Richard Clark wrote:
Figure 5.28(a) shows a complete shielding. Of course this is entirely
electric, and arguably magnetic. However, magnetic flux can penetrate
thin shields, electric flux cannot.

Only when the shield is thin compared to the skin depth.
When the shield is thick relative to skin depth nothing gets through.

This is very easy to prove. I have been making measurements of a ~ .032
inch thick copper wall and with 0dB reference on a small resonant pick
up loop my analyzer is in the noise (-90dB) on the side directly
opposite that loop.

The same is true for a direct soldered connection to the wall on each
opposite side.

One inch to the side on the same side levels are -10dB. That would be a
two inch long path shorted by the copper the entire way. Go to the
direct opposite side through only .032 thick copper and levels are not
even detectable.

This is part and parcel to the world of isolated and shielded
circuits. The electrostatic shields are as effective as they are
complete in their coverage. Their contribution is measured in mutual
capacitance between the two points being isolated.

I don't have that reference and so cannot see that shield, but the only
thing the shield can do is reduce field impedance by changing the ratio
of electric to magnetic fields. In order to take either one to zero the
other must also be at zero.

I think the confusion comes from misapplying a grid forming a shunt
capacitance to reduce direct capacitance between two objects (forming a
"T" divider) to the shielded loop antenna or shielded link.

Consider a loop of any size, even a link in a tank coil. That conductor
has a field impedance and radiation characteristics largely set by the
diameter of the coil.

Once we put a wall around that conductor more than a few skin depths
thick NOTHING goes through that wall. The "shield" actually becomes the
coupling coil, the link inside simply develops a voltage across the
shield to drive that external coil. Both electric and magnetic fields
are present on the outer wall of the shield, and while they may be
modified by shield balance they really are not much different than we
had with just the inner conductor alone. We really just change the
balance.

With the grid, we have substantial air gap segmenting the "wall".
Naturally the coupling mechanism is different than we have with a solid
wall. We, in effect, have dozens of long gaps.

Each conductor in that grid is indeed excited by the magnetic and
electric fields, and each conductor has a potential difference across
area and a current flowing. Part of the field, both electric and
magnetic, leaks through. Part is reradiated by the currents and
voltages in the grid.

I think at some point of time MRT or Dave Saloff patented a right angle
grid of two layers with opposing ends in each adjacent conductor in
each layer grounded that I designed. The idea was to more evenly
distribute the fields and prevent "hot spots". This was for a medical
application.

Rest assured this applicator produced both time-varying electric and
magnetic fields, but the balance was so much improved tuning was more
stable. The improved balance and evenly distributed field meant the
feedline did not radiate any significant amount when brought near the
patient, unlike a regular multiple turn loop.

I still have some prototype applicators here, as well as the field
measurements required by the FDA. The applicator actually had to match
with lowest return loss into the buttocks of an average size 30 year
old female.

73 Tom