Structured Design
of Switching Power
Transformers
Transformer Design Procedure
Design of switching power transformers can be accomplished in a relatively simple manner by limiting magnetic
configurations to a few core and coilform structures.
These structures have been chosen both for their versatility and their low cost. Dimensional information as well as
design information in the form of design curves for the
chosen structures may be found at the end of this document. By using these curves, the complete transformer
can be designed.
Step 3. Secondary turn count
Secondary turn count is a function of duty cycle and
primary turn count.
For a flyback system:
For a forward converter:
Step 1. Structure size
The first step in the design is choosing a minimum structure size consistent with the output power required. The
approximate power capabilities of each structure are
provided in Table1. If five or six outputs are required,
a larger structure may be required to allow the copper
along with insulation and winding crossovers to fit in the
available winding area.
Step 2. Primary turn count
For a given core size, the ability of an inductor to operate without saturating is directly proportional to its turn
count NP. The normal saturation specification is ET or
volt-time rating. The E-T rating is the maximum voltage,
E, which can be applied over a time of T seconds. (The
E-T rating is identical to the product of inductance L and
peak current I.) Equation 1 defines a minimum value of
NP for a volt-time product of E-T:
mine NP. This value for NP should allow non-saturating
operation to 100C with reasonable core loss.
(1)
Where:
E-T = the minimum volt-time rating in volt-seconds
B = the maximum allowable flux density
AE = the effective cross sectional core
Equation 1 is plotted for the specific chosen core structures shown in Figure1. These plots are for B = 3000
Gauss, which will prevent the core from saturation and
typically will provide low core loss suitable for operation in
the range of 200kHz to 400kHz. For higher frequencies,
a higher primary turn count should be used to ensure low
core loss. To use this chart, locate the required E-T rating
on the vertical axis. Move horizontally to the curve. From
this point drop vertically to the horizontal axis and deter-
(2)
(3)
Where:
NP = the primary turn count.
NS = the secondary turn count.
VS = the secondary output voltage.
VD = the voltage drop across the rectifier and
choke in the secondary.
D = the duty cycle.
VP = the voltage across the primary.
For the flyback system, D is seldom greater than 0.5. For
the forward converter, D is the duty cycle of the rectified
output, and can approach 0.9 for a wave rectified output.
Known conditions should be used to calculate NS. For
example, at minimum input voltage and maximum output
power, the supply will operate at maximum duty cycle.
This is a good point to use to determine NS.
Step 4. Wire size
Once all the turn counts have been determined, wire size
must be chosen for each winding.
Power losses in the transformer windings cause a temperature rise, T, in the transformer. The amount of loss
depends on how much current is being drawn from the
winding, the length of wire and what wire size is used. The
power loss is a function of the amount of resistance in
the wire. This resistance is composed of a DC resistance
(RDC) and an AC resistance (RAC). At low frequencies
and small wire sizes, for example #30 AWG at 250 kHz,
RDC>>RAC, and RAC can effectively be ignored. For larger
wire sizes and high frequencies, >500 kHz, it may be
necessary to use stranded wire or foil. Lets assume RAC
Document 627-1 Revised 09/13/07
is small and can be ignored. Then all of the copper losses
will be due to RDC. This power loss should be distributed
among the primary and secondary windings to minimize
hot spots in the transformer. The temperature rise in a
given structure is a function of thermal resistance and
surface area.
In order to calculate the wire size, it is necessary to assign an acceptable power loss PL to each winding. The
power loss is then used to calculate the required DCR.
For flyback transformers:
(4)
For the forward converter:
(5)
Where:
PL = the maximum allowable power loss in the
winding.
I = the peak current in the winding.
D1 = the duty cycle of the current ramp
IAV = the average current in the winding
To determine winding area, use the curves in Figure2.
From this chart the number of turns per layer nl can be
read directly. Next find the number of layers, l, by dividing turns N by nl.
(6)
Next, find the area per layer, al. Multiply the winding width
as shown on the individual coilform drawing (Figures 4
and 5) by the wire diameter. Multiply this value by the
number of layers to find the area occupied by the winding.
(7)
Repeat this procedure for each winding and add to find
the total area occupied by the windings.
(8)
The total winding area, at, added to the area of insulation and shielding if required, ai, must be less than the
allowed structure area, aw.
aw at + ai (9)
Step 6. Inductance
Once RDC for all the windings has been calculated, the
wire size can be determined using the average length
per turn (Table1) for each coil form. Divide the RDC by
the product of the average length per turn times the turn
count. This value of RDC per foot can be found listed in any
wire manufacturers data sheet. From their data, choose
a wire size that has the same or slightly less RDC per foot.
