CMYK

DS ES N DE I IGGN

HIGH-FREQUENCY POWER TRANSFORMERS
Design a high-frequency power transformer based on flyback topology
DINESH KUMAR

E

ssentially, switched-mode power supplies, or SMPS, act as AC-to-DC converters. These rectify the AC input voltage (85V-265V AC) to convert it into DC. Depending upon the design considerations, these chop the rectified voltage (DC) at very high frequencies. SMPS find use in computer power supplies, TV sets, CD players, battery chargers, adaptors, etc. Their major advantages are ligher weight, smaller size, higher efficiency and lesser cost. Let’s now consider the merits of SMPS individually. Lighter weight and smaller size are due to operation at a significantly higher frequency range and use of smaller inductive elements. Rapid switching of the power transistor between saturation and cut-off regions of its operation results in very little energy dissipation and hence reduced heat-sink requirements. Costs are reduced owing to the absence of large bulky power transformers, a huge reduction in volume and power dissipation, smaller material requirements and smaller semiconductor devices. SMPS have a complex circuitry. Making a traditional 12V DC power supply providing 3A current for a stereo cassette player is not a difficult job for electronic hobbyists with some experience. But designing an SMPS for the same application is quite painstaking. Power transformer is a crucial part of SMPS. Other components are controllers (PWM ICs), power switches, input/output rectifier and bulk capacitors.

Topology selection
The circuit topology (Fig. 1) has a great impact on the transformer design. Flyback circuits are used primarily at power levels of

0 to 150 watts, forward converters at 50 to 500 watts, half-bridge at 100 to 1000 watts, and fullbridge usually over 500 watts. Full-bridge and half-bridge topologies with full-bridge secondaries have the highest transformer efficiency because the core and the windings are fully utilised. Let’s assume that you need 12V DC output at 2A of current for your stereo cassette player from 220V AC, 50 Hz. Since the output power rating is 12V×2A=24W, the right topology for this design is the flyback. To fully understand a flyback power supply design, it is useful to review the theory of flyback topology and the general aspects of a switched-mode power supply such as continuous and discontinuous operation modes of a high-switching frequency transformer design. The power-transformer design is the biggest stumbling block in developing switched-mode power supplies. In Fig. 1(a) when power switch T1 is ‘on’ with the application of ‘on’ pulse from the control circuit (not shown in the figure), the current flows through the primary winding and energy stores within the core. Note that no current can flow through the secondary because of opposite dot polarity (and hence blocked diode D1). When power switch driving pulse from the control circuit is removed (during ‘off’ time), the polarity reverses and the current flows in the secondary winding. The current flows Fig. 1: Various topologies including (a) flyback topology, (b) forward topology, (c) half-bridge topolgy and (d) full- in either the primary or secondary winding but never in both bridge topology
MARCH 2004 ELECTRONICS FOR YOU

Thus the so-called flyback transformer is not a transformer but a coupled inductor. In the discontinuous mode all the energy stored in the primary during the power switch ‘on’ time is completely transferred to the secondary and to the load before the next cycle. is seldom used for low-power applications. A circuit designed for the discontinuous mode will move into the continuous mode when the output current is increased beyond a certain value. The reverse recovery time of the output diode is not critical because the forward current is zero before the reverse voltage is applied. At these power levels. the fewest components. The discontinuous mode has higher peak currents and therefore higher output voltage spikes during the turn-off. but it has different characteristics. . ELECTRONICS FOR YOU MARCH 2004 and there is also a dead time between the instant the secondary current reaches zero and the start of the next cycle. Primary inductance calculation: Primary inductance Lp = VDC min×Dmax Ipp×fs …. it has faster load transient response and lower primary inductance. VACmin and VACmax (f) Maximum duty cycle (Dmax). even if it has lower peak currents. To design a flyback transformer. the waveforms of primary and secondary currents through the transformer are shown in Fig. and therefore lower output voltage spikes. Define the power supply parameters pertaining to the transformer design: (a) Derive output power (Po) (b) Output voltage (Vo) (c) Bias voltage (Vb) (d) AC mains frequency (fL) (e) Minimum and maximum AC mains voltage. the total component cost is less than with other techniques. On the other hand. At lower power levels. even with higher component counts. and therefore the transformer can be made smaller in size. Conducted EMI noise is reduced in discontinuous mode because the transistor turns on with zero collector current. It is possible for flyback to operate in both modes. The continuous mode. between 75 and 150W. 2: Primary and secondary currents in (a) discontinuous and (b) continuous mode 50-70W 70-100W windings simultaneously.75-0.DESIGN TABLE I Core Size Selection on the Basis of Power Handling Capacity Output power level 0-10W Recommended core types EFD15 SEE16 EF16 EPC17 EE19 EF(D)20 EPC25 EF(D)25 EE19 EPC19 EF(D)20 EE or EI22 EF(D)25 EPC25 EI25 EF(D)25 EPC25 EPC30 EF(D)30 ETD29 EER28(L) EI28 EER28(L) ETD29 EF(D)30 EER35 EER28L ETD34 EER35 ETD39 ETD34 EER35 ETD39 EER40 E21 10-20W 20-30W 30-50W Fig. In the continuous mode there is still some energy left in the secondary at the beginning of the next cycle.and current-stress levels (such as the forward converter) are more costeffective. Higher voltage and current spikes are not desirable because these exert extra electrical stress on the output diode and the power switch connected in the primary of the transformer. However.5 (g) Estimated power supply efficiency (η) at 0. recommended maximum is 0. you need to go through the following steps: Step 1. topologies with lower voltage.(1) Flyback transformer design The flyback topology is used extensively because flyback power supplies require where VDC min=√2×VAC min and Ipp is the peak primary current.85 Step 2. 2. Discontinuous and continuous modes of operation A flyback converter has two different modes of operation: discontinuous mode and continuous mode. increasing voltage and current stresses cause flyback-component cost to increase significantly. Both modes require the same circuit.

