Improved PFC Boost Choke using a Quasi-Planar Winding ...

1 Abstract- A novel approach to Boost inductordesign using a Quasi-Planar Winding configurationconsisting of a helical wound flat copper coil todevelop high ampere-turn capability for use in highcurrent power factor correction Boost circuits ispresented. By comparing this approach with fourother commonly used Boost Choke configurations ,the advantage of this approach is INTRODUCTION As the requirement for power factor correction(PFC) has become more widespread do to such interna-tional regulations as IEC1000-3-2 which limit inputline current harmonics, the PFC Boost converter hasbecome a necessary part of most power suppliesintended for international use. Active PFC Boost converters can be divided into twotypes : the continuous conduction mode fixed frequencytype and the discontinuous conduction mode variablefrequency (constant on-time ) type. The continuousconduction mode type (CCM) requires a very fastoutput rectifier with good forward and reverse recoverytimes since the diode must switch from zero to fullcurrent in tenths of CCM PFC and the discontinuous conductionmode( DCM )PFC both require a Boost Choke that canwithstand high voltage (typically 400V) square waveswhich necessitates the selection of a core material withlow core loss at high flux levels and high frequency(typically from 20kHZ to 500kHz).

2 Improved PFC Boost Choke using aQuasi- planar Winding ConfigurationDave ShontsSchott Corporation1000 Parkers Lake RoadWayzata, MN 55391 There is a big difference in the current ripplerequirement between the two approaches : the DCMPFC Boost Choke has much higher ripple currentlevels at the same output power level. This isbecause it must switch from zero to line currentlevels each PWM cycle unlike the CCM PFC boostchoke which does not have its current drop to zeroon each DCM PFC requires a very high quality boostinductor since it must switch extremely high peakripple currents and voltages at a wide variation offrequencies. Core manufacturers have developedsome useful materials over the last decade whichhave proved useful in the construction of highfrequency Boost inductors. Some examples of thesematerials for cores are Mag Inc. s Kool Muavailable in toroidal cores and high frequencygapped ferrite available in ETD cores.

3 But a goodwinding configuration has not been available :round wire does not meet the needs of highfrequency ; foil wound inductors are better as far asskin effect is concerned but trap heat in the innerwindings , litz wire is hard to terminate and alsotraps heat in the inner windings especially aroundthe location of the center leg gap. This paper introduces a new approach for boostinductor windings : the Helical Wound Technology(HWT) which is a helical wound flat copper coilwith a baked on thermal epoxy insulation. This newconfiguration allows the construction of a boostchoke with flat copper helical windings which areperpendicular to the center leg of the ferrite ETDcore. Each Winding has an outer surface whichallows internally generated heat to escape to thesurface. The Winding does not require a bobbinsince it is a self supporting structure which allows alarger portion of the window area to be used forcopper .Since the foil Winding is typically 20 milsthick or more it is a self leaded device which lendsitself to easy pc mount on through hole boards (seeFigure 1).

4 A detailed analysis of this new compo-nent is made using a CCM PFC test circuit whichallows the variation of frequency from 20kHz to200kHz and load from 100 Watts to 10kWatts (seeFig. 2). This circuit along with the PFC CCMcontrol circuit (see Fig. 3) allows the test andmeasurement of all the criticalFigure 1. A new type of Boost Choke construction:gapped ETD ferrite with the new helical wound technology (HWT)Figure 2. Continuous Conduction Mode PFC Test circuitFig. 3. Continuous Conduction Mode PFC Control Circuitparameters associated with the Boost Choke perfor-mance such as 1) Boost inductor Winding temperatureversus frequency (measured at the internal hot spot ofthe Winding 2) Boost inductor Winding temperature asa function of current and 3) Boost inductor windingtemperature as a function of input to output temperature measurements were made with anisolated thermal sensor placed in the inner Winding ofthe Boost Choke under test.

5 By measuring the tempera-ture at which the inductor reaches 100 degreesCentigrade, a quantifiable limit to the to the inductorscurrent carrying capacity at a given frequency andinput voltage can be found. It is worthy to note that theCCM PFC circuit was chosen for the empirical testvehicle because it allows the precise adjustment of itsfixed frequency where as the DCM PFC circuittypically sweeps out a wide range of frequencies as afunction of line and load. By adjusting the load andfrequency on the CCM PFC it is possible to simulatethe ripple current , DC bias current, and voltage thatthe Boost inductor in the DCM PFC (also called thetransition mode PFC or constant on-time PFC). Thepower loss variations between different Boost chokescan be measured by noting the difference in efficiencyof the converter with each type of Boost Choke insertedinto the test circuit. It is essential that an extremelyrobust test circuit be used in order not to have effi-ciency variations masked by other elements in theconvertor.

