Protection from Tempest attacks
Korejwa
I am looking for advice on how to build a secure (number theory) research
lab. Basically, I need a secure location to develop some ideas using "off
the shelf" computers in total privacy. It should be impossible to monitor
EM transmissions from outside the lab using satelites or tempest sensitive
equipment.
I have some radar-absorbing plasic material. (which I obtained from
machinists who worked on government projects) But I don't even know if
it's leagal for me to own, much less buy sheets of it to surround a room.
Does anyone know where I can buy radar-absorbing material?
I had an aquantence who installed a SR-71 flight simulator at Area-51 back
in the 80's. The simulator was in a building within a building within a
building, so Russian satellites couldn't detect EM transmissions and learn
about the SR-71's capabilities. Does anyone know why they would use three
barriers to block EM transmissions? Is there a technical reason, or was
this just redundency to protect national security?
I'd like some input from anyone who has experience with tempest or
anti-tempest equipment. When looking at the (supposed) security companies
on the internet, I get the feeling they're ready to install spy cameras in
any room I pay them to protect. I feel like I'd be better off just
surrounding a room with chicken-wire to block EM transmissions, and hope
for the best.
I want to protect myself from tempest attacks. My preference is to build
a secure building or room that will block all EM transmissions from
typical electronics, since I am doing research and don't know what
electronics I will need in the future. I assume there is a way to do this
by installing some EM absorbing material in the walls. Is this possible?
If so, where can I buy such material?
Any input is much appreciated. Thanks.
-John Korejwa
Massachusetts, USA
Tom St Denis
>
> I am looking for advice on how to build a secure (number theory)
> research lab. Basically, I need a secure location to develop some ideas
> using "off the shelf" computers in total privacy. It should be
> impossible to monitor EM transmissions from outside the lab using
> satelites or tempest sensitive equipment.
Why not use LCD monitors instead of
blast-your-head-and-rob-you-of-your-child-jelly CRTs?
> I have some radar-absorbing plasic material. (which I obtained from
> machinists who worked on government projects) But I don't even know if
> it's leagal for me to own, much less buy sheets of it to surround a
> room. Does anyone know where I can buy radar-absorbing material?
>
> I had an aquantence who installed a SR-71 flight simulator at Area-51
> back in the 80's. The simulator was in a building within a building
> within a building, so Russian satellites couldn't detect EM
> transmissions and learn about the SR-71's capabilities. Does anyone
> know why they would use three barriers to block EM transmissions? Is
> there a technical reason, or was this just redundency to protect
> national security?
Dude don't forget the tinfoil for your head so they don't suck your
brainwaves out.
> I'd like some input from anyone who has experience with tempest or
> anti-tempest equipment. When looking at the (supposed) security
> companies on the internet, I get the feeling they're ready to install
> spy cameras in any room I pay them to protect. I feel like I'd be
> better off just surrounding a room with chicken-wire to block EM
> transmissions, and hope for the best.
Easy.
Build an anechoic chamber suspended in a faraday cage surrounded by 2'
thick walls mixed with lead.
Should be no sound/RF leakage ;-)
Though all this for a "number theory" project? x^3 == 5 mod 7, oh no I
need national security for this one!
Tom
Louis Scheffer
>I am looking for advice on how to build a secure (number theory) research
>lab. Basically, I need a secure location to develop some ideas using "off
>the shelf" computers in total privacy. It should be impossible to monitor
>EM transmissions from outside the lab using satelites or tempest sensitive
>equipment.
For some non-classified but technically competent approaches to this problem,
check out what radio astronomers have done. They have almost exactly the
same problem - they would like to use computers, but not have any appreciable
radiation leak out.
>I have some radar-absorbing plasic material. (which I obtained from
>machinists who worked on government projects) But I don't even know if
>it's leagal for me to own, much less buy sheets of it to surround a room.
>Does anyone know where I can buy radar-absorbing material?
This is the wrong approach, in general. You are much better off with a
shielded room. Basically the idea is very simple - make all the walls with
a conductive material, and all the radiation stays inside. The biggest
problem is stuff that penetrates the barrier - doors, electrical lines,
air ducts, and so on. There are well known techniques for fixing each
of these.
>I had an aquantence who installed a SR-71 flight simulator at Area-51 back
>in the 80's. The simulator was in a building within a building within a
>building, so Russian satellites couldn't detect EM transmissions and learn
>about the SR-71's capabilities. Does anyone know why they would use three
>barriers to block EM transmissions? Is there a technical reason, or was
>this just redundency to protect national security?
