CMYK

DS ES N DE I IGGN

HIGH-FREQUENCY POWER TRANSFORMERS

Design a high-frequency power transformer based on flyback topology
DINESH KUMAR

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. Lets 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. Lets 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 12V2A=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

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. 2: Primary and secondary currents in (a) discontinuous and (b) continuous mode

50-70W

70-100W

windings simultaneously. Thus the so-called flyback transformer is not a transformer but a coupled inductor.

Discontinuous and continuous modes of operation

A flyback converter has two different modes of operation: discontinuous mode and continuous mode. Both modes require the same circuit; the waveforms of primary and secondary currents through the transformer are shown in Fig. 2. A circuit designed for the discontinuous mode will move into the continuous mode when the output current is increased beyond a certain value. 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,
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. In the continuous mode there is still some energy left in the secondary at the beginning of the next cycle. It is possible for flyback to operate in both modes, but it has different characteristics. The discontinuous mode has higher peak currents and therefore higher output voltage spikes during the turn-off. On the other hand, it has faster load transient response and lower primary inductance, and therefore the transformer can be made smaller in size. The reverse recovery time of the output diode is not critical because the forward current is zero before the reverse voltage is applied. Conducted EMI noise is reduced in discontinuous mode because the transistor turns on with zero collector current. The continuous mode, even if it has lower peak currents, and therefore lower output voltage spikes, is seldom used for low-power applications. 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.

the fewest components. At lower power levels, the total component cost is less than with other techniques. However, between 75 and 150W, increasing voltage and current stresses cause flyback-component cost to increase significantly. At these power levels, topologies with lower voltage- and current-stress levels (such as the forward converter) are more costeffective, even with higher component counts. To design a flyback transformer, you need to go through the following steps: Step 1. 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, VACmin and VACmax (f) Maximum duty cycle (Dmax); recommended maximum is 0.5 (g) Estimated power supply efficiency () at 0.75-0.85 Step 2. Primary inductance calculation: Primary inductance Lp = VDC minDmax Ippfs .(1)

Flyback transformer design

The flyback topology is used extensively because flyback power supplies require

where VDC min=2VAC min and Ipp is the peak primary current.

CMYK

DESIGN
TABLE II EE and EF Core Specifications Bmax = NpIpALG Ae

tesla or webber/m2

Type A EE10 EE10A EE13 EE13A EE12.6 EE16A EE16B EF16 EE19A EE19B EF20 EE20 EE20A EE20/20 EE23 EE25A EE25B EE25C 10.2 0.3 10.2 0.3 13.0 0.4 13.0 0.4 12.7 0.4 16.0 0.4 16.0 0.4 16.0 +0.7 0.5 19.0 0.4 19.0 0.4 20.0 0.6 20.5 0.5 20.3 0.5 20.1 0.4 22.7 0.3 25.0 0.5 25.0 0.5 25.4 0.6 B 5.5 0.2 5.5 0.2 6.0 +0.2 0.1 6.5 0.2 6.4 0.2 7.3 0.3 12.4 0.3 8.2 +0 0.3 8.0 0.3 13.6 0.3 9.3 +0 0.4 10.7 0.3 8.4 0.3 10.15 0.25 11.0 0.3 9.9 0.3 9.8 0.2 9.6 +0.3 0.1

Dimensions (mm) C 5.0 +0 0.3 9.8 +0.3 0.1 5.9 0.3 9.8 +0.3 0.1 3.3 0.2 5.1 +0 0.4 5.1 +0 0.4 4.7 +0 0.4 5.9 +0.1 0.5 5.1 +0 0.5 5.9 +0.4 0.5 7.0 0.3 4.8 0.2 9.2 0.25 10.2 0.2 6.35 0.25 6.1 0.3 6.75 0.2 D 2.50.2 2.50.2 2.8 +0 0.4 3.6 0.2 3.7 0.2 4.0 0.2 4.0 0.2 4.7 +0 0.3 5.0 +0 0.5 5.1 +0 0.5 5.9 +0.4 0 5.0 0.3 4.8 0.2 4.5 0.2 7.5 +0 0.3 6.35 0.2 5.8 0.2 6.35 0.25 E min 7.5 7.6 9.8 8.5 8.8 11.7 12.4 11.3 13.8 14.0 14.1 14.7 15.3 15.2 16.4 +0.5 0 18.6 18.6 18.6 F 4.30.2 4.30.2 4.6 +0.3 0.1 4.6 0.2 4.65 0.3 5.2 +0.2 0 10.4 0.2 5.7 +0.4 0 5.6 +0.4 0.1 11.3 0.3 6.1 +0.4 0 7.0 0.3 6.2 0.2 7.9 0.2 7.5 +0 0.3 6.9 0.3 6.8 0.2 6.5 +0.3 0.1

