Part Number Hot Search : 
TLSE156P DS24B33 39202B NJU7221U LTC3112 03800 HA124 4S161N
Product Description
Full Text Search
 

To Download AN-22 Datasheet File

  If you can't view the Datasheet, Please click here to try to view without PDF Reader .  
 
 


  Datasheet File OCR Text:
  introduction many topswitch flyback power supply applications require two or more outputs to supply a variety of secondary circuits. typical consumer applications of these multiple output converters include television and related products such as set- top decoders and video cassette recorders (vcrs). industrial applications generally require a number of outputs to supply analog and digital low voltage circuitry. motor control applications often require several separately isolated outputs to supply half-bridge drivers and control circuitry. when compared to single output flyback supplies, multiple output applications demand further design considerations to designing multiple output flyback power supplies with topswitch application note AN-22 figure 1. schematic diagram of 85-265 vac, 25 w power supply using top223. pi-2123-120297 5 v rtn br1 400 v c1 68 f 400 v c4 0.1 f u1 top223 r2 100 ? 1/2 w d2 d3 1n4148 c10 1000 f 35 v t1 d1 byv26c c7* 1.0 nf y1 c11 100 f 35 v u2 nec2501 u3 tl431 r4 10 k ? r5 10 k ? c9 0.1 f r1 100 ? vr1 p6ke200 l1 3.3 h f1 1.0 a j1 c8 0.1 f l2 33 mh l n * two series connected, 2.2 nf, y2-capacitors can replace c7. d4 l3 3.3 h d5 c2 47 f 50 v 30 v c3 470 f 35 v c6 100 f 35 v 12 v c5 47 f d s c control r3 6.2 ? topswitch-ii r6 10 ? march 1998
AN-22 c 5/98 2 table 2. choice of feedback technique depends on requirements for output regulation. table 1. outline power supply specification. main output 10% 5% 5% input voltage: output 1 output 2 output 3 total output power: 85-265 vac 5 vdc 5% 0.40 a to 2.00 a 12 vdc 10% 0.12 a to 1.20 a 30 vdc 10% 0.01 a to 0.02 a 25 w primary (basic or enhanced) opto/zener opto/tl431 power supply specifications voltage current voltage current voltage current optimize the performance. the design of multiple output power supplies always requires some breadboarding to verify transformer designs, feedback techniques and system behavior. this application note provides guidelines to streamline the decision making process and to reduce development effort for an optimized design. an example multiple output power supply design illustrates the procedure. all essential aspects are considered. the design begins with system specifications that define regulation requirements, followed by selection of an appropriate feedback scheme. it then moves to calculation of transformer parameters and application of construction techniques specific to multiple output supplies, aided by reference to application notes an-17 and an-18 for detailed descriptions. a discussion of output cross regulation includes measurements and test results. additional emi considerations are presented with reference to an-15 and an-16. there is also a listing of general tips which may be appropriate to specific designs. appendix a provides some additional reminders for use of the transformer design spreadsheet, while appendix b contains special techniques for use with output voltages of 3.3 v and 5 v. appendix c gives complete construction details of the transformer used in the hardware examples. design procedure the design procedure for multiple output power supplies is a simple extension of the single output case. the circuitry on the primary side of the transformer is the same for either application. additional steps in the design for multiple outputs are needed only to calculate turns ratios and wire sizes for the extra windings. transformer construction has more degrees of freedom than in the single output case. the designer can apply several circuit techniques to adjust output regulation characteristics as needed. other outputs wider than 10% wider than 10% tighter than 10% notes any lightly loaded output may be post-regulated to get 5% or better regulation with 2% zener proportional feedback from two or more outputs optional power supply feedback techniques output regulation feedback technique
c 5/98 AN-22 3 regulation requirements specification of the regulation requirements on all outputs is essential to successful design of the circuit configuration and transformer. requirements differ significantly depending on the application. one output usually requires tighter regulation than the others. usually the 5 v supply for logic circuitry requires regulation of 5% or less, while other outputs have a wider tolerance of typically 10%. many applications now require both 3.3 v and 5 v outputs, with 5% regulation specifications. there are several techniques which can be used to achieve this performance, and they are discussed in more detail in appendix b of this application note. while a 5 v output may have the most stringent regulation specification, a different winding often has a higher output load specification. consideration must therefore be given to the required cross regulation between these outputs, because it will influence the transformer winding technique for an optimum design. table 1 gives an outline specification for a 25 w power supply with three outputs. note that the 5 v output has the highest current and the tightest regulation, but the 12 v output delivers the highest power. the techniques presented here can be extended to any number of outputs. some specific considerations for more outputs are discussed later. the next step of the design is to determine the most appropriate feedback technique. as a quick reference for deciding the optimum feedback technique, table 2 provides broad design rules which can be used, based on the required output tolerances of a specific application. if no tighter than 10% tolerance is required on all outputs, a primary side feedback scheme may be employed. this technique eliminates the need for an optocoupler by using the primary bias winding of the transformer to derive information about the regulated output on the secondary. this type of feedback scheme is detailed in an-16. it is difficult, however, to achieve the output voltage tolerance of 5% with this scheme alone. if outputs requiring 5% are only lightly loaded, primary side feedback may be used with a linear post regulator on these outputs at the expense of some drop in efficiency. from the specification in table 1, however, the 2 a peak load on the 5 v output would lead to excessive dissipation in a linear regulator; therefore, the remainder of this application note will concentrate on feedback that uses an optocoupler. there are two common techniques to generate a secondary reference with optocoupler feedback. the first uses a simple zener diode as a secondary reference. this technique is described in the supporting literature for power integrations rd5 reference design board. the output voltage is determined by the zener voltage, the forward voltage of the optocoupler s led and the series resistor that sets the loop gain. a 2% tolerance zener diode allows 5% tolerance on the regulated output voltage. however, it is often necessary to improve cross regulation by providing feedback from more than one output. the second technique uses a tl431 precision shunt regulator to offer more flexibility in such cases. the tl431 precision shunt regulator integrates an accurate 2.5 v bandgap reference with an amplifier and driver into a single device. it is popular as a secondary referenced error amplifier. the tl431 also introduces the possibility of combining feedback from two or more outputs simultaneously to its reference pin. this can be a useful technique when it is required to employ one output as the primary source of feedback but also introduce a proportion of the feedback from another output. this advanced technique is described in more detail later. this application note, therefore, focuses on the use of the tl431 shunt regulator. figure 1 shows a schematic in a typical application with an optocoupler to provide tight regulation on the 5 v output of a multiple output power supply. transformer design the choice of topswitch and calculation of the primary transformer characteristics is independent of the number of outputs. as such, the power integrations standard transformer design spreadsheets (available from your local power integrations representative or on the power integrations web site at www.powerint.com ) can be used to define the basic transformer specification in terms of the transformer core, primary inductance, primary turns and the output volts per turn. this basic design can then be extended to define the turns and wire selection on other outputs. two spreadsheets are available: one for discontinuous conduction mode (dcm) designs and one for continuous conduction mode (ccm) designs. refer to an-16 and an-17 in the power integrations 1996-97 data book and design guide for further explanation of converter operation and use of the spreadsheets. operation in dcm results in smaller transformer core sizes for a given output power, but the smallest size is often not the most desirable choice in multiple output power supplies. transformer hardware is usually selected to allow optimum circuit board layout. this motivation drives the selection of a transformer bobbin with the best arrangement of the number of pins and the pin spacing. designing for ccm provides the optimum utilization of the topswitch silicon for a given output power. therefore, this