Certainly two layers, or at least an airlock type structure, makes sense.
Otherwise you can't open the door while the stuff is running. Or perhaps
it was too hard to get a level of attenuation they were absolutely sure
was safe with just one or two layers. Often the specs for this kind of
stuff are drastic overkill.
>I'd like some input from anyone who has experience with tempest or
>anti-tempest equipment. When looking at the (supposed) security companies
>on the internet, I get the feeling they're ready to install spy cameras in
>any room I pay them to protect.
That's why I suggest looking at what radio astronomers have done. They
are a pretty open bunch, will tell you how to make your own measurements,
and have no security angle themselves.
>I feel like I'd be better off just
>surrounding a room with chicken-wire to block EM transmissions, and hope
>for the best.
This is in general the right approach, but chicken wire almost surely
won't do - it's not conductive enough, the holes are too big for modern
clock frequencies, and it's hard to make the seams good enough.
>I want to protect myself from tempest attacks. My preference is to build
>a secure building or room that will block all EM transmissions from
>typical electronics, since I am doing research and don't know what
>electronics I will need in the future. I assume there is a way to do this
>by installing some EM absorbing material in the walls. Is this possible?
>If so, where can I buy such material?
I'd go for sheets of copper soldered at the corners and overlaps. Sheets of
copper are available from many sources (roofing supplies, I think,
often have them). You will need good filters (perhaps several in series,
they are mainly intended to prevent gross interference, not meet
tempest requirements) where your power goes through
the walls. These filters must be well connected to the shield walls. Any
electronics catalog will have these. You will need to install a fairly
tight (few mm) mesh over all the air ducts, and make sure it's well
grounded to the walls. Your door(s) will need to be covered in copper, with
the little metal interlocking fingers all the way around to make a solid
connection when it's closed. I don't know where to get these fingers, but
they are used all the time in computer cases so they must be available.
Then you will need to test it - fire up your computers inside, then take
a good receiver just outside and see if you can detect any feedthrough. If
so fix the problem(s).
>Any input is much appreciated. Thanks.
This is not rocket science, but it's a lot of work to build an RF tight room
and keep it tight.
Lou Scheffer
Mok-Kong Shen
Korejwa wrote:
>
[snip]
> It should be
> impossible to monitor EM transmissions from outside the lab using
> satelites or tempest sensitive equipment.
If you are really interested in TEMPEST, you should anyway
study:
http://www.cl.cam.ac.uk/TechReports/UCAM-CL-TR-577.pdf
M. K. Shen
Simon Johnson
"Tom St Denis" <tomst...@iahu.ca> wrote in message
news:PK2pc.994$0qd...@twister01.bloor.is.net.cable.rogers.com...
> Korejwa wrote:
>>
>> I am looking for advice on how to build a secure (number theory) research
>> lab. Basically, I need a secure location to develop some ideas using
>> "off the shelf" computers in total privacy. It should be impossible to
>> monitor EM transmissions from outside the lab using satelites or tempest
>> sensitive equipment.
>
> Why not use LCD monitors instead of
> blast-your-head-and-rob-you-of-your-child-jelly CRTs?
Child Jelly? We call it man-jam over here.. well up north in england anyway.
Si.
flip
news:opr7znhc...@news.west.earthlink.net...
> I want to protect myself from tempest attacks. My preference is to build
> a secure building or room that will block all EM transmissions from
> typical electronics, since I am doing research and don't know what
> electronics I will need in the future. I assume there is a way to do this
> by installing some EM absorbing material in the walls. Is this possible?
> If so, where can I buy such material?
http://www.eskimo.com/~joelm/tempest.html
Douglas A. Gwyn
> I am looking for advice on how to build a secure (number theory) research
> lab.
Why? Surely the main aim of research in general is to
propagate knowledge. If, however, you're working on some
actual "classified" project you need to contact your SSO
rather than guess what the actual security requirements are.
d...@florence.edu
wrote:
I know i've been watching too many Star trek episodes, but seems as
though state of the art low emission equipment might involve super
conducting thin films and photon based peripheral switching
interfaces?
Peter Fairbrother
> I know i've been watching too many Star trek episodes, but seems as
> though state of the art low emission equipment might involve super
> conducting thin films and photon based peripheral switching
> interfaces?
