Page 117 - Power Electronics Handbook

110 Electromagnetic compatibility
Because very low reflections occur in a magnetic field at the
air-conductor interface, the field inside the conductor is high. In fact,
using the magnetic equivalent of equation (4.11), H2 will be twice the value
of the incident field HI, for Z2 much smaller than Z1. the absorption
If
losses in the conductor are low then most of this energy will be reflected at
the conductor-air interface, and it will result in multiple reflections within
the conductor. This is illustrated in Figure 4.S(e), where a small amount of
field is shown to leak away at each reflection. Eventually, in the ideal case
of zero absorption loss, half the field will emerge to the right and half to the
left, so that there are no reflection losses. To allow for this multiple-
reflection phenomenon a term BI is introduced into the shielding
effectiveness equation (4.3).
The value of the correction term B, is given by equation (4.19).
B, = 20 log (1 - e-w6) (4.19)

It is always negative to show that reflection loss predictions without
considering multiple reflections are too optimistic. BI can be omitted if the
absorption loss is high, greater than about lOdB, so that multiple
reflections are minimised, although for low-frequency magnetic fields, and
thin shield, this term is almost always needed.

4.5.2 Shield design
The shielding effectiveness is often less than that predicted by the
equations given in the previous section due to discontinuities in the shield,
such as seams and holes. These discontinuities impede the flow of induced
current in the shield, which is responsible for generating a field opposing
the interfering field. Wherever possible, the location of the discontinuity
must be such as to minimise its effect on these currents. For example, the
seam should be located such that the circulating current does not have to
flow across it, as in Figure 4.6(b), but flows parallel to it, as in Figure
4.6(a).
The joints between the different parts of the shield represent its weakest
point, and the transfer impedance of these joints is a useful concept for
characterising them. If a current Z flows on the outside of the shield, as in
Figure 4.6(c), and it induces a voltage V on the inside of the shield, then
the transfer impedance is equal to the ratio VU. The impedance at a seam is
due to the contact impedance and the surface impedance of any gasket
which is used. The contact impedance can be represented by a
parallel-connected RC circuit whilst the gasket impedance is a series-
connected RL circuit. Therefore if there is poor contact resistance this will
usually show up as a drop in transfer impedance as the frequency increases.
The contact resistance at a joint must be kept as low as possible, and this
can be done by applying sufficient pressure; by using materials with low
contact resistance; by having a large surface area of contact; and by
avoiding corrosion. Figure 4.6(d) shows how simple shaping techniques
can be used to increase the contact area of a joint, and corrosion can be
reduced by not placing dissimilar materials in contact with each other.