EM-nature (was: infra-org)

Bill Stewart bill.stewart at pobox.com
Sun Dec 8 21:01:32 PST 2013


>JYA wrote:
> > The celebrated EM components of TEMPEST are commonly
> > used as subterfuges, ruses and ploys to divert attention from
> > the non-EM -- which an almost limitless number of animals and
> > other creatures use for SIGINT and COMINT...

Back in the 80s I ran a TEMPEST-shielded computer room,
with a VAX 11/780 and later some Sun-2 workstations,
doing studies for the government that they didn't want
Commie Spies listening to.

The shielding was basically plywood with sheet metal on both sides,
with special metal joints in between them and on the corners,
packed with wire mesh in the joints to plug any extra leaks,
making a nice big Faraday cage.  The air vents were metal grates
an inch or two deep with zig-zaggy airflow paths,
and our data connections used fiber optics going through
waveguide holes that were about 3" long and 1/8" diameter.
Power feeds used big inductive low-pass filters.

We measured the RF-tightness using meters that ran at 450 MHz,
trying to keep things 120dB tight.  You had one person carry a
transmitter on one side of the wall, another person a reader on the other side.
One experiment we'd occasionally do was to stick a wire
through the fiber waveguides, and if you got more than halfway through,
that was enough for 450MHz to leak out.  I'm not sure what frequencies we
*really* needed to protect against, but the Vax ran at something like 10 MHz,
so 450MHz was way overkill for any harmonics that might happen.

What occurred to me recently was that 10 MHz computers may have been
state of the art for the mid-80s, but that was a lot of iterations of
Moore's Law ago, and I doubt the technology of the time is much use
for current 2-3 GHz laptops.  The main Faraday cage should be fine,
plus or minus a bit of extra copper tape to plug minor leaks,
but all the air vents are going to be awfully leaky at those frequencies.
Any idea what people use today?





More information about the cypherpunks mailing list