The calculated Bmax should be 0.2 to 0.3 tesla. If you get the flux density more than 0.3 tesla, go back to turns per volt (Te). Slightly increase Te to get higher values of Ns and Np and a lower value of ALG. If still you get Bmax more than 0.3 tesla, again increase Te and repeat the process until you get Bmax less than 0.3 tesla. Now calculate the required air gap, which means you first need to calculate the relative permeability of the ungapped core (r). This is calculated from core parameters Ae (effective cross-sectional area in cm2), Le (effective magnetic path length in cm2) and AL (inductance factor in nH/turn2) as follows: r = ALLe 0.4Ae

The gap length (Lg) can now be calculated. The gap should be ground only in the centre leg of the core. If the gap is put into the outer legs, it will need to be half that calculated here. The minimum limit for Lg is 0.051 mm, and Lg is calculated from the following equation: Lg= Ae Le ( 0.4N )10 Lp
2 p r 3

mm

Po Average primary current IAV = VDC min .(2) 2 Ipp= IAV Dmax = 2Po VDC minDmax .(3)

Now let us define turns per volt (Te) to decide Np and Ns. If some conditions (explained later) are not satisfied, well have to come back to modify Te. Ns=TeVo (5) Now from Eq. 4, we can calculate Np. Step 4. Calculation of the required core size and core air-gap: Table I explains the power handling capacity of various ferrite cores. After selecting the appropriate core, refer to the manufacturers datasheet (see Table II) to know the required parameters of the core such as AL, Ae and Le. ALG= Lp N2p H/turn2

Step 3. Calculation of the number of turns in primary, secondary and biasing windings: The numbers of secondary turns (Np) is calculated as follows: Dmax VDC min ..(4) Np =Ns Vo+VD 1Dmax where VD is the forward voltage drop of the output diode (D1) as shown in Fig. 2(a) and Ns is the number of secondary turns.

Step 5. Selection of wire area for primary and secondary windings: For primary and secondary, choose a wire that doesnt generate too much heat in the winding at the desired current. For that, use the current density (J, in amp/ mm2) to calculate the area of the conductor. The accepted value of J is 3A/mm2 to 6A/mm2. A good value of J is 4.5A/mm2 as this gives a smaller wire size without undesirable temperature rise in the winding and the core. The area of primary winding conductor: Input rms current (Irms) J