>
I want one. Who makes them?
--
Peter Fairbrother
d...@florence.edu
Lots of links on the web for super conductors, this company looks like
they sell old school stuff.
http://www.smartmoney.com/att/eqsnaps/index.cfm?story=snapshot&symbol=ISO&fk=13056
Peter Fairbrother
> On Sun, 16 May 2004 17:54:30 +0100, Peter Fairbrother
> <zenad...@zen.co.uk> wrote:
>
>> d...@Florence.edu wrote:
>>
>>> I know i've been watching too many Star trek episodes, but seems as
>>> though state of the art low emission equipment might involve super
>>> conducting thin films and photon based peripheral switching
>>> interfaces?
>>>
>>
>> I want one. Who makes them?
>
> Lots of links on the web for super conductors, this company looks like
> they sell old school stuff.
:)
But I still want ... a room temperature superconducor covered laptop, with
no EM, electric or magnetic emanations except the light from the screen,
which should be projected only towards the pupils of my eyes.
I use an old 6mm welded steel fuel tank, with one end made into a door,
about 3 feet wide, seven feet high and five feet deep. No electrical through
connections, lead-acid batteries and an invertor inside for power, but two
three foot long four inch dia steel air in and out pipes, quad (two fillets
on each side) welded to the tank, filled with wire wool and with a 1/16 inch
copper mesh cover brazed on. Fans from old cpu's on the external ends.
The door has copper fingers I got from Maplins UK, and a further wire-wool
curly thing. Hard to describe, but easy to imagine - wire wool meets wire
wool.
All the wire wool has a teeny tiny bit of very diluted silver conductive
paint on it, but ... it is inside a cupboard, but it hasn't corroded.
I only use it for keygen tho'.
Actually, I've never needed it yet ...
--
Peter Fairbrother
Joe Durusau
To the O.P.: If you are really serious about all EM radiation, get
a good book on electromagnetic theory. (A basic one will do to show
the principles involved). Then decide on how serious you are, because
perfect never exists except in theory. You will presumeably want
to be able to stay inside your development area for significant time,
which will imply ventilation and non-perfect shielding. You will more
than likely need external electrical connections, which will mean
even more plumbing. The details of true "Tempest" shielding rules
are classified, but the underlying principles are well-known. BTW,
make sure that there are no exterior walls (and especially no windows)
in the area, because there are other attacks available that way.
For the truely serious, there are many consultants that will build you
such a faciility, who also have the equipment to test it, since without
a proper test, everything is just a "maybe".
Speaking only for myself,
Joe Durusau
d...@florence.edu
<zenad...@zen.co.uk> wrote:
According to Ted Kaczynski at his trial, the international superpowers
can monitor brain waves from low earth orbit. Since brain waves would
be magnetic in nature they would diminish by the cube of the distance.
Kaczynski was reasonably bright and good at math so he would have had
to assume superconductor technology would not suffer from divide by
zero overflow.
Of course the hand cipher Kaczynski used to document his deeds was
broken quite easily by the FBI. That would cast doubts on his ability
to estimate his adversaries.
Peter Fairbrother
> Why not use LCD monitors instead of
> blast-your-head-and-rob-you-of-your-child-jelly CRTs?
They aren't much better.
Markus Kuhn (the "optical tempest" guy) has just done a paper,
"Electromagnetic Eavesdropping Risks of Flat-Panel Displays", though it
isn't on the 'net yet afaik. Drop him a line and maybe he'll put it up,
though he's very busy just now.
From the "conclusions" section: "the eavesdropping risk of flat panel
displays ... is at least comparable to that of CRTs". There are also some
nice eavesdropped images, including readable text snooped from 10 metres
through 2 walls.
For general advice on Tempest, the technical report that came out of
Markus's PhD thesis is the nearest thing to a bible of publicly-available
Tempest knowledge that exists:
http://www.cl.cam.ac.uk/TechReports/UCAM-CL-TR-577.pdf
though be warned, it's about 8 Megs.
TEMPEST attacks are expensive to mount, so unless you have valuable secrets
to protect you aren't in much danger from them. At the moment.
--
Peter Fairbrother
Mok-Kong Shen
Peter Fairbrother wrote:
[snip]
> TEMPEST attacks are expensive to mount, so unless you have valuable secrets
> to protect you aren't in much danger from them. At the moment.
