Vous êtes sur la page 1sur 5

Power-Inductors with Multi-layer Windings

Single layer toroids are inexpensive and popular but not the only option. For higher powers or inductors
requiring a large number of turns in a compact volume, core+bobbin assemblies are frequently used.
One advantage of segmented cores is that air-gaps can be incorporated easily, allowing the designer greater
control over the effective permeability.
Unlike the single-layer toroid in which the winding is constrained by a linear dimension (the inner diameter),
multiple-layer windings are constrained by an available winding area, called the window area
Core halves slide into plastic
sleeve or bobbin onto
which the wire is wound.

Shaded region is the


bobbin window area Ab
available for winding (a
function of both the core
and bobbin choices)
Manufacturers usually
specify the window area
and also the MLT for a
particular bobbin-core
combination.

We will develop two commonly-used approaches for core selection that can be easily extended to other
core types:
The area-product method (similar to the area-diameter product for single-layer toroids)
The core geometry method

Bob York

Back to TOC

Area-Product (Ap) Method


Most of the same constraints hold for all geometries. The only change is the available area for windings:

Ab

Since the wires are round we can never quite fill


the available area. Some insulating layers are
sometimes added between windings to reduce
capacitance or improve the voltage breakdown.
Thus it is common to write:

NAbw K w Ab
Window
utilization factor

Single bare
wire area

The wire size is again related to a maximum rms current density:

Different winding orientations and insulation


schemes affect the window utilization

I rms
J max
Abw

dbw

4I rms
J max

So the inductor design is governed by the following:


1) Inductance:

A
L = e N e = N 2 AL
le

2) B<Bmax

I max

NBmax Ae
L

NAbw
N I rms
=
Kw
K w J max

3) Window area:

Ab

4) Winding Loss

PCu = Cu

N ( MLT ) 2
I rms
Abw

The product of core cross sectional area and window area is the area product Ap:

LI max I rms
Ap Ae Ab
Bmax K w J max

This gives a quick method for choosing cores. Some


manufacturers specify the area product for various cores.
Typically the window utilization factor is assumed to be:

K w = 0.4 0.5

Note: look for the area product in the bobbin specification!

Bob York

Back to TOC

Gapped Cores
Once the core geometry is chosen, the required
inductance factor can be determined from (1) and (2):

AL

2
Bmax
Ae2

Ferroxcube E-cores

2
LI max

Usually the cores are made from a high permeability ferrite, and an air-gap is used
to engineer the desired effective permeability or inductance factor. From earlier
work we can get a quick estimate of the required gap:

AL

0
g

Ae

2
0 LI max

Gapped E-core set

2
Bmax
Ae

Some manufacturers sell some pregapped core sets with certain predetermined values of inductance factor.
Often one of those will work fine.

Core relative
permeability

r
L I 2 [J]

For customized gaps the manufacturer will often provide information like the
chart at right. This aids core selection as well. This data should account for
fringing flux around the gap which increases the effective cross section.
The effect can be approximated by assuming the fringing
increases the cross-sectional dimensions by the gap length:

Agap

Rgap

1 g
0 Agap

Bob York

Ae + g

Ae + 2 g Ae
NI max
Bmax Ae
Rgap

Flux in gap = flux in core

L 0 N 2

Agap
g

2
LI max
Ae

g
1 + 2 g / Ae

Fringing explains why the


data above is sub-linear

Back to TOC

Design Example
Circuit Spec:

Assumptions:

L = 200 H
I dc = 2 A

Bmax = 0.3T

I = 0.4 A

J max = 400A/cm 2

Ferroxcube E25/10/6 bobbin and core Datasheet Parameters:

Ku = 0.5

Required Area Product:

Ap > 1520mm4
E25/10/6 core satisfies this criterion.
Using this core, the design requires:

AL < 145nH
For 3C81 ferrite (a general purpose
material for <200kHz) a standard gapped
core is available with:

AL = 100nH
From this we find:

N=

L
= 47turns
AL

Bob York

PCu 0.3Watt
Back to TOC

Core-Geometry (Kg) Method of Core Selection


Unlike the area-product method, the core-geometry method explicity includes the winding loss as part of the
core selection process.
The condition (2) for B<Bmax can be written as:

Using this, the inductance relation (1) gives:

Condition (3) gives:

Abw

NI max
Bmax
le
NAe Bmax
L
I max

K w Ab
N

Using the last two results the winding loss is:

PCu Cu

L I max
Ae Bmax

MLT 2 L I max
I rms

K w Ab
A
B
e max

This result combines all the important constraints on the inductor design. If
we group all the parameters relating to the core geometry on the left we get:

Kg

2
L I max
Cu I rms

PCu K w Bmax

where

Ab Ae2
Kg =
MLT

Kg is the core-geometry constant. It can usually be calculated for any core-bobbin combination, and thus gives a
method for chosing a core when power loss is specified.
Once the core is selection, the air-gap is chosen as outlined previously.

Bob York

Back to TOC