One thing about the calculations for skin effect that is hard to figure
in is the affect of conductors that are not round or multiple conductors
placed near each other.
Transformer manufacturers have to deal with 'proximity effect' which is
basically a form of skin effect. In a bobbin wound transformer, the
currents in the wires don't run in a 'skin' around the wires, but tend
to run predominately either on the inside (toward the bobbin center) or
the outside (away from bobbin center), forming sort of a 'coil skin' on
one side of the winding. Exactly where it runs depends on the layering
of the windings etc. So the skin effect is overly optimistic for
Fundamentally, the current will tend to run as spread out as possible,
regardless of the conductor shape. For a strap, the current will not run
evenly in a 'skin' around the strap, but will prefer to run at the edges
of the strap, increasing the impedance beyond what you'd get from purely
a skin effect calculation.
Round wire and tubing is easy to calculate re: the skin effect because
of the symmetry of the conductor, i.e. the best place for the current is
a skin around the circumference.
Two wires placed in parallel work well since they will add more surface
area. But note that if they are placed right next to each other, the
surface of the wires that are near each other will not conduct much
current ('proximity effect').
TowerTalk mailing list