If one does want to protect against such attacks, what are
the appropriate measures to take in practice? (Someone
mentioned Farady cage. In that case I like to know what
(concrete) kind of Farady cage to use.) Thanks.
M. K. Shen
Andrew Swallow
"Mok-Kong Shen" <mok-ko...@t-online.de> wrote in message
news:c9qlmi$vf8$00$1...@news.t-online.com...
mesh that covers the top, bottom, left, right, front and back
of the object being protected. There should be a low
electrical resistance path between any two points on the
cage. The entire cage should be earthed.
Andrew Swallow
nemo
Loosely speaking you are correct but it's not quite that simple.
(Historically the Farady cage was originally applied for static
electricity.) Applying it at high frequencies (e.g., 3 GHz) is
non-trivial. Grounding at such frequencies is largely
superfluous (and a bitch to do well!).
Regards,
Mok-Kong Shen
Andrew Swallow wrote:
> Faraday cages are oversized bird cages. They are a metal
> mesh that covers the top, bottom, left, right, front and back
> of the object being protected. There should be a low
> electrical resistance path between any two points on the
> cage. The entire cage should be earthed.
I know that and had also seen an experiment in which it
protected a person from a lightening created in lab with
high voltage. However, as layman in physics I 'conjecture'
that the mesh size plays a certain essential role. Anyway,
if the mesh were of, say, 1m*1m size, I couldn't imagine
that it would prevent e.g. ordinary radio waves going through.
So, to more concretely formulate my question: What's the
appropriate mesh size for a Farady cage that well protects
computer (PC/workstations) work? Should other practical
matters also play a role, it would be fine to be able to
know them too. Thanks.
M. K. Shen
lyal
It is generally found that the best option is 2 separate (insulated from
each other, except at 1 earthing point) layers of 1.5 mm steel or copper
sheet - not mesh, if the goal is looking for radiation supporession abot a
few Mhz. I forget the exact numbers, but openings and surface flaws can
become good re-radiators when a linear dimension is more than about 1/6th -
1/10th of the wavelength of interest.
But don't waste money or time on this level of construction on unless
conducted energy is also effectively addressed - e.g. data and powerlines.
See the tempest docs at crytome.org for hints, and all radio engineering
books, and EMC suppression texts for more details.
Lyal
Mok-Kong Shen
lyal wrote:
>
> It is generally found that the best option is 2 separate (insulated from
> each other, except at 1 earthing point) layers of 1.5 mm steel or copper
> sheet - not mesh, if the goal is looking for radiation supporession abot a
> few Mhz. I forget the exact numbers, but openings and surface flaws can
> become good re-radiators when a linear dimension is more than about 1/6th -
> 1/10th of the wavelength of interest.
> But don't waste money or time on this level of construction on unless
> conducted energy is also effectively addressed - e.g. data and powerlines.
>
> See the tempest docs at crytome.org for hints, and all radio engineering
> books, and EMC suppression texts for more details.
Thanks for the informations. That suffices for my current
curiosity. The two-layered metal sheet reminded me of a discussion
quite long ago about shielding a mobile phone, where it was found
that wrapping the phone with tin foils several times (without the
isolation you mentioned!) didn't work. (There is BTW a firm
manufaturing portable bags, claimed to be able to shield mobile
phones.) A problem I could imagine, though, is that, since metal
sheets don't let air through, ventilation might be an issue for
the person working in the cage.
M. K. Shen
Louis Scheffer
>Andrew Swallow wrote:
>> Faraday cages are oversized bird cages. [...]
>I know that and had also seen an experiment in which it
>protected a person from a lightening created in lab with
>high voltage. However, as layman in physics I 'conjecture'
>that the mesh size plays a certain essential role. Anyway,
>if the mesh were of, say, 1m*1m size, I couldn't imagine
>that it would prevent e.g. ordinary radio waves going through.
Your intuition is exactly right. To a first approximation,
waves shorter than the hole size pass though, waves larger
are reflected and cannot pass. This is exactly why your AM
radio fades out when going under an overpass (AM uses about
200 meter waves) and FM does not (FM uses 3 meter waves)
>So, to more concretely formulate my question: What's the
>appropriate mesh size for a Farady cage that well protects
>computer (PC/workstations) work?