AN-22 c 5/98 4 figure 2. spreadsheet to design transformers for single output and multiple output flyback converters. AN-22.xls 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 abcde f rev 2.1 input output contr2p1.xls: topswitch continuous fl y back transformer desi g n s p readsheet enter application variables AN-22 vacmin 8 5 volts minimum ac input voltage vacmax 265 volts maximum ac input voltage fl 5 0 hertz ac mains frequency fs 100000 hertz topswitch switching frequency vo 5 volts output voltage po 2 5 watts output power n 0.8 efficiency estimate z 0 . 5 loss allocation factor vb 1 2 volts bias voltage tc 3 mseconds bridge rectifier conduction time estimate cin 6 8 ufarads input filter capacitor enter topswitch variables vor 110 volts reflected output voltage ilimitmax 1.65 top224 amps from topswitch data sheet vds 1 0 volts topswitch on-state drain to source voltage vd 0.7 volts output winding diode forward voltage drop vdb 0.7 volts bias winding diode forward voltage drop krp 0.45 ripple to peak current ratio (0.4 < krp < 1.0) enter transformer core/construction variables etd29 core type ae 0.76 cm^2 core effective cross sectional area le 7.2 cm core effective path length al 2100 nh/t^2 ungapped core effective inductance bw 1 9 mm bobbin physical winding width m 3 mm safety margin width (half the primary to secondary creepage distance) l 2 number of primary layers ns 4 number of secondary turns dc input voltage parameters vmin 9 0 volts minimum dc input voltage vmax 375 volts maximum dc input voltage current waveform shape parameters dmax 0.58 duty cycle at minimum dc i nput voltage (vmin) iavg 0.35 amps average primary current ip 0.78 amps peak primary current ir 0.35 amps primary ripple current irms 0.46 amps primary rms current transformer primary design parameters lp 1339 uhenries primary inductance np 7 7 primary winding number of turns nb 9 bias winding number of turns alg 225 nh/t^2 gapped core effective inductance bm 1771 gauss flux density at po, vmin bp 3767 gauss peak flux density (bp < 4200) bac 3 9 9 gauss ac flux density for core loss curves (0.5 x peak to peak) ur 1583 relative permeability of ungapped core lg 0.38 mm gap length (lg >> 0.051 mm) bwe 2 6 mm effective bobbin width od 0.34 mm maximum primary wire diameter including insulation ins 0.06 mm e stimated total insulation thickness (= 2 * film thickness) dia 0.28 mm bare conductor diameter awg 3 0 awg primary wire gauge (rounded to next smaller standard awg value) cm 1 0 2 cmils bare conductor effective area in circular mils cma 219 cmils/amp primary winding current capacity (200 < cma < 500) transformer secondary design parameters isp 14.98 amps peak secondary current isrms 7.62 amps secondary rms current io 5.00 amps power supply output current iripple 5.75 amps output capacitor rms ripple current cms 1667 cmils secondary bare conductor minimum circular mils awgs 1 7 awg secondary wire gauge (rounded up to next larger standard awg value) dias 1.15 mm secondary minimum bare conductor diameter ods 3.25 mm secondary maximum insulated wire outside diameter inss 1.05 mm maximum secondary insulation wall thickness voltage stress parameters vdrain 626 volts maximum drain voltage estimate (includes effect of leakage inductance) pivs 2 4 volts output rectifier maximum peak inverse voltage pivb 5 5 volts bias rectifier maximum peak inverse voltage additional outputs vx 1 2 volts auxiliary output voltage vdx 0.7 volts auxiliary diode forward voltage drop nx 8.91 auxiliary number of turns pivx 5 5 volts aux iliar y rectifier maximum peak inverse volta g e pa g e 1
c 5/98 AN-22 5 example uses the spreadsheet for continuous conduction mode. the techniques described in the following sections to extend the standard single output transformer design to multiple outputs are the same for either spreadsheet. spreadsheet transformer design figure 2 shows the spreadsheet for a transformer that meets the output power and input voltage specification of table 1. a full explanation of the use of the spreadsheet is provided in an-17, but a brief overview will suffice for this explanation. the first section of the spreadsheet is used to input the application variables. note that only the 5 v output is needed to determine the number of turns of the primary, while the total output power for all outputs is specified in this section to select the transformer core, primary inductance and wire gauge. initial design requirements may not be firm enough to determine which topswitch will be used in the final product. the designer usually has to choose between two likely candidates (see an-21). in all designs, whether single or multiple output, the transformer design should accommodate the largest topswitch that might be used with it. a designer may find it necessary to use the larger topswitch (with a lower on-resistance) to permit the use of a smaller heatsink, for example. thus, although the circuit of figure 1 specifies the top223y, the spreadsheet uses the upper current limit value for the top224y/top224p. the higher value is used here to ensure flexibility to allow the use of the top224 should the application require it. the change may be necessary if mechanical restrictions in the available space of the power supply s enclosure force the use of a smaller heatsink. the upper current limit is subsequently used in the spreadsheet to determine the peak flux density b p , which should be limited to prevent excessive core saturation under overload and start up conditions. the ferrite core used here is the industry standard etd29. this is used as an example only. other standard cores such as the ee or eer families can be substituted as desired. the design is based on a margin wound construction, where 3 mm margins are provided at each side of the bobbin to give a total of 6 mm primary to secondary creepage distance. this is the standard creepage distance allowed for mains input power supplies meeting iec950 (or equivalent) isolation. local safety agency requirements for creepage and clearance should be obtained before committing a design to manufacture. other transformer construction techniques, such as slotted bobbin, concentric bobbin or the use of triple insulated wire, are equally applicable. the bobbin style does not influence the calculation of the primary inductance, but specific bobbin width must be input to determine the physical space available for the primary winding. although triple insulated wire techniques are not normally favored in applications requiring a high number of secondary turns, transformer suppliers should be consulted for advice on the optimum construction technique in a particular application. the spreadsheet defines two layers for the primary winding to minimize construction costs. if other cores with reduced bobbin widths are used, additional layers may be necessary to satisfy recommendations for current capacity (cma). it should be noted that an even number of layers will ease construction because the start and finish of the primary winding will be at the same side of the bobbin. the remaining sections of the spreadsheet provide the transformer design that results from the input variables described above. the key parameters that must be checked before a design can be deemed acceptable are detailed in an-17 and summarized in appendix a. since the spreadsheet is written for single output supplies, the transformer secondary design parameters show values assuming the total output power is provided by the 5 v output. it is therefore necessary to extend these calculations to account for the partitioning of output power defined in the power supply specification of table 1. the following section provides the equations necessary to assign appropriate numbers of turns and wire gauges to each output. calculation of secondary turns from the spreadsheet, the 5 v output winding is defined as having 4 turns. the voltage on the cathode of d2 in figure 1 is 5 v. therefore, 4 turns produce the output voltage plus the forward drop of the output diode d2. the volts per turn v pt is defined as: v vv n pt od s = + () (1) where: v pt = volts per turn v o = output voltage (5 v) v d = output diode forward voltage drop (typically 0.7 v for ultra fast pn power diodes and 0.4 v for schottky diodes)
AN-22 c 5/98 6 77 t 0.3 mm (29 awg) 22 t 0.5 mm (24 awg) 30 v (a) separate winding (b) stacked winding pi-2743-120297 rtn 9 t 0.3 mm (29 awg) 4 t 0.5 mm (24 awg) 3 in parallel 9 t 0.5 mm (24 awg) 2 in parallel 12 v rtn 5 v rtn 77 t 0.3 mm (29 awg) 13 t 0.5 mm (24 awg) 30 v 9 t 0.3 mm (29 awg) 5 t 0.5 mm (24 awg) 2 in parallel 12 v 5 v rtn 4 t 0.5 mm (24 awg) 4 in parallel n s = number of secondary turns (4 turns for the 5 v output) substitution of these values into (1) gives: v pt = 1.43 v per turn this value is used to calculate the turns required by the other outputs. simple rearrangement of (1) gives: n vv v s od pt = + () (2) for the 12 v output, v o = 12 v v d = 0.7 v substituting in (2): n s 12 12 0 7 143 = + = (.) . 8.9 turns a practical transformer requires integer numbers of turns; therefore, the 12 v output uses 9 turns. for the 30 v output, v o = 30 v v d = 0.7 v substituting in (2) gives: n s 30 30 0 7 143 21 5 = + = (.) . . turns select 22 turns for the 30 v winding. this last result highlights a frequently encountered problem in multiple output transformers. an integer number of turns, such as 21 or 22, will make the output voltage lower or higher respectively than desired. since this is a high voltage output with a large number of turns, the difference between the desired value and the integer value amounts to only about 2%. the resulting change in output voltage is not significant, and will be masked by other factors such as cross regulation and diode characteristics. however, it is worth mentioning the options available should this problem be encountered with figure 3. transformer winding diagrams showing two techniques for the secondary winding.
