single power stage, Proceedings of 22nd IEEE Power Electronics Specialists Conference, p. 769 779, 1991 Madigan, M.; Erickson, R.; Ismail, E. Integrated high quality rectifierregulator, Proceedings of 23rd IEEE Power Electronics Specialists Conference, vol. 2, p.1043-51, 1992 Redl, R.; Balogh, L.; Sokal, N.O. A New family of single-stage isolated power-factor correctors with fast regulation of the output voltage, Proceedings of 25th IEEE Power Electronics Specialists Conference, vol.2, p. 1137 1144 Sharifipour B., Huang J.S., Liao P., Huber L., Jovanovic M.M., Manufacturing and cost analysis of power-factor-correction circuits, Proceedings of 13th IEEE Applied Power Electronics Conference, , vol.1, p.490-494 , 1998 Schutten M.J., Steigerwald R.L., Kheraluwala M.H., Characteristics of Load Resonant Converters Operated in a High Power Factor Mode, Proceedings of 6th IEEE Applied Power Electronics Conference, pp. 5 16, 1991 Sippola M., Sepponen R., A new AC/DC converter topology with high power factor, International Journal of Electronics, vol. 89, no. 3, pp. 245 258, 2002

[2]

[3] [4]

[5]

Fig. 9. Conducted EMI without switching frequency variation


[6]

[7]

VIII. BIOGRAPHIES

Fig. 10. Conducted EMI with switching frequency variation VI. SUMMARY AND CONCLUSION The output voltage ripple reduction based on the mains voltage feedforward driven switching frequency variation was demonstrated and validated successfully. Rms output voltage ripple was reduced from 1.4V to 0.3V, input current waveform was shaped from EN61000-3-2 Class D to A and

Mika Sippola was born in Helsingin mlk, Finland in 1971. He received M.Sc, Lic.Sc and Dr.Sc in electrical engineering from Helsinki University of Technology, Espoo, Finland, in 1997, 2001 and 2003, respectively. He has been working as a design engineer for an UPS manufacturer, researcher at Applied Electronics laboratory (Helsinki University of Technology) and as a consultant for inductive components manufacturers. Since 2002 he has been with Schaffner were he works on Research & Technology. His research interests include power conversion, electronics packaging, electromagnetic compatibility and power quality.

También podría gustarte