3 tesla.0 ±0.0 ±0.4 22.4 5.4 3.8 ±0.2 ±0.5 ±0.3 –0.5 +0 –0. which means you first need to calculate the relative permeability of the ungapped core (µr).4 –0.4 –0 7.8 ±0.25 10.2 +0. If the gap is put into the outer legs.3 8.4 11.2 ±0.8 ±0.5 5.25 11.2 4.2 7.1 14.5 19.2 ±0. If still you get Bmax more than 0. ALG= Lp N2p H/turn2 Step 3.3 –0.3 6.2 9. Le (effective magnetic path length in cm2) and AL (inductance factor in nH/turn2) as follows: µr = AL×Le 0.0 ±0. refer to the manufacturer’s datasheet (see Table II) to know the required parameters of the core such as AL.4 ±0.7 ±0.2 2.2 4.7 ±0.2 5.0 +0 –0.5 20.3 tesla.1 +0.2 7.1 +0 –0.7 +0.3 6.8 +0.6 ±0.4 –0 5.1 The calculated Bmax should be 0.3 9.35 ±0.4 ±0.2 –0 10. If some conditions (explained later) are not satisfied.7 15.4 ±0.3 10.5 5.0 ±0.2 ±0.4 4.(2) 2 Ipp= IAV× Dmax = 2Po VDC min×η×Dmax ….25 E min 7.3 9.4 5.2 6. it will need to be half that calculated here.1 +0 –0. 2(a) and Ns is the number of secondary turns.6 ±0.6 18.5 ±0.4 –0. choose a wire that doesn’t generate too much heat in the winding at the desired current. use the current density (J.1 4.5 8.4 +0.8 ±0.0 +0.2 6.3 8.2 D 2.5 ±0.7 –0.7 12.(4) Np =Ns× × Vo+VD 1–Dmax where VD is the forward voltage drop of the output diode (D1) as shown in Fig.4 16. If you get the flux density more than 0. Calculation of the number of turns in primary.3 13.2 4. Calculation of the required core size and core air-gap: Table I explains the power handling capacity of various ferrite cores.4π×Ae The gap length (Lg) can now be calculated.2 7.4 20.3 ±0.3 –0.8 ±0.9 ±0.4 10.3 8.5 +0 –0.0 ±0.3 4.3 –0.2 16.1 3. we’ll have to come back to modify Te.3 10.7 +0 –0. Now calculate the required air gap.65 ±0.6 B 5.4π×N )×10 Lp µ 2 p r –3 mm Po Average primary current IAV = η×VDC min ….1 6. Ae and Le.1 ±0.6 +0.2 4.9 +0.2 4. and Lg is calculated from the following equation: Lg= ×Ae Le – ( 0.5 25.3 5.0 ±0.0 ±0.2 5.2 5.3 +0 –0.2 6.9 +0.5 5.0 ±0.7 ±0. The area of primary winding conductor: Input rms current (Irms) J Ap = mm2 Similarly.5 7.0 ±0.4 12.6 +0.1 Dimensions (mm) C 5.0 ±0.5 5.5A/mm2 as this gives a smaller wire size without undesirable temperature rise in the winding and the core.3 tesla.3 12.2 4.1 ±0.051 mm.6 18.6 ±0.5 7.3 6.3 –0.6 20.0 ±0.3±0. Slightly increase Te to get higher values of Ns and Np and a lower value of ALG.2 2. Selection of wire area for primary and secondary windings: For primary and secondary.6 9.8 11.4 13.CMYK DESIGN TABLE II EE and EF Core Specifications Bmax = Np×Ip×ALG Ae tesla or webber/m2 Type A EE10 EE10A EE13 EE13A EE12..3 25.0 ±0.1 –0.3 ±0.1 5.0 ±0. Ns=Te×Vo …………(5) Now from Eq.25 6. 4.8 14.7 +0 –0.3 6.2 to 0. secondary and biasing windings: The numbers of secondary turns (Np) is calculated as follows: Dmax VDC min ……….3 9. After selecting the appropriate core.0 +0.4 ±0.5 ±0.7 ±0.0 ±0.0 ±0.3 13.1 11.75 ±0. Step 4.5±0.3 ±0.3 ±0.2 +0 –0.3 tesla. The minimum limit for Lg is 0.0 ±0.2 7.2 4.4 19.3 13.35 ±0.9 +0.2 9.3 9.3±0.4 16.8 +0 –0.35 ±0. we can calculate Np.3 5. Step 5. For that.2 5.5 –0 18.3 6. in amp/ mm2) to calculate the area of the conductor.4 16.9 ±0.8 8.3 15.2 –0.5 ±0.5 20.4 –0 5.3 4.2 6.9 ±0.9 ±0. The gap should be ground only in the centre leg of the core.2 6.6 +0.5±0.2 3.15 ±0.2 ±0.0 +0 –0. This is calculated from core parameters Ae (effective cross-sectional area in cm2). go back to turns per volt (Te).1 +0 –0.8 +0.3 9.5 25. the area of secondary winding conductor (As): Output current (Io) J Now calculate the maximum flux density Bmax using the effective cross-sectional area for the selected core: As= mm2 After calculating area of the primary MARCH 2004 ELECTRONICS FOR YOU . The accepted value of J is 3A/mm2 to 6A/mm2.6 EE16A EE16B EF16 EE19A EE19B EF20 EE20 EE20A EE20/20 EE23 EE25A EE25B EE25C 10. again increase Te and repeat the process until you get Bmax less than 0. A good value of J is 4.6 F 4.5 +0.0 14.1 +0 –0.(3) Now let us define turns per volt (Te) to decide Np and Ns.4 ±0.3 6.