c 5/98 AN-22 7 lower voltage outputs where the requirement for integer numbers of turns can introduce a significant deviation from the desired value. 1. if the output in question requires a high degree of accuracy, then a higher output voltage can be defined in equation (2) and a linear post regulator employed to achieve the output voltage. 2. if the tolerance is less critical, a series resistor and a zener diode of appropriate value can be used as a shunt regulator for low power outputs. 3. the fundamental transformer design could be modified such that the main 5 v output uses a number of turns which yields an integer number of turns on the other windings when calculated using equations (1) and (2). 4. the choice of rectifier on the main regulated output can be used to influence the volts per turn. if a schottky diode with a forward voltage of typically 0.4 v were employed on the 5 v output, the v pt from (1) would be 1.35. a standard pn diode on the 30 v output would from equation (2) yield 22.7 turns, which is closer to the integer number 23. use of the schottky diode with 4 turns on the 5 v output, however, would decrease the accuracy of the 12 v output. the required number of turns would move farther away from an integer value, from 8.9 to 9.4 turns. the designer can investigate alternative integer turns ratios with both schottky and pn diodes by repeating the spreadsheet design for other values of secondary turns. if a need for higher efficiency calls for a schottky diode on the 5 v output, then 3 turns on the 5 v output with 7 and 17 turns for the 12 v and 30 v outputs respectively may give acceptable results. designers often use the "golden ratios" of 3:7:9 with a schottky diode for the 5 v output and a pn diode for the 12 v output, or 4:9:11 with all pn diodes to achieve outputs of 5, 12 and 15 v. another useful ratio is 2:3 for outputs of 3.3 and 5 v with schottky diodes on each. the turns could be in the ratio of 3:4 if the 3.3 v output uses a pn diode and the 5 v uses a schottky diode. all designs need to be tested thoroughly to verify acceptability. in practice, if tight tolerance is required on windings other than the main feedback output, some form of post regulation or combined feedback circuitry is often necessary. these issues of cross regulation are discussed later in the section on circuit performance. in this case, as mentioned above, the choice of 22 turns for the 30 v output will not introduce a significant inaccuracy. the final choice of turns on each output is therefore shown in figure 3(a), and summarized as follows: 5 v 4 turns 12 v 9 turns 30 v 22 turns figure 3 illustrates two winding diagrams: one with separate windings for each output and one with stacked output windings. these two configurations are discussed in detail later in the section on transformer construction. choice of output wire gauge appropriate wire gauge for the outputs is determined on the basis of the maximum continuous rms current rating for each winding. the analysis of the distribution of current in the various outputs can be very complex, but a few reasonable assumptions make the task easy. the waveshapes of the currents in the individual output windings are determined by the impedances in each circuit. leakage inductance, rectifier characteristics and capacitor values are some of the parameters that affect the magnitude and duration of the currents. the average currents are always equal to the dc load current, while the rms values are functions of peak magnitudes and conduction times. the rms values determine the power dissipation in the windings. for ordinary multiple output designs it is valid to make the reasonable simplifying assumption that all output currents have the same shape as for the single output case. this is the case of greatest dissipation. ultimately the final design of the transformer has to be decided on the basis of tests and consultation with transformer suppliers. however, the first order analysis that assumes the same waveshape for all output currents provides a start point for the choice of wire gauge. the single output design of the spreadsheet calculates the rms current in the secondary as if the 5 v winding supplied all the power. however, from the specification of table 1, the 5 v output supplies a maximum of 10 w. the actual currents in the multiple output application are computed from quantities on the single output spreadsheet. since we assume the currents in the output windings have the same shape, each will have the same ratio of rms to average as the single output case. if k ra is the ratio of rms to average current, then k i i ra srms o == = 762 5 1 524 . . a a (3)
AN-22 c 5/98 8 493 and 197 cma respectively) allows acceptable power dissipation in the majority of applications, depending on the conditions of maximum ambient temperature and efficiency requirements. in the united states, it is common to use the reciprocal of current density expressed as circular mils per ampere (cma). one mil is 0.001 inch, and the area in circular mils is the square of the wire diameter in mils. one circular mil is 7.854 10 -7 in 2 or 5.067 10 -4 mm 2 . based on 9 a/mm 2 (219 cma), using the rms current calculated above, the minimum bare copper diameter for each output is: 5 v output 0.66 mm (22 awg) 12 v output 0.51 mm (24 awg) 30 v output 0.07 mm (41 awg) the above calculations define the minimum wire diameter specifications. however, practical considerations of transformer manufacture determine the actual wire gauges used. for example, two or three parallel windings on the higher current outputs can reduce the required wire diameter while optimizing coverage of the bobbin. these issues are discussed in detail next. transformer construction primary winding techniques are well documented in an-18 where i srms and i o are from the spreadsheet. to find the rms current in a winding, we simply multiply its average current by k ra . iik rmsx x ra = (4) hence, the rms current in the 5 v winding is i rms 5 20 1524 305 = . a . = . a and the rms current on the 12 v winding is i rms 12 12 1524 183 == ... a a similar calculations for the 30 v output yield i rms 30 30 5 = . ma the wire diameter can be chosen on the basis of the total dissipation in the output winding. one can find the resistance of the winding from the resistance per unit length of a particular wire gauge and the length of the wire associated with each output winding. however, a calculation based on the current density can be used to make a first estimate of the required wire gauge on each output. a current density between 4 and 10 a/mm 2 (corresponding to table 3. comparison of secondary winding techniques in margin wound transformers. disadvantages 1. poor regulation of lightly loaded outputs due to peak charging. 2. generally higher manufacturing costs. 3. more pins on bobbin. 1. winding with lowest or highest voltage output must be placed closest to the primary winding ?no flexibility to reduce leakage inductance of outputs with higher currents. winding technique separate output windings stacked output windings advantages 1. flexibility in winding placement; output with highest current can be positioned closest to primary to minimize energy lost from leakage inductance. 1. improved cross regulation. 2. generally lowest cost manufacturing technique.
