From: Michael Motyka <mmotyka@lsil.com>
Date: Thu, 24 Dec 1998 09:49:40 +0800
To: Igor Chudov <ichudov@Algebra.COM>
Subject: Re: Question about 'TEMPEST': UPS
In-Reply-To: <199812240040.SAA07841@manifold.algebra.com>
Message-ID: <36819772.503@lsil.com>
MIME-Version: 1.0
Content-Type: text/plain



Igor,

These were commercial cages used in the consumer products design
business. Things like radios, TVs, 900MHz cordless phones etc...do I
know for a fact that two or three layers of fine mesh copper ( BTW about
an inch apart ) were necessary? No. I just took the word of the guys who
had been doing the stuff for years. It wouldn't surprise me if the
commercial cages were overkill. Neither would I be surprised if anything
less leaked significantly. RFI is not an easy subject area. Find a
commercial cage vendor and talk to an apps eng.

Lacking an actual test facility, I suppose a really basic test would be
to use a hand-held AM radio tuned to a space between stations and try
out your PC with and without a chicken-wire cage. Sometimes a particular
piece of SW will have a pattern that you can recognize, maybe following
a keypress. I've been able to differentiate between button scan,
watchdog, servos, processor etc... It's a relatively cheap, quick
experiment. Try different frequencies and orientations of the ferrite
loop antenna. For higher frequencies ( CPU clocks + other harmonics )
you'll need other methods. I've been meaning to take a look at a
standard PC keyboard. Someday soon...if I do I'll let you know what I
find.

As for looks - tough - a PC generates lots of RFI, a partially
functional cage is probably worse than none at all because of the false
sense of security.

Trying to shield the computer and the monitor is an entirely different
can of worms. Probably an order of magnitude harder than making a good
cage.

Regards,
Mike

In real life even the easy things are not easy.

Igor Chudov @ home wrote:
> 
> Michael Motyka wrote:
> >
> >
> > >
> > Now... Just curious, if I simply make my study into a faraday cage via
> > use of fine (< 1/2 inch) chicken wire, and insulate my power system from
> > the computer emenations, would that be enough? Chicken wire is not that
> > expensive, and I think that I could make it look nice on walls.
> > >
> >
> > The cages I've used were framed in wood and had several layers of copper
> > screening. The mesh was like standard window screen, <.125". There were
> > two, I think sometimes three, layers. The corners/joints were copper
> 
> Michael,
> 
> Was the amount of work involved really necessary? What do three layers
> accomplish compared to just one?
> 
> > sheet. The screen was soldered to the corners. The doors were stepped or
> > beveled and had several sets of fingers - indside edge, outside edge,
> > perhaps one intermediate. Screen makes ventilation easier. Power should
> > be brought in through a separate, partitioned, shielded box with
> > filtering in each partition. Phone lines, same thing. You trust your
> > apps, os and TCP/IP stack to not rat you out? If not, work off-line
> > always.
> 
> Right.
> 
> > Al foil if it were thick enough would be OK but how do you reliably join
> > edges?
> 
> Glue? Perhaps some weak glue that would not destroy the underlying paint.
> 
> > Converting your study is non-trivial. The door would look like crap. You
> > would need the euivalent of a door over the window.
> 
> My window has a mesh on it. Is that not enough?
> 
> > I wouldn't trust 1/2" galvanized chicken wire for containing anything
> > but chickens or other non-gnawing furry woodland creatures.
> 
> Is that confirmed by evidence?
> 
> The reason for my question, I think that I somewhat value TEMPEST
> protection, but not enough as to cover my whole room with expensive and
> ugly looking copper or three layers of mesh and ferrite. I don't want
> all of it to be too conspicious.
> 
>         - Igor.