Considering that a 3 GHz clock is a 10 cm wave, and contains
many harmonics of shorter wavelength yet, a 1mm mesh would
be a reasonable guess. (Note: I have not measured this, nor
mounted any Tempest attacks myself. This is just a first
guess from the physics involved.)
Should other practical
>matters also play a role, it would be fine to be able to
>know them too. Thanks.
An important one is: watch the seams. A bad seam looks like
a hole to a properly polarized wave.
A second one is to properly filter any power or other signal
that passes through the cage. (This requires much stricter
filters than are normally used.)
Finally, test your resulting cage. Like all security, your
first attempt probably won't be perfect, and since leaks are
not obvious, only explicit testing will find them.
Lou Scheffer
Michael Amling
> Mok-Kong Shen <mok-ko...@t-online.de> writes:
>
>>Andrew Swallow wrote:
>>
>>>Faraday cages are oversized bird cages. [...]
>
>>I know that and had also seen an experiment in which it
>>protected a person from a lightening created in lab with
>>high voltage. However, as layman in physics I 'conjecture'
>>that the mesh size plays a certain essential role. Anyway,
>>if the mesh were of, say, 1m*1m size, I couldn't imagine
>>that it would prevent e.g. ordinary radio waves going through.
>
> Your intuition is exactly right. To a first approximation,
> waves shorter than the hole size pass though, waves larger
> are reflected and cannot pass. This is exactly why your AM
> radio fades out when going under an overpass (AM uses about
> 200 meter waves) and FM does not (FM uses 3 meter waves)
That's most of it. With AM the volume goes down as the signal
strength goes does so it's more obvious when you're under that overpass.
With FM, the volume stays the same as the signal strength goes down (but
the signal to noise ratio gets worse).
--Mike Amling
Allan Herriman
wrote:
AM radios have an AGC (automatic gain control) to partially counteract
that effect over the expected range of signal strength.
Decreasing the signal strength will cause the volume to remain roughly
constant until the AGC is at maximum gain. Only then will the volume
decrease (e.g. under the overpass).
Regards,
Allan.
nemo
..snip...
>Finally, test your resulting cage. Like all security, your
>first attempt probably won't be perfect, and since leaks are
>not obvious, only explicit testing will find them.
>
> Lou Scheffer
Your advice is vaguely in the right direction, but is about as
helpful in practice as saying "to make a nuclear bomb, first get
a pound of plutonium."
The questioner would be, if anything, even less capable of
testing a Tempest-shielding Faraday cage than building one!
Moreover, he is unlikely to have any idea of what attenuation is
"good enough."
Regards,
Lou Scheffer
> In article <40c3fcd6$1...@news.cadence.com>, Louis Scheffer <l...@cadence.com> wrote:
> ..snip...
> >Finally, test your resulting cage. Like all security, your
> >first attempt probably won't be perfect, and since leaks are
> >not obvious, only explicit testing will find them.
>
> helpful in practice as saying "to make a nuclear bomb, first get
> a pound of plutonium."
>
> The questioner would be, if anything, even less capable of
> testing a Tempest-shielding Faraday cage than building one!
> Moreover, he is unlikely to have any idea of what attenuation is
> "good enough."
>
yourself, can be useful. If you hire someone to build you a Tempest
proof cage, they should be able to show you how they will test it.
Better yet, if you are seriously worried about security, hire company
A to build the cage, and then ask company B to test it, each with
incentives for finding errors in the other's work.
I may have no idea how to get a pound of plutonium, but if I'm going
to buy a supposed atomic bomb, I'd make very sure it had at least a
pound of plutonium inside.
Lou Scheffer
nemo
As an intellectual exercise, knowing how to build and test a
Tempest-shielding cage would be interesting.
Interesting, not useful :-)
Regards,
PS My point was that building an effective Tempest shielding
cage is quite difficult and likely to be well beyond the
capability of the inquirer who started this thread. (If it isn't
difficult then there are some specialized manufacturers who are
grossly overcharging :-)
But, properly testing such a cage would require equipment and
expertise **far beyond** merely building one. Equipment and
expertise that few outside the military, the TLAs, and a few
specialized manufacturers are likely to have.
Coupling the difficulties of building and testing such a device
with the dearth of current information in the public domain re
capabilities and threats, and the required level of performance
in the face of hostile attempts to electronically eavesdrop,
means there isn't even a clear idea of realistic objectives, let
alone methods of achieving them.
Donald Hines
>Actually, I've never needed it yet ...