274 Ae (mm2) 12.804 0.1 23.1 70.396 0.4 83.515 0.1 26.0 655 1080 750 914 1450 1505 18.132 0.102 61.3 1.516 8.0 39.9 43.629 2.50 0.0 14.8 0. of which E-core is commonly used in .1 69.4 35.3 5.6 29.16 9.9 46 63.7 354 540 Le (mm) 26.086 1.95 0.92 2.721 18.8 96.2 0.1 17.81 0.03 1.340 1.6 0.61 0.7 0.2 1.073 0.4 77.4 184.23 7.8 97.9 1.38 0. ETD.1 108.2 0.9 49.0 4.4 0.168 0.831 5.6 59.3 1.5 49.066 4500 2990 3680 4250 4600 4704 4200 5400 6500 and the secondary (in mm2).5 10.4 70.6 39 22 41.7 3 5 17.6 3. High losses reduce the overall efficiency of the power supply and increase the size of the heat sink required to dissipate the heat.471 4. EFD.1 42. But it also increases transformer losses and switching losses of the switch.46 0.02 0.6 49.8 2042 3469 1963 1823 1823 2091 1500 2004 2850 3488.5 3.0 76.7 49.8 87. ETD are good for high-power and EER are good for multiple-output designs.34 2.40 5.2 62.344 53.4 33.51 0.8 11 13.95 2.5 10 35 23.6 99. Step 1.06 3.6 0.4 12.4 57. size and cost of magnetic components and output capacitors.5 55 60 61 72 90 120 62 216 410 Wire gauge size OOOO OOO OO O 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 TABLE IV Wire Parameters SWG diameter (mm) 10.83 0.7 23.581 38.3 69.3 4.57 0.8 2.583 1.032 70.2 44.694 15.5 33.50 0.298 3.2 9.20 1.89 5.1 0.7 611.8 49.3 23. Power supply specifications: (a) Output voltage (Vo)=12V (b) Output current (Io)=2A (c) Total output power Po= (Vo+VD)×Io=(12+1)×2=26W Core selection Ferrite cores are available in many shapes.6 123 147 Ve (mm2) 330.154 27. RM.29 SWG area (mm²) 81.50 0.75 1.57 0.83 1. An example For a better understanding of the entire power transformer design.1 73.940 10.166 0.8 1.DESIGN TABLE III EE and EF Core Specifications Type Ct (mm-1) EE10 EE10A EE13 EE13A EE12.68 2.1 4508 4801 6391 6350.47 4.38 4.0 86.6 39.5 55.7 0.25 1.5 6269.233 22.56 0.35 0.1 1.096 0.6 36.25 2.246 0.1 1.2 37.0 147. EFD are good for low-profile.2 57.1 96.7 49.64 2.50 942.84 8.7 65.21 1.6 369.8 1107.170 45. and therefore the total size of the transformer.5 77.6 EE16A EE16B EF16 EE19A EE19B EF20 EE20 EE20A EE20/20 EE23 EE25A EE25B EE25B/20 EE25C EF25 EE25D EE25E EE26 EE28A EE28B EE28C EE28Q EE30 EE30/30 EF32A EF32B EE33 EE35A EE35B EE40A EE40B EE41 EE42A EE42B EE42C EE46 EE55 EE65 2.4 84.7 48.5 40.92 0.292 6.22 1.2 88.01 6.2 47.7 3879 4295 6140 8192 3973 6121 9922 11710 11722 14275 17843 22889 11583 43542 79000 SK ≥700 ≥700 ≥860 ≥860 ≥650 ≥950 ≥680 1000 1050 1000 1800 2100 1100 2100 ≥1600 ≥1400 ≥1400 ≥1600 ≥1630 ≥1600 ≥1600 ≥1600 3400 4500 3200 3200 2200 2200 AL(nH/N2)±25% SP3 810 810 1000 1000 900 1100 750 800 1200 840 1550 1900 1900 2000 2300 1900 1600 1600 1800 1900 1800 1850 1850 3060 4000 2806 2806 1800 1900 2500 2400 4000 2600 3200 3700 4000 4200 3700 4700 5700 3700 6700 8000 SP4 810 1000 1000 1500 800 1100 750 800 1200 840 1450 1800 1800 1800 2300 1900 1600 1600 1800 1900 1800 1850 1700 2850 4000 2650 3500 1900 1900 2500 2400 4000 2600 3000 3700 4000 4200 3600 4700 5700 3700 6700 8000 Weight (gm) 1.6 31 31.4 18.5 2.45 8.31 0.2 37.8 9.6 0.71 0.6 7.8 1.63 1.235 2.664 0.62 7.9 36. flyback transformers because of its low cost and easy availability.6 513.817 0. Other types such as EF. select the wire gauge from Table IV.65 1.87 1.292 0.88 4.5 7.4 30.3 127 153 152.0 9.5 26 20 22 21.0 48.3 74 66.5 1.575 32. here’s an example.6 20.15 1. total transformer losses.1 64.0 10. A high switching frequency reduces the output capacitor value and the inducELECTRONICS FOR YOU MARCH 2004 tance of the primary and secondary windings.0 29.89 0.8 71.1 83 124.35 0.41 0.1 42. EER and EI can also be used depending on particular requirements such as height restrictions.42 1.7 49.7 35.113 0.0 3.4 1.92 0.204 0.82 0.4 19. Switching frequency selection The operating frequency of the power supply should be selected to obtain the best balance between switching losses.66 3. toroid and Pot cores are not suitable because of the safety isolation required.3 235 140.685 12.5 61.5 30 40.2 75.