Ap =

mm2

Similarly, 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

DESIGN
TABLE III EE and EF Core Specifications
Type Ct (mm-1) EE10 EE10A EE13 EE13A EE12.6 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 2.1 1.1 1.7 0.8 2.4 1.92 2.8 1.87 1.68 2.65 1.2 1.21 1.9 1.2 0.75 1.25 1.3 1.3 1.15 1.340 1.20 1.2 0.6 0.95 0.57 0.82 0.8 0.5 1.1 0.92 0.89 0.57 0.804 0.83 0.6 0.50 0.50 0.7 0.50 0.4 0.6 0.35 0.274 Ae (mm2) 12.7 23.1 17.2 37.4 12.4 18.4 19.6 20.1 23.3 23.4 35.6 39 22 41.1 70.0 39.6 36.9 36.9 43.4 33.5 40.7 49.2 75.8 71.6 99.8 87.4 83.1 108.6 59.4 70.1 83 124.0 86.2 88.3 127 153 152.0 147.4 184.3 235 140.7 354 540 Le (mm) 26.1 26.4 30.0 29.6 29.7 35.5 55.2 37.6 39.2 62.1 42.2 47.1 42.8 49.7 49.6 49.5 49.7 49.7 48.2 44.9 49.2 57.9 46 63.0 48.1 73.1 64.4 57.7 65.5 61.3 74 66.1 69.3 69.4 77.0 76.5 77.1 96.8 96.8 97.4 84.6 123 147 Ve (mm2) 330.7 611.6 513.8 1107.6 369.0 655 1080 750 914 1450 1505 18.50 942.8 2042 3469 1963 1823 1823 2091 1500 2004 2850 3488.1 4508 4801 6391 6350.5 6269.7 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) 1.8 1.5 3.6 3.5 2.0 3.3 5.3 4.0 4.6 7.5 7.8 9.7 3 5 17.5 10.2 9.0 9.0 10.0 14.8 11 13.5 10 35 23.5 33.5 26 20 22 21.5 30 40.6 31 31.5 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) 10.16 9.45 8.84 8.23 7.62 7.01 6.40 5.89 5.38 4.88 4.47 4.06 3.66 3.25 2.95 2.64 2.34 2.03 1.83 1.63 1.42 1.22 1.02 0.92 0.81 0.71 0.61 0.56 0.51 0.46 0.41 0.38 0.35 0.31 0.29 SWG area (mm) 81.032 70.102 61.344 53.170 45.581 38.575 32.154 27.233 22.721 18.694 15.685 12.940 10.516 8.292 6.831 5.471 4.298 3.235 2.629 2.086 1.583 1.168 0.817 0.664 0.515 0.396 0.292 0.246 0.204 0.166 0.132 0.113 0.096 0.073 0.066

4500 2990 3680 4250 4600 4704 4200 5400 6500

and the secondary (in mm2), select the wire gauge from Table IV.

Switching frequency selection

The operating frequency of the power supply should be selected to obtain the best balance between switching losses, total transformer losses, size and cost of magnetic components and output capacitors. A high switching frequency reduces the output capacitor value and the inducELECTRONICS FOR YOU MARCH 2004

tance of the primary and secondary windings, and therefore the total size of the transformer. But it also increases transformer losses and switching losses of the switch. High losses reduce the overall efficiency of the power supply and increase the size of the heat sink required to dissipate the heat.

flyback transformers because of its low cost and easy availability. Other types such as EF, EFD, ETD, EER and EI can also be used depending on particular requirements such as height restrictions. RM, toroid and Pot cores are not suitable because of the safety isolation required. EFD are good for low-profile, ETD are good for high-power and EER are good for multiple-output designs.

An example
For a better understanding of the entire power transformer design, heres an example. Step 1. 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, of which E-core is commonly used in

CMYK

DESIGN

Flyback Transformer Application in An SMPS

CD players need 3V to 12V DC to play music and mobile phones need around 4.5V DC to recharge their battery. But we plug the CD player or the cellphones battery charger into 220V AC, 50Hz through an adaptor (which generally comes with CD player or the cellphone). Now the question is how this 220V AC is converted into 3V or 4.5V DC? Or what is there inside the CD player adaptor or cellphone charger to do this job? Devices like CD players, cellphone battery chargers, fax machine power supplies, TV power supplies, stereo power supplies, PC power supplies, electronic toy power supplies and PDA adaptors contain nothing but the SMPS shown in Fig. 3. The figure shows the application of a flyback transformer in an SMPS. The equipment take 220V, 50Hz AC power from the wall-mounted power socket in our home or office. This 220V AC is rectified and filtered into DC voltage without an input isolation transformer. The rectified and filtered DC voltage (in the range of 260V DC to 360V DC) is given to the flyback transformer, where the rectified high-voltage DC is stepped down to 3.3V, 4.5V DC or some other value to feed the device. This is done

Fig. 3: SMPS block diagram

with the help of a power-switch control pulse-width modulation (PWM) circuit. When you connect the device to the output DC voltage from the flyback transformer after output rectifier, this voltage may vary from the nominal value because of fluctuations in the input AC voltage or overloading. To maintain the output voltage at the desired level, a sampling network is required. The sampling network gives a sample of the output voltage to check with the reference voltage. For any difference between the sampled voltage and the reference voltage, the PWM circuit takes action to maintain the output voltage constant. In the block diagram of the SMPS, the power transformer need not be a flyback transformer only. It could be a forward transformer or a half-/ full-bridge transformer. The selection of the topology (type of the transformer) depends on the output power level.