For a forward converter it is generally preferable to have
as much inductance as possible. For a flyback, however it
is necessary to calculate the desired primary inductance
based on the amount of energy stored per cycle (1/2 LI2)
required to deliver an average output power Po.
Step 5. Will it fit?
Where:
L = inductance
= transformer efficiency (assume >0.9)
V = minimum DC input voltage applied to the
primary
ton = maximum on time for input voltage
f = switching frequency
Po = output power
After the wire sizes have been determined, it is necessary
to check fit, to see if the available winding area will accommodate the copper calculated in the previous steps.
The available winding area for the structure chosen is
shown in Table1. The area each winding occupies must
be determined, then the areas occupied by each winding
are added. To this total, an area allowing for interwinding insulation should be added. Assume about 10% of
the available area for insulation. The grand total is then
compared to the available winding area. If the area of the
windings is greater than the allowed structure area, either
wire size must be reduced, or a larger structure must
be chosen. Of course, a reduction in wire size results in
increased losses in the transformer.
A fit of 100% of the available space should not be used.
To accommodate winding crossovers, interwinding insulation and spaces between turns, it is recommended
that a fit factor no greater than about 80% be used. For
transformers with multiple outputs this factor may need
to be reduced further.
L=
(V ton )2 f
2Po
(10)
Step 7. Air gap
Now that the structure size and winding turn counts have
been found, the air gap must be determined. For most
forward converters, inductance should be as large as
possible and no air gap is required. However for flyback
transformers and chokes, an air gap is necessary. For
economic reasons and for the best reduction of EMI,
cores are usually gapped in the center leg. The amount
of gap can be found in Figure3. To use this chart it is necessary to convert inductance, L, to inductance factor, AL.
(11)
Document 627-2 Revised 09/13/07
Summary
Appendix Core data
The transformer should now be ready to wind. Sufficient
insulation must be used for the required voltage isolation. Windings must be uniform and carefully wound to
minimize leakage inductance. Some small adjustments
may be necessary to obtain the exact results desired,
but this procedure will provide a good first draft design.
EP and EFD series cores have proven to be both versatile and cost effective for transformer design of about 50
Watts or less. EP cores are suitable for lower power and
offer the smallest pcb area. EFD cores offer the lowest
profile (height). The recommended core sizes are shown
in Table 1.
Table 1. Core Data
EP7
EP10
EP13
EFD15
EFD17
EFD20
EFD25
Power capacity @ 100 kHz
10 W
12 W
20 W
20 W
25 W
30 W
50 W
ae (effective cross-sectional area) cm2
0.10
0.11
0.20
0.14
0.21
0.31
0.59
le (mean magnetic path length) cm
1.57
1.92
2.47
3.29
3.88
4.61
5.65
aw (bobbin winding area) cm2
0.045
0.122
0.141
0.173
0.198
0.286
0.4175
13.2 10.9
15.2 12.7
17.8 13.5
22.0 17.2
24.1 17.4
30.0 20.6
32.7 26.8
Maximum height mm
9.0
11.0
12.3
8.5
10.0
11.4
14.0
Average length per turn cm
1.79
2.15
2.38
2.60
3.15
3.90
4.64
Required board space mm
EP core styles
EFD core styles
Figure 2. Turns per layer vs wire size
Figure 1. Volt time product vs primary turns
Figure 3. Inductance factor vs air gap
Document 627-3 Revised 09/13/07
EP13
EP10
EP7
12.32 max
11.0 max
9.02 max
12.5
10.6
12.7
1.48
15.4
14.6
2.30
0.7 typ
10.46
17.0
2.10
0.6 typ
11.5
0.7 typ
12.8
2.50 typ
2.50 typ
2.50
3.60
5.8
Dimensions are in mm
7.6
Figure 4. EP core dimensions
EFD17
EFD15
10.0 max
8.01 max
17.6
20.3
21.25
23.3
0.7 typ
2.13
0.7 typ
2.20
17.0
17.0
2.50 typ
2.50 typ
9.0
11.0
EFD25
EFD20
14.0 max
11.43 max
24.8
29.7
28.5
31.5
2.45
0.7 typ
2.85
1.1 typ
20.0
3.00 typ
25.0
4.00 typ
Dimensions are in mm
13.64
Figure 5. EFD core dimensions
Document 627-4 Revised 09/13/07