c 5/98 AN-22 9 and are not influenced by the number of output windings. there are, however, two secondary winding techniques commonly used in margin wound transformers. these are described below and summarized in table 3. other transformer constructions such as slotted bobbin and concentric bobbin designs may demand other considerations. the designer should consult with the specific transformer supplier to insure the optimum technique in each case. separate output windings the winding diagram of figure 3(a) shows each output wound as a separate coil. in this way each winding conducts only current associated with the specific load on that output. since each output is wound as a separate operation, this construction technique provides flexibility in the placement of output windings relative to the primary winding. this freedom can be an important consideration in multiple output transformers to minimize the leakage inductance. the leakage inductance of a transformer is the inductance associated with flux which does not link all windings. as such, this flux does not contribute to the transfer of energy. in single output transformer structures, all the leakage is usually measured on the primary by shorting the output winding and measuring the resulting inductance of the primary. this provides a good estimate of the energy which the primary clamp circuitry will dissipate. in figure 1, components d1 and vr1 are specified for clamping the leakage energy. however, in a multiple output design, there are leakage inductances associated with each output winding according to its coupling to the primary and to other secondary windings. placement of the output windings should be made to minimize the leakage inductance associated with outputs that provide the most current. for example, in the circuit design of this application note, the 5 v and 12 v outputs handle most of the power with 2 a and 1.2 a respectively, while the 30 v output has a load of only 20 ma. the windings therefore should be figure 4. schematic of multiple output 25 w power supply with stacked secondary windings. 5 v rtn br1 400 v c1 68 f 400 v u1 top223 r2 100 ? 1/2 w d8 mbr745 d3 1n4148 c2 1000 f 35 v t1 d1 byv26c c7* 1000 pf 250vac y1 c3 120 f 25 v u2 cny 17-2 u3 tl431 r4 10 k ? r5 10 k ? c9 0.1 f r1 75 ? vr1 p6ke200 l1 3.3 h f1 1.0 a j1 c8 0.1 f l2 33 mh l n * two series connected, 2.2 nf, y2-capacitors can replace c7. d4 mur420 l3 3.3 h d5 uf4004 c12 100 f 50 v 30 v c11 100 f 35 v c6 100 f 50v 12 v c5 47 f d s c control r3 5.2 ? topswitch-ii c10 390 f 35 v pi-2125-121197 c4 0.1 f
AN-22 c 5/98 10 arranged such that the 5 v and 12 v outputs have the best coupling to the primary winding. an arrangement that has the 30 v winding closest to the primary may show the same primary leakage inductance as the preferred structure when measured with the standard technique of shorting all outputs together. in the application, however, efficiency will be reduced since the leakage inductance associated with the 5 v and 12 v outputs will be higher. the use of separate output windings provides complete flexibility in the winding arrangement. in this case the optimum configuration for separate layers might be to wind the 5 v output first followed by the 12 v winding and finally the 30 v output. that is, the winding with the greatest output current would go next to the primary. an even better arrangement would have the two highest current windings share a single layer using the nesting technique illustrated in appendix c. separate windings, however, tend to increase the cost of the transformer since every output winding is a separate operation. the alternative stacking technique described below improves the regulation, particularly on lightly loaded outputs. stacked output windings figure 3(b) shows a stacked output winding configuration, which is generally favored by transformer manufacturers. the windings of the 5 v output provide the return and part of the windings for the 12 v output. similarly, the 30 v output uses the turns of the 5 and 12 v outputs and additional turns to make up the full winding. the wire for each output must be sized to accommodate its output current plus the sum of the currents for the other outputs stacked on top of it. the stacked configuration improves cross regulation while reducing construction costs. consider this example where the 5 v output is fully loaded but the 12 v and 30 v outputs have minimum load applied. with separate output windings, the capacitors on the 12 v and 30 v outputs would tend to peak charge under the influence of leakage inductance. however, with a stacked winding, the fact that the 5 v output forms part of the 12 v and 30 v windings reduces the impedance of these windings and reduces the effect of peak charging. the only disadvantage of this winding technique is that there is little flexibility in the placement of the windings relative to the primary. either the 30 v or 5 v winding must form the start of the output windings closest to the primary. in this case, since the 5 v has the highest loading, it is defined as the start of the secondary winding. since the stacking technique generally offers the best cross regulation, the winding construction of figure 3(b) was chosen for the example circuit in this application note, as illustrated in figure 4. the only difference between t1 in figures 1 and 4 is the use of the stacked winding technique on the transformer in figure 4. construction to improve cross regulation the cross regulation is a measure of how well the output voltages regulate under the influence of varying load conditions on other outputs. the quality of cross regulation depends on the coupling between the various output windings. the better coupled these windings are, the better the cross regulation. as such, it is recommended that each individual winding is wound to cover the complete bobbin width. therefore, the easiest way to wind the transformer is to use several parallel wires of the same gauge to insure the bobbin is well covered. in this case, the total copper area used by the 5 v winding must handle the total rms current of all outputs. the total output rms current is: iiii rmstot rms rms rms =+ + = 5 12 30 503 . a this summation is possible only when the currents have the same shape, which is a valid simplifying assumption for the design. based on a current density of 9 a/mm 2 (219 cma), the copper diameter of a single wire would need to be 1.03 mm (20 awg). however, if the wire is split into several parallel sections, each carrying an equal share of the current, we may use a smaller diameter wire which is much easier to handle during manufacture. figure 5. cross section of bobbin showing five interleaved turns of four parallel conductors on a single layer. turn 1 finish } turn 5 } turn 2 } turn 4 } turn 3 } start pi-2128-120297
c 5/98 AN-22 11 also, the multiple parallel strands of thinner wire can be placed flat for good coverage of the bobbin as shown in figure 5. this will insure that the winding is well coupled to the primary and to the other secondaries that are wound afterwards. in this example we chose to split the 5 v winding into six conductors to fit the pin arrangement of the bobbin. one pin can accommodate three wires. since each wire carries one sixth the current, or 0.84 a rms, we may use a wire diameter of 0.4 mm (27 awg), which is much easier to handle during manufacture. the 12 v winding must handle a total of 1.86 a rms (i rms12 + i rms30 ). to maintain a maximum current density below 9 a/mm 2 (219 cma) we can use the same 0.4 mm (27 awg) wire with the number of parallel strands reduced to 2. again this should be wound evenly across the bobbin with turns distributed to provide the optimum coupling with the 5 v and primary winding. appendix c shows how to put both windings on the same layer for best coupling. finally, the 30 v winding is added across the entire bobbin width. this winding carries the current for only the 30 v load; therefore, we can use a single strand of the 0.4 mm diameter (27 awg) wire. if desired, a thinner wire gauge may be specified to reduce the volume occupied by the winding. the same wire may be used in all windings to reduce cost. appendix c illustrates these methods with complete construction details of the transformer used in this application note. the techniques detailed above should be used in the transformer construction to optimize cross regulation. however, additional external circuit techniques to further enhance cross regulation are discussed in the section on circuit performance. output rectifier specification as in single output converters, the proper choice of output rectifiers in multiple output converters is essential to achieve desired performance and reliability. it is important to use only schottky and ultra fast pn junction rectifiers. the effects of the reverse recovery characteristics on the primary circuit are amplified in multiple output applications because the output rectifiers are effectively in parallel. refer to an-19 for a discussion of how the selection of output rectifiers influences efficiency. the specification on each output rectifier diode is determined on the basis of the required voltage and current rating. the peak inverse voltage (piv) on each diode is given by: piv v v n n x x max x p =+ ? ? ? ? ? ? (5) where v x is the voltage of the particular output, n x is the number of output turns on the particular output and n p is the transformer primary turns. v max is the maximum primary dc rail voltage, which for 230 vac input applications is typically 375 vdc (peak value of 265 vac). for the transformer in this example, n p = 77 turns v max = 375 v hence, for the 5 v output, figure 6 (a). cross regulation with feedback from 5 v only. response to a 5 v load. 95 0.50 0.75 1.00 1.25 1.50 (a) 5 v load (a) output voltage (% of nominal) pi-2142-121997 105 100 5 v 12 v 30 v 1.75 2.00 95 0.50 0.75 1.00 1.25 1.50 (b) 5 v load (a) output voltage (% of nominal) pi-2144-121997 105 100 1.75 2.00 5 v 12 v 30 v figure 6 (b). cross regulation with feedback from 5 v and 12 v. response to a 5 v load.
AN-22 c 5/98 12 piv 5 5 375 4 77 25 =+ ? ? ? ? = v for the 12 v output, piv 12 12 375 9 77 56 =+ ? ? ? ? = v for the 30 v output, piv 30 30 375 22 77 137 =+ ? ? ? ? = v the diodes chosen for each output should have a reverse voltage rating 1.25 piv x . this insures that the peak reverse voltage never exceeds 80% of the rating of a particular diode. hence, in this case, the diode on the 5 v output should be rated for more than 30 v, the 12 v output more than 70 v, and the 30 v output more than 171 v. peak reverse voltages should be measured on all diodes under maximum load and startup conditions to ensure that ratings are not exceeded. the rule of thumb for the diode current rating is to choose a device with a dc current rating at least three times the average dc output current of the particular output. from the current specifications of table 1 and the voltage requirements above, the following minimum ratings should be defined in this case: 5 v output diode 6.0 a, 30 v 12 v output diode 3.6 a, 70 v 30 v output diode 60 ma, 171 v figure 7. modified schematic with feedback from both 5 v & 12 v outputs. 5 v rtn br1 400 v c1 68 f 400 v c4 0.1 f u1 top223 r2 100 ? 1/2 w d8 mbr745 d3 1n4148 c2 1000 f 35 v t1 d1 byv26c c7* 1000 pf 250vac y1 c3 120 f 25 v u2 cny 17-2 u3 tl431 r4 21 k ? r5 10 k ? c9 0.1 f r1 75 ? vr1 p6ke200 l1 3.3 h f1 1.0 a j1 c6 0.1 f l2 33 mh l n * two series connected, 2.2 nf, y2-capacitors can replace c7. d4 mur420 l3 3.3 h d5 uf4004 c12 100 f 50 v 30 v c11 100 f 35 v c6 100 f 50v 12 v c5 47 f d s c control r3 5.2 ? topswitch-ii c10 390 f 35 v pi-2131-121197 r6 75 k ?
c 5/98 AN-22 13 for reverse voltage ratings less than 100 v, schottky diodes can be used to minimize power losses. as discussed earlier, schottkys can also be used to improve the relative accuracy of output voltages when calculating the number of turns. schottky diodes are more expensive than pn junction diodes. the circuit of figure 1 uses ultra fast recovery pn diodes for the lowest cost, while the circuit in figure 4 uses a schottky diode on the 5 v output with the same transformer design. circuit performance may be improved with a transformer designed specifically for a schottky diode on the 5 v output. in this example many possible diodes are available to achieve the required characteristics. the devices in the example of figure 4 are: 5 v output: mbr745 7.8 a, 45 v motorola 12 v output: mur420 4.0 a, 200 v motorola 30 v output: uf4004 1.0 a, 400 v general semiconductor other suitable diodes are available from different manufacturers. tests with a number of diodes are recommended to verify the optimum devices in each application. circuit performance the volts per turn defined in equation (1) is an approximation based on the forward voltage of the output diode. this value changes with load current and temperature. as the outputs have varying loads, the output diodes will exhibit different forward voltages depending on the load conditions on the particular output. changing load conditions on the 5 v output, for example, will inherently influence the voltages on the other outputs. in addition, secondary effects such as voltage spikes from leakage inductance and quality of coupling between output windings, lead to reduced voltage accuracy on outputs which do not provide feedback through the optocoupler. the basic circuit of figure 4 derives feedback only from the 5 v output. as a consequence, the other output voltages vary as the 5 v output current changes. the influence on the 12 v output is shown in figure 6(a). use of a schottky diode in a circuit designed for a pn diode emphasizes the effect of a change in voltage drop, as illustrated in this example. the 5 v output voltage is well controlled since it exclusively provides the feedback signal. the 12 v output, however, is seen to vary by 2% as the 5 v load is varied between 25% and 100% (0.5 amps to 2.0 amps). for this test the 12 v output load was held constant at 0.6 amps. the 12 v and 30 v outputs are also below their nominal values because of the lower drop of the schottky diode. transformer construction techniques to optimize output cross regulation were discussed earlier. however, it is often necessary to further enhance cross regulation using external circuit techniques. for example, if improved regulation is required on the 12 v output, a simple technique is to derive the feedback from both 5 v and 12 v outputs. in this example, as in most applications, higher accuracy is required on one of the outputs. here it is assumed that the main output is still the 5 v, but some feedback may be drawn from the 12 v output to improve its figure 8(a). cross regulation with feedback from 5 v only. response to variation of 12 v load 95 0.00 0.20 0.40 0.60 0.80 (a) 12 v load (a) output voltage (% of nominal) pi-2146-121997 105 100 1.00 1.20 5 v 12 v 30 v 95 0.00 0.20 0.40 0.60 0.80 (b) 12 v load (a) output voltage (% of nominal) pi-2148-121997 105 100 1.00 1.20 5 v 12 v 30 v figure 8(b). cross regulation with feedback from 5 v and 12 v. response to variation of 12 v load
AN-22 c 5/98 14 load regulation. the schematic of figure 7 illustrates a simple modification to the original circuit of figure 4, where resistor r6 is introduced from the 12 v output to the reference pin of the tl431 shunt regulator. figure 6(b) illustrates the improvement obtained by employing this new feedback scheme where the load regulation on the 12 v output is improved to 1.5%. the effect would be more dramatic if the transformer had greater leakage inductance on the output windings. the value of r6 is generally determined through iteration and depends on the degree of feedback desired from the second output. introducing feedback from a second winding has a detrimental effect on the regulation of the main output. in this example the change in the 5 v output increases from effectively 0% in figure 6(a) to 0.75% in fig 6(b). a good rule of thumb as a start point for tests is to choose r6 such that it yields about 10% of the current in r4 (with the tl431 reference pin at 2.5 v). in this example the current in r4 before modification is: i r 4 525 10 250 = ? = (.) v k a ? to emphasize the effect, we let the 12 v output provide 50% of this amount through r6. assuming that the tl431 reference pin is still at 2.5 v r 6 12 2 5 125 76 = ? () = . v a k ? a standard resistor value of 75.0 kohm was chosen for r6 in figure 7. figure 9. modified schematic of figure 4 with isolated 30 v output and c13 for common mode current return. pi-2129-121197 5 v rtn br1 400 v c1 68 f 400 v c4 0.1 f u1 top223 r2 100 ? 1/2 w d2 mbr745 d3 1n4148 c10 1000 f 35 v t1 d1 byv26c c7* 1.0 nf y1 c11 120 f 25 v u2 cny 17-2 u3 tl431 r4 10 k ? r5 10 k ? c9 0.1 f r1 75 ? vr1 p6ke200 l1 3.3 h f1 1.0 a j1 c8 0.1 f l2 33 mh l n * two series connected, 2.2 nf, y2-capacitors can replace c7. d4 mur420 l3 3.3 h d5 uf4004 c2 100 f 50 v 30 v c3 390 f 35 v c6 100 f 35 v 12 v c5 47 f d s c control r3 6.2 ? topswitch-ii isolated rtn c13 1 nf 500 v c12 100 f 50 v