6×10–6 =0.48×10–7 H/turn2 902 . Calculation of the required core size and core air-gap: From Table I. Now calculate the maximum flux density as follows: Bmax= = Np×Ip×ALG Ae 90×1.2×10–3 H 1.397×10–8 H/turn2 Now Bmax= 0. Step 4. If Te= 1. 16V to 20V) (e) AC mains frequency fL=50 Hz (f) Minimum AC mains voltage VACmin=85V and maximum AC mains voltage VACmax=265V (g) Maximum duty cycle Dm=0. the flux density has become very MARCH 2004 ELECTRONICS FOR YOU (where VDC min=√2×85=120V DC) = 26 =0.85 (i) Switching frequency fs=40 kHz Step 2. The selection of the topology (type of the transformer) depends on the output power level. TV power supplies. let’s consider Np=90. The sampling network gives a sample of the output voltage to check with the reference voltage.5 mm Lp ALG = 2 H/turn2 Np = 1.45 =90. If we assume Te=1. Calculation of the number of turns in primary and secondary windings: Ns=Te×Vo Taking Te= 1 turn/volt. For any difference between the sampled voltage and the reference voltage. The figure shows the application of a flyback transformer in an SMPS. fax machine power supplies.5V DC or some other value to feed the device.6 mm2.CMYK DESIGN Flyback Transformer Application in An SMPS CD players need 3V to 12V DC to play music and mobile phones need around 4.45 VAC min Table II shows that the effective crosssectional area for the selected core (EE25A) is 39. 50Hz AC power from the wall-mounted power socket in our home or office.197 tesla Now.48×10–7 39. we get: Ns=15 Np=113 ALG=9.62 120 × 12+1 1–0.379 tesla This value of Bmax will saturate the core. 4. Now the question is how this 220V AC is converted into 3V or 4. cellphone battery chargers.5V DC? Or what is there inside the CD player adaptor or cellphone charger to do this job? Devices like CD players. In the block diagram of the SMPS.2×10–3 =1. where the rectified high-voltage DC is stepped down to 3.302 tesla As it again shows that the core is operating on the edge of saturation. Primary inductance calculation: IAV = Po η×VDC min Step 3. we find that an appropriate core for this design is EE25A. (d) Bias voltage Vb=18V (generally.128×40×103 Since turns take a round figure. It could be a forward transformer or a half-/ full-bridge transformer.45 VDC min×Dmax Primary inductance Lp= Ipp×fs =0.25. The equipment take 220V.128×1. This is done Fig. But we plug the CD player or the cellphone’s battery charger into 220V AC. When you connect the device to the output DC voltage from the flyback transformer after output rectifier.128A 0.5. this voltage may vary from the nominal value because of fluctuations in the input AC voltage or overloading.85×120 2 Dmax Ipp=IAV× 2 =1. we get: Ns=23 Np=173 ALG=4×10–8 H/turn2 Bmax= 0. 3.5V DC to recharge their battery. From Tables II and III. the PWM circuit takes action to maintain the output voltage constant. This 220V AC is rectified and filtered into DC voltage without an input isolation transformer.45 = =1. the power transformer need not be a flyback transformer only. 3: SMPS block diagram with the help of a power-switch control pulse-width modulation (PWM) circuit.3V. PC power supplies. we get Ns=11×12=12 turns Now primary turns (Np) are calculated as follows: Dmax Np= Ns× × Vo+VD 1–Dmax =12× 0. hence we have to go back to increase Te and check what happens. core specifications are: Ae=39. The rectified and filtered DC voltage (in the range of 260V DC to 360V DC) is given to the flyback transformer.5 (h) Estimated power supply efficiency η=0. a sampling network is required. we calculate Bmax once again with an increased value of Te.6 mm2 AL= 1900 nH/turn2 Le= 49.254× 120×0.254A 0. 50Hz through an adaptor (which generally comes with CD player or the cellphone). stereo power supplies. To maintain the output voltage at the desired level. electronic toy power supplies and PDA adaptors contain nothing but the SMPS shown in Fig.