Problem as I see it, is that building (or attempting to build) a
tempest safe computer room is like putting up a neon sign that says "I
HAVE SECRETS".
You attract attention, and if the "powers that be" decide they might
actually be INTERESTED in what you are working on, they'll just get a
warrant and go in and TAKE the computer right out of the tempest safe
room and into a comfortable office at the NSA or homeland security.
If you are going to whisper secrets, meld in with a crowd, don't put
up some elaborate cone of silence. Your best safety is in obscurity
because there isn't much that will actually stop a goverment agency
that decides it has an interest in what you are doing.
Donald Hines
lyal
> In article <3a6089b6.04060...@posting.google.com>,
> l...@cadence.com (Lou Scheffer) wrote:
> >nemo ou...@erewhon.com (nemo outis) wrote in message
> > news:<3T6xc.664909$Pk3.179779@pd7tw1no>...
> >> In article <40c3fcd6$1...@news.cadence.com>, Louis Scheffer
<l...@cadence.com>
> > wrote:
> >> ..snip...
> PS My point was that building an effective Tempest shielding
> cage is quite difficult and likely to be well beyond the
> capability of the inquirer who started this thread. (If it isn't
> difficult then there are some specialized manufacturers who are
> grossly overcharging :-)
>
> But, properly testing such a cage would require equipment and
> expertise **far beyond** merely building one. Equipment and
> expertise that few outside the military, the TLAs, and a few
> specialized manufacturers are likely to have.
Most of the screened enclosures I've seen ahve been tested with commercial
and off the shelf hardware. Not cheap hardware, but only couple of items are
expensive enought that you'd hire the unit for a few days - the rf
generation and antenna needs are pretty easily obtained, little rf power is
invovled.
The big hassle is getting the test environment right electrically and RF
wise, physical space in which to move the antennas, and the major hassle of
fixing problems may mean starting over, in the worst case.
Attentuation levels stem from 2 things imho - simple rf engineering (e.g.
the noise floor in a given bandwidth, how many harmonis of signals of
interest are you concerned about - text book stuff and a few minutes
thought) and ones level of paranoia. Mind you, if you go to the effort
building then testing a screened enclosure, paranoia = max anway :-)
lyal
Lyal
d...@florence.edu
wrote:
Some tempest level 4 hardware would not necessarily be used to process
secret information. New parallel processing hardware might be used
for super fast table handling of non sensitive data. Manufacturers
would naturally prefer that such equipment could not be reverse
engineered. Perhaps implementing tamper reactive black box designs
for business applications would be possible?
Intel is gearing up to roll out the 4 GHZ Prescott processor next
year. At 4 GHZ these silicon chips are getting kind of power hungry.
Intel also dabbles in parrallel processing with research partners like
Ovonyx. There have been very few hints about any Super Chip actually
making it to the commercial markets however.
nemo
You bring what I would call 'good engineering knowledge' and
'intelligent speculation' to the problem. But, in doing so, you
reinforce my primary message: the original questioner is unlikely
to have access to suitable equipment and, more particularly, the
expertise to use it well. Nor would he have a realistic way of
specifying acceptable performance, except as wild conjecture. He
might as well buy a 'canned solution' from some Tempest-shield
manufacturer.
But, going along with your approach (reasonably strong
theoretics with pragmatics a decade or two behind the likely
state of the art), I will concede, arguendo, that it might be
possible to build and test a Tempest-shielding cage that would
thwart, say, a serious amateur.
But is that the likely adversary who would be using Tempest. I
think not!
If you truly need Tempest protection, and are not just a hobbyist
or dilettante, you are likely dealing with serious adversaries
(something on the scale of a national police force or
intelligence agency - a TLA). And for them such a home-built
cage would not likely be sufficient. It could, in fact, be a
liability if false confidence caused over-reliance on it. At
best - and I'm being charitable - it might be a mildly useful
auxiliary or supplemental defence.
Regards,
Louis Scheffer
>If you truly need Tempest protection, and are not just a hobbyist
>or dilettante, you are likely dealing with serious adversaries
>(something on the scale of a national police force or
>intelligence agency - a TLA). And for them such a home-built
>cage would not likely be sufficient.
The TLAs are good, but they have no god-like powers. Radio astronomers
use computers near radio telescopes, and they do it with exactly this
kind of engineering. And radio astronomers need to keep their