c 5/98 AN-22 15 note that since the sum of the currents through r4 and r6 is a constant equal to 2.5v divided by r5, the additional feedback that r6 introduces from the 12 v output will tend to reduce the regulated value of the 5 v output. this requires that r4 is in turn adjusted to retain the 5 v output at the desired level. to retain the voltage at the reference pin of the tl431 at 2.5 v, the value of r4 must therefore be increased to reduce its current by 50%. r 4 525 250 125 20 = ? () ? () = . v a k ? a slightly larger precision 21.0 kohm resistor was specified in the circuit of figure 7 to compensate for the small penalty in regulation on the 5 v output. figure 8 shows load regulation measurements before and after these circuit modifications with the 5 v output load held constant at 1 a and the 12 v output load varied from 10% to 100%. in figure 8(a), the 5 v output is very stable since this is exclusively providing output feedback, while the 12 v output drops by 4% over the load range. in figure 8(b), the introduction of r6 maintains tighter regulation on the 12 v output ( 1.5% variation with load), whereas this additional feedback introduces a 0.75% variation in the 5 v output voltage over the same load range. the degree of feedback required from each output can thus be determined depending on the application requirements for output voltage tolerance. breadboard evaluation is necessary to adjust component values for the desired performance. emi considerations in general, the emi considerations in a multiple output topswitch power supply do not differ from those of a single output supply, and are covered in detail in an-15. there are, however, specific multiple output power supplies where figure 10. modified schematic of figure 4 with soft-start capacitor c15 added. 5 v rtn br1 400 v c1 68 f 400 v c4 0.1 f u1 top223 r2 100 ? 1/2 w d8 mbr745 d3 1n4148 c2 1000 f 35 v t1 d1 byv26c c7* 1000 pf 250vac y1 c3 120 f 35 v u2 cny 17-2 u3 tl431 r4 10 k ? r5 10 k ? c9 0.1 f r1 75 ? vr1 p6ke200 l1 3.3 h f1 1.0 a j1 c8 0.1 f l2 33 mh l n * two series connected, 2.2 nf, y2-capacitors can replace c7. d4 mur420 l3 3.3 h d5 uf4004 c12 47 f 50 v 30 v c11 120 f 35 v c6 100 f 50v 12 v c5 47 f d s c control r3 5.2 ? topswitch-ii c10 470 f 35 v pi-2132-121897 c15 22 f 25 v
AN-22 c 5/98 16 figure 11 (a), (b). two configurations to get negative outputs. additional measures are necessary to optimize the emi performance. this is particularly true when the outputs are galvanically isolated from each other. in motor control circuits, for example, several isolated outputs may be required to supply high side drivers in an inverter output stage. in these cases it is important that displacement currents driven by the topswitch drain node through the transformer s interwinding capacitance have a low impedance return path from a specific output to the primary side of the power supply. this consideration demands that each isolated output provide a low impedance path for common mode displacement currents to return from its own return to the primary return ( topswitch source potential). this low impedance path can usually be provided from the output s return through a capacitor (suitably rated for the isolation voltage required on a particular output) to the main secondary return, from where a safety y capacitor is connected to the primary return rail. this configuration is shown in figure 9, where the isolated 30 v output has a 500 v capacitor, c13, connected between its return rail and that of the main power supply output. if these low impedance capacitive paths are not provided on each isolated output, then the common mode displacement currents transferred through the transformer interwinding capacitance will return to their source on the primary of the transformer through any alternative route that is available. the common mode currents may split many times on their route to the drain node. if a capacitive return path is not present, there is the risk that enough of the displacement current will flow through the ac input conductors to fail regulatory emission specifications. the need for additional capacitors in this type of circuit depends on the transformer s interwinding capacitance. additional capacitors from an isolated output may not be necessary if its capacitance to the primary is low enough. however, tests are essential to verify the necessity of additional components. one other emi consideration related to output diode snubbers is worthy of note. output diodes are always a source of additional noise that depends on their forward and reverse recovery characteristics, particularly the di/dt and dv/dt during recovery. many diodes are now available with so called soft recovery characteristics which are designed to limit switching noise. it is often desirable, however, to further snub the diode characteristics with external components. these external snubbers are usually a single capacitor, or series resistor and capacitor in parallel with the output diodes. in many cases the snubbing circuitry can be limited to a single output diode to achieve the desired reduction in switching noise. in such cases, the highest voltage winding with significant loading should be chosen for the snubber circuitry. in this example, the 12 v output diode would be chosen since the capacitors on that output have lower esr than the capacitors on the 30 v output. it also has the best overall coupling with the primary winding because it is physically closest. during the primary switching events, these snubber components are an ac current path in series with the output electrolytic capacitors. they therefore provide a low impedance ac path across the transformer output winding and the output diode to confine the noise currents created by primary switching events. +v output (a) (b) pi-2130-120297 output rtn v output dc rail drain primary rtn bias +v output output rtn v output dc rail drain primary rtn bias
c 5/98 AN-22 17 additional tips following are some tips which can be considered and tested where necessary to improve circuit performance. optocoupler connection in multiple output power supplies, the current for the optocoupler led is often supplied via the loop gain setting resistor from an output other than the main feedback voltage. in figure 4, this connection of r1 is made to the 5 v winding. this technique introduces some ac feedback from the 5 v winding, which helps reduce variation on that output during transient load conditions. r1 and r2 may be connected to the 12 v output instead of the 5 v output (with their values changed appropriately). ripple current from this output has a path to the tl431 reference pin via r1 and c9. this type of connection, however, will often introduce loop instability with very light loads on the 12 v output. the reason is that the 12 v output is subject to peak charging from energy in leakage inductance as its load approaches zero. peak charging effectively uncouples the output so that it is no longer related to the 5 v output by the turns ratio. if instability is observed during light or no load conditions on the 12 v output, two options are available: 1. the optocoupler led should be supplied from the 5 v winding (with the value of r1 selected to maintain acceptable ac gain) or 2. a dummy or minimum load resistor can be added to the 12 v output to eliminate the effects of peak charging. dummy loads are usually added to improve regulation at light loads. r2 is used for this purpose on the 5 v output in figure 4. r2 might be moved to the 12 v output if one dummy load is sufficient to meet specifications. the value of this resistor should be adjusted as necessary to allow for the load range of a particular application. soft start circuitry soft start circuitry is often useful to avoid output voltage overshoot