Step 5. which means the core is underutilised. Secondary number of turns (Ns) with 21 SWG insulated copper wire=17 4. Primary number of turns (Np) with 28 SWG insulated copper wire=128 3. Let’s calculate once more with Te =1.5 The flyback topology is used extensively because flyback power supplies require the fewest components.6×10–6 = 0. select the wire gauge from Table IV.444 mm2 4.20)×10–3=0.4–26.889×10–3 ) = Input rms current (Irms) J mm2 ELECTRONICS FOR YOU MARCH 2004 .6 mm ❑ The author is a hardware engineer at MRO-TEK Ltd 0. Selection of wire area for primary and secondary windings: Area of primary winding conductor (Ap): After calculating area (in mm2) of primary and secondary.26 tesla This value is within the acceptable limits. The relative permeability is calculated as follows (if it is not specified in the core datasheet by the core vendor): µr = = AL×Le 0.051 mm. the total component cost is less than with other techniques ×10–3 mm =(679.4π×N )×10 Lp µ 2 p r –3 mm = ( This gap will be in the centre of the core.5×10–3 – –3 1.0970 mm2 4. Air-gap in the EE core = 0.889×10–3 Air-gap length (Lg)= = ×Ae Le – ( 0.653 mm The rms value of the primary current (Irms) is given by: Ipp× √ Dmax 3 As= 2 =0. Core type: EE25A 2.6×10–6 49.5×10–3 0.35.4π×39.2×10 1. At lower power levels.324×10–8 H/turn2 Bmax= 0.5 =1. The SWG for primary conductor is 29 and the SWG for secondary conductor is 21 or 20. The minimum limit for Lg is 0.4π×1282×39.4π×Ae 1900×10–9×49. We get: Ns=17 Np=128 ALG=7. The design results can be summarised as follows: 1.436 =0.DESIGN low.