(d) Bias voltage Vb=18V (generally, 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=0.5 (h) Estimated power supply efficiency =0.85 (i) Switching frequency fs=40 kHz Step 2. Primary inductance calculation: IAV = Po VDC min

Step 3. Calculation of the number of turns in primary and secondary windings: Ns=TeVo Taking Te= 1 turn/volt, we get Ns=1112=12 turns Now primary turns (Np) are calculated as follows: Dmax Np= Ns Vo+VD 1Dmax =12 0.45 =90.62 120 12+1 10.45 VAC min

Table II shows that the effective crosssectional area for the selected core (EE25A) is 39.6 mm2. Now calculate the maximum flux density as follows: Bmax= = NpIpALG Ae

901.1281.48107 39.6106

=0.379 tesla This value of Bmax will saturate the core, hence we have to go back to increase Te and check what happens. If Te= 1.25, we get: Ns=15 Np=113 ALG=9.397108 H/turn2 Now Bmax= 0.302 tesla As it again shows that the core is operating on the edge of saturation, we calculate Bmax once again with an increased value of Te. If we assume Te=1.5, we get: Ns=23 Np=173 ALG=4108 H/turn2 Bmax= 0.197 tesla Now, the flux density has become very
MARCH 2004 ELECTRONICS FOR YOU

(where VDC min=285=120V DC) = 26 =0.254A 0.85120 2 Dmax

Ipp=IAV

2 =1.128A 0.45 VDC minDmax Primary inductance Lp= Ippfs =0.254 1200.45 = =1.2103 H 1.12840103

Since turns take a round figure, lets consider Np=90. Step 4. Calculation of the required core size and core air-gap: From Table I, we find that an appropriate core for this design is EE25A. From Tables II and III, core specifications are: Ae=39.6 mm2 AL= 1900 nH/turn2 Le= 49.5 mm Lp ALG = 2 H/turn2 Np = 1.2103 =1.48107 H/turn2 902

DESIGN
low, which means the core is underutilised. Lets calculate once more with Te =1.35. We get: Ns=17 Np=128 ALG=7.324108 H/turn2 Bmax= 0.26 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): r = = ALLe 0.4Ae 190010949.5103 0.439.6106 = 0.436 =0.0970 mm2 4.5

The flyback topology is used extensively because flyback power supplies require the fewest components. At lower power levels, the total component cost is less than with other techniques
103 mm =(679.426.20)103=0.653 mm

The rms value of the primary current (Irms) is given by: Ipp

Dmax 3

As=

2 =0.444 mm2 4.5

=1.889103 Air-gap length (Lg)= = Ae Le ( 0.4N )10 Lp

2 p r 3

mm

This gap will be in the centre of the core. The minimum limit for Lg is 0.051 mm. Step 5. Selection of wire area for primary and secondary windings: Area of primary winding conductor (Ap):

After calculating area (in mm2) of primary and secondary, select the wire gauge from Table IV. The SWG for primary conductor is 29 and the SWG for secondary conductor is 21 or 20. The design results can be summarised as follows: 1. Core type: EE25A 2. Primary number of turns (Np) with 28 SWG insulated copper wire=128 3. Secondary number of turns (Ns) with 21 SWG insulated copper wire=17 4. Air-gap in the EE core = 0.6 mm The author is a hardware engineer at MRO-TEK Ltd

0.4128239.6106 49.5103 3 1.210 1.889103

Input rms current (Irms) J

mm2

ELECTRONICS FOR YOU

MARCH 2004