6 The test circuit shown in and 2 usedlarge Isotop packaged fets and diodes that wereheatsinked to a large baseplate, which eliminated thevariation in RDS (ON) and diode conduction andswitching losses during the testing of each individualboost Choke . II. COMPARISON WITH 4 OTHER Boost . Choke Winding CONFIGURATIONSIn addition to the empirical measurements made on theboost Choke , it and four other Winding configurationswere analyzed on Proxy, A Proximity and Skin EffectAnalyzer ( ) . This program provides an AC lossanalysis of different Winding configurations. It analyzesplanar, litz, foil and round wire Winding configurationsusing Fourier analysis. This program allows the input ofthe DC bias current and the triangular waveform uses the DC bias current to determine the minimumloss condition, and uses the ripple current content tocalculate the minimum loss for a given Winding thick-ness. It allows the determination of the optimum windingconfiguration to reduce the ratio of AC to DC 5 different Boost Choke configurationsanalyzed:1) toroidal core with Litz wire ( Mag.

7 Mu toroid, 77109-A7 with 420strands of AWG 36 Litz wire)2) copper foil wound on an ETD49-3C85core (Phillips largest ETD core andbobbin conventionally wound with 10 milcopper foil and ) ETD49-3C85 core and bobbin woundwith 420 strands of AWG36 Litz wire4) Toroid core with conventional wire (Mag. Incs. Kool Mu toroid 77109-A7with Awg 12 wire)5) Improved PFC Boost Choke using aQuasi- planar Winding Configurationconsists of a 20 mil thick by .225 incheswide helical wound copper coil with of .925 inches , an of wound coated with 3M thermalepoxy as interwinding insulation. Thereare 44 spiral turns on an the purposes of the analysis each of theinductors were designed for 80uH with 10A of DCbias current. As can be seen from Figure 4 andFigure 5 there are large differences in DCR,copper weight per window area and skin depthsbetween the various Winding configurations. Inorder to achieve a low loss Boost inductor it isessential that 1) DC loss is low 2) AC loss ratio islow and 3) core loss is low.

8 In addition to this theideal Winding configuration would have a meansfor allowing any heat generated to radiate outfreely from the center and not trap heat in theareas around the gap . The inductor temperatureswere all measured at ambient air temperatures(25C) and the inductors were not tied to a baseplate or heatsink The temperatures of the induc-tors were measured at 500 watts of output powerinto a resistive load ,and the input voltage RESULTSAs can be seen from Fig. 6 and 7 the HWTinductor measured the lowest temperatures and thehighest efficiency of the inductors tested. Byadjusting the frequency, load, and input voltagethe test circuit allows the adjustment of the ACripple magnitude as well as the DC bias level toany value desired. It is in this way that theconditions the Boost Choke would see operating ina DCM PFC or Transition Mode PFC can also ofWinding methodsHelical Coil (HWT)16 X .0393 (8757 Circ-Mils)DCR = mOhmCU Wt = mLBMagnet Wire16 Turns#12 AWG (6528 Circ-Mils)DCR = mOhmCU Wt = mLBCopper Foil16 X (6493 Circ-Mils)DCR = mOhmCU Wt = mLBFigure 4.

9 Comparison of Winding methods via a normalized one inch by one inch Winding cross sectionSKIN EFFECT EXAMPLEFREQUENCY = 200kHzTEMPERATURE = 100 CSKIN DEPTH = X X X 5. Skin Effect ComparisonsFigure 6. Normalized power lossFigure 7. Empirical measurements of Boost Choke temperaturesIV. SUMMARYIn this paper we have looked at a new Winding configura-tion for Boost chokes and made comparisons to othertypical methods of the past. This paper has shown that theHWT configuration is capable of producing the lowestDCR per unit volume which relates to the lowest minimalwinding loss at DC bias have also found that its AC loss ratio is very low incomparison to the other Winding structures because of itsquasi- planar Winding structure. These characteristics haveexplained the low temperature readings under comparableconditions of the HWT Quasi-Planar Boost Choke incomparison with the other CONCLUSIONA novel approach to Boost inductor design using aspiral wound flat copper coil shows great promise inmeeting the needs of the power supply designer:1) a printed circuit mountable inductor that is selfleaded2) an inductor utilizing a flat, thin conductorwhich accommodates 10 mil skin depthrequirements, making for low ac losses in theconductor at frequencies from 10kHz ) A spiral or helical wound coil which lendsitself to either planar , low profile inductors, orquasi- planar coils mounted on ETD type coresyielding a very high ampere turn per linearcm.

10 Capability in a very dense ) This new coil structure leads to an inductorwith 50% higher energy storage capabilitythan the other approaches analyzed (in otherwords, higher inductance at a higher currentin a given space).5) This new coil structure leads to an inductorwith significantly lower copper losses thanthose wound using traditional ) As magnetic core manufacturers continue toimprove inductor cores to operate at higherfrequencies it is also important to look at newcoil Winding techniques such as this one forhigh ampere turn would like to thank Tom Cochran ofSchott Corp for the comparison data ofthe different Winding techniques (Figs. 4and 5) and for his help in the design andobtaining of materials for the variousboost chokes. REFERENCE[1] Proxy A Proximity and Skin EffectAnalyzer by KO SYSTEMS , DavidO mear 10437 Laramie Ave,Chatsworth ,Ca 91311