during power supply turn on. this is achieved simply by introducing a capacitor from the tl431 cathode to anode as shown by c15 in figure 10. note a discharge path is required for this capacitor to insure the soft start function is reset when the output voltages decay at turn off. this function is provided in figure 10 by the minimum load resistor r2. when introducing soft start, it is useful to supply the optocoupler led from a higher voltage output, such as the 12 v rail in this case, since this will insure that c15 begins to charge and provide the soft start function as soon as possible after the power supply starts to operate. the issues of minimum load on the higher voltage output, discussed above, must be considered when doing this to insure loop stability under all conditions. improving regulation in lightly loaded outputs some outputs, such as the 30 v output in this example, can have very light loads even under maximum load conditions. these are prone to peak charging, which can produce output voltages much higher than expected by the turns ratio of the transformer output. the degree of this peak charging is strongly influenced by the loads on the other outputs. the output in question can simply be clamped with a zener diode between the output and secondary return. however, a lower cost and more efficient solution is to provide some low pass filtering that will reject the short voltage spikes from leakage inductance to prevent charging of the output capacitors. the introduction of a resistor in series with d5 in figure 4 will provide this function. values from 10 to 100 ohms should be tested to determine the optimum. see r6 in figure 1. negative outputs negative outputs are often required in a system for operational amplifiers or other analog circuitry. two simple configurations generally used to provide these outputs are shown in figure 11. figure 11(a) shows the most usual configuration, where the direction of the output diode is reversed such that that diode's cathode is connected to the transformer s output pin. the other end of the negative winding is connected to the common secondary return using the same dot convention as the other output windings. an alternative technique connects the anode of the output diode to the return end of the winding with the cathode connected to the common secondary return as shown in figure 11(b). the alternate, however, is not available with stacked windings. the calculation of the number of output turns is identical to that for positive outputs, and the same transformer construction techniques are used to optimize cross regulation. since negative outputs are often lightly loaded, the techniques to improve regulation in lightly loaded outputs detailed above are often useful. alternatively, the output can simply be post regulated with a linear regulator.
AN-22 c 5/98 18 appendix a key spreadsheet variables. the following key variables in the transformer design spreadsheet of figure 2 should be checked before a transformer design can be deemed acceptable: d max must be less than the topswitch data sheet minimum value of 64% (0.64). i p to allow for thermal effects, this should be no greater than 90% of the data sheet minimum current limit specification for the chosen topswitch at 25 c. in this example, the minimum current limit for the top223y is specified as 0.9 a, so the spreadsheet value of 0.78 meets the above criterion. b p this must be below the recommended value of 4200 gauss to avoid excessive core saturation at the peak topswitch current limit. here the value of 3767 gauss is well within this requirement. l g although the guidance of transformer vendors should be sought, airgaps of <0.051mm are not recommended because such small gaps make it difficult to hold a reasonable tolerance on the specified primary inductance. cma values between 200 and 500 allow reasonable temperature rise in the windings. smaller values indicate higher temperatures from greater losses in the copper. the value of 219 circular mils per amp in figure 2 meets the recommended lower limit of 200. again, an-17 should be consulted for full details on the use of the spreadsheet of figure 2.
c 5/98 AN-22 19 appendix b 3.3 v and 5 v outputs an increasing number of applications require that both 3.3 v and 5 v outputs in multiple output power supplies, both requiring 5% regulation to supply digital control circuitry. several commonly used techniques to achieve this performance are described below. linear regulator the simplest, though least efficient technique, is to design only a 5 v output winding with wire capable of supplying the rms current for both the 5 v and 3.3 v outputs. a linear regulator is then placed on this 5 v output, regulating down to 3.3 v as shown in figure 1. integrated 3.3 v regulators are now available from a number of suppliers with varying current capabilities. a simple emitter follower regulator could also be employed using discrete components. the disadvantage of this technique is reduced power supply efficiency, although it simplifies the transformer construction and reduces the number of output pins. transformer turns ratio two techniques are commonly used to design separate transformer windings for each output. each has the required turns ratio relationship to provide the regulation required. 1. copper wire if 3 turns are defined for the 3.3 v output and an ultra fast pn junction diode is specified for this output, the calculation of the volts per turn provides a solution where 4 turns are used with a schottky diode for the 5 v output. v vv n pt od s = + () (1) if v o = 33 . v v d = 07 . v n s = 3 turns then from (1) find v pt = 133 . v per turn rearranging (1) to calculate the turns required for the 5 v output yields: n vv v s od pt = + (2) if v s = 5 v v d = 04 . v figure 1. derivation of 3.3 v output from 5 v with linear regulator. pi-2133 -121597 5 v rtn 3.3 v linear regulator
AN-22 c 5/98 20 v pt = 133 . v per turn from (2), the turns required on the 5 v output are: n s = 406 . turns this result demonstrates that this choice of turns and output diodes yields an almost perfect integer turns ratio between the 3.3 v and 5 v outputs. it is a very popular solution for this reason. the coupling between the output windings is still a crucial factor to insure that the turns ratio calculated above does indeed result in the required output cross regulation. since so few turns are involved in these outputs, it is usual for multiple parallel wire strands to be used on each output winding, and for the 3.3 v and 5 v outputs to be constructed as separate windings. stacked windings are not appropriate in this case. as discussed in the body of the application note, windings should be constructed in 2 layers and interleaved across the bobbin width to optimize coupling with the primary winding. 