Gregory L. Hansen said:
As I understand it, the EMP comes from the gamma flash stripping electrons
from atmospheric molecules, and they can bounce around for many minutes
before settling down.
Hardening equipment against nuclear EMPs is sort of a specialized topic, I
don't know much about it. But you should really design the equipment from
the beginning with that in mind. And you can forget about just slapping a
Faraday cage on to it. "Cage" won't work. Aluminum foil won't work.
Think more in terms of alternating layers of aluminum and a magnetically
permeable material like steel, mu-metal, or one of the MetGlas alloys.
I find no evidence in the literature that anything other than a Faraday cage
is required for complete protection, although it is possible that the
environment close to a nuclear blast might require what you describe.
You are probably reasoning that powerful magnetic fields could induce
potentials INSIDE the Faraday cage. However, although a Faraday cage is
permeable to constant or slowly changing magnetic fields, it rapidly becomes
impermeable as the frequency goes up, due to the skin effect. And the
induced EMF is proportional to the rate of change of the flux.
Check out
http://www.tfd.chalmers.se/~valeri/EMP.html
One author, with uncertain credentials, claims that antennas less than 30"
in length, are "safe". The context is not given.
But in this case, I don't have a Faraday cage. I'm not trying to harden
against nearby nuclear explosions, because, as others have remarked, there
would be other things to worry about. Nevertheless, one can approach a
Faraday cage in little steps. By isolating a local system from long
antenna-like structures, such as power lines and unshielded network cables,
it is possible to reduce the potential difference that can be induced
between parts of the system. All this has been proven in terms of protection
from lightning strikes. Two computers in different parts of a structure but
connected by a network cable can have lightning damage over the data cable,
because of ground currents, some of them caused by transzorb-based surge
protectors.
In particular, this means that CAT-5 cable should be replaced by optical.
Any cable that can serve as an antenna should be shortened to the minimum or
bundled correctly. And I have isolated the cable modem with a 3' fiber optic
link.
Power protection is more vexing. Transzorb based protectors limit the spike,
but tend to "light up" the ground. Current-dumping protectors provide better
control of lightning transients, but in the presence of a prolonged surge
simply follow the waveform up.
Isolation transformers provide the best common mode suppression, but
typically incorporate rudimentary or no normal mode suppression. I have
added two isolation transformers in front of my UPS'es, which contain
tranzorbs. If I knew something about the kind of waveform induced by EMP,
I'd have a better idea how to optimize power protection.
If futile against EMP, my efforts are hardening my home network against the
occasional "superbolt", which is a particular kind of strike that has about
100X the energy of a typical lightning strike.