2. foil windings an alternative technique is to use foil instead of multiple strands of copper wire. using this technique, the turns ratio of 3 turns on the 3.3 v output and 4 turns on the 5 v output is retained. the foil is cut to the required length with appropriate termination points included prior to winding. the foil is then wound as a single operation. termination to the transformer pins is performed afterwards. figure 2 illustrates this technique. the foil is prepared to fit the bobbin width of the chosen transformer exactly, and is backed with insulation material which is wrapped around the foil to provide creepage distances appropriate to the required isolation requirements of the application. although this technique may add some cost to the transformer construction, the fact that the foil is prepared to the exact bobbin width provides excellent coupling with the primary winding. in addition, the 3.3 v and 5 v windings have very good mutual coupling that improves cross regulation. this mutual coupling makes the stacked winding construction the preferred technique when using foil windings. as shown in figure 3, subsequent output windings can be stacked on the foil windings, though the total rms current requirements must accounted for in the choice of the foil. independent of whether copper wire or foil winding techniques are used, the output feedback configuration must be determined according to the load and regulation requirements. figure 4 shows the use of a tl431 where the feedback is derived from both the 3.3 v and 5 v outputs. the proportion of feedback from each output can be adjusted as required, and is discussed in detail in the body of the application note. figure 2. preparation of foil windings for 5 v and 3.3 v outputs. pi-2134 -121897 foil prepared to exact bobbin width return termination point 3.3 v output termination point 5 v output termination point insulated backing wrapped to provide creepage distance
c 5/98 AN-22 21 figure 3. winding arrangement with foil and wire for multiple outputs. figure 4. use of feedback from both outputs with tl431 to improve regulation on 3.3 v output. pi-2136 -121997 foil foil primary wrapped insulation additional secondary windings 5 v 3.3 v pi-2138 -121997 rtn
AN-22 c 5/98 22 appendix c transformer construction details pi-2154-020598 1 2 3 4 5 6 7 8 core, etd29 bobbin, etd29-1s-13p, 13 pin wire, 30 awg heavy nyleze wire, 27 awg heavy nyleze tape, epoxy 2.5 mm wide tape, polyester 14 mm wide tape, polyester 19 mm wide varnish parts list for transformer design example 1pr. 1ea. a/r a/r a/r a/r a/r a/r item amt. description part # manufacturer 4312 020 3750* 4322 021 3438 #10 #1298 #1298 philips philips 3m 3m 3m *gap for a l of 225 nh/t 2 5% pi-2140-121997 77 t #30 awg 9 t 3x #27 awg 13 t #27 awg core# - etd29 (philips) gap for a l of 225 nh/t 2 bobbin# 4322 021 3438 (philips) 6 5 3 1 13 1 3 5 6 7, 8 9, 10 11 12 13 high-voltage dc bus topswitch drain v bias primary-side common return +5 v output +5 v output connection +12 v output +30 v output pin function 11 electrical strength creepage primary inductance resonant frequency primary leakage inductance 3000 vac 5.0 mm (min) 1340 h, 10% 1 mhz (min) 34 h (max) 60 hz, 1 minute, from pins 1-6 to pins 7-13 between pins 1-6 and pins 7-13 all windings open all windings open pins 7 through 13 shorted electrical specifications note: all inductance measurements should be made at 100 khz 1 7 7 1 6 13 10 9 8 7 4 t #27 awg x6 5 t #27 awg x2 12 margin wound transformer
c 5/98 AN-22 23 pi-2152-020498 primary and bias margins double primary layer basic insulation bias winding reinforced insulation output margins +5 v and +12 v winding basic insulation +30 v winding outer assembly final assembly tape margins with item [5]. match height with primary and bias windings. start at pin 3. wind 39 turns of item [3] from left to right. wind in a single layer. apply 1 layer of tape, item [6], for basic insulation. wind remaining 38 turns in the next layer from right to left. finish on pin 1. 1 layer of tape [6] for insulation. start at pin 5. wind 9 parallel trifilar turns of item [4] from left to right. wind uniformly, in a single layer, across entire width of bobbin. finish on pin 6. 3 layers of tape [7] for insulation. tape margins with item [5]. match height with all output windings start with two sets each containing three wires item [4], and one pair of wires item [4]. terminate first set of three wires to pin 9 and the second set of three wires to pin 10. terminate the pair of wires to pin 12. wind the combination of eight wires in parallel right to left evenly across the bobbin, with the pair of wires closest to the right side of the bobbin. after four turns of the combination of eight wires, terminate the first set of wires to pin 8 and the second set of wires to pin 7. continue to wind the pair of wires one more turn for five turns total. finish at pin 11. 1 layer of tape [6] for basic insulation. start at pin 13. wind 13 turns of item [4] from right to left. wind uniformly, in a single layer, across entire width of bobbin. finish on pin 12. 3 layers of tape [7] for insulation. assemble and secure core halves. impregnate uniformly with varnish. winding instructions 6 5 1 3 12 10 margin wound transformer construction tape tape margins (4 places) bias primary tape +30 v +5, +12 v tape tape 9 8 7 11 13 12
AN-22 c 5/98 24 korea power integrations international holdings, inc. rm# 402, handuk building 649-4 yeoksam-dong, kangnam-gu, seoul, korea phone: +82-2-568-7520 fax: +82-2-568-7474 e-mail: koreasales@powerint.com world headquarters americas power integrations, inc. 5245 hellyer avenue san jose, ca 95138 usa main: +1 408-414-9200 customer service: phone: +1 408-414-9665 fax: +1 408-414-9765 e-mail: usasales@powerint.com for the latest updates, visit our web site: www.powerint.com power integrations reserves the right to make changes to its products at any time to improve reliability or manufacturability. power integrations does not assume any liability arising from the use of any device or circuit described herein, nor does it convey any license under its patent rights or the rights of others. the pi logo, topswitch , tinyswitch and ecosmart are registered trademarks of power integrations, inc. ?copyright 2001, power integrations, inc. japan power integrations, k.k. keihin-tatemono 1st bldg. 12-20 shin-yokohama 2-chome kohoku-ku, yokohama-shi kanagawa 222-0033, japan phone: +81-45-471-1021 fax: +81-45-471-3717 e-mail: japansales@powerint.com taiwan power integrations international holdings, inc. 17f-3, no. 510 chung hsiao e. rd., sec. 5, taipei, taiwan 110, r.o.c. phone: +886-2-2727-1221 fax: +886-2-2727-1223 e-mail: taiwansales@powerint.com europe & africa power integrations (europe) ltd. centennial court easthampstead road bracknell berkshire, rg12 1yq united kingdom phone: +44-1344-462-300 fax: +44-1344-311-732 e-mail: eurosales@powerint.com china power integrations international holdings, inc. rm# 1705, bao hua bldg. 1016 hua qiang bei lu shenzhen, guangdong 518031 china phone: +86-755-367-5143 fax: +86-755-377-9610 e-mail: chinasales@powerint.com india (technical support) innovatech #1, 8th main road vasanthnagar bangalore, india 560052 phone: +91-80-226-6023 fax: +91-80-228-9727 e-mail: indiasales@powerint.com applications hotline world wide +1-408-414-9660 applications fax world wide +1-408-414-9760


▲Up To Search▲   

 
Price & Availability of AN-22

All Rights Reserved © IC-ON-LINE 2003 - 2022  

[Add Bookmark] [Contact Us] [Link exchange] [Privacy policy]
Mirror Sites :  [www.datasheet.hk]   [www.maxim4u.com]  [www.ic-on-line.cn] [www.ic-on-line.com] [www.ic-on-line.net] [www.alldatasheet.com.cn] [www.gdcy.com]  [www.gdcy.net]


 . . . . .
  We use cookies to deliver the best possible web experience and assist with our advertising efforts. By continuing to use this site, you consent to the use of cookies. For more information on cookies, please take a look at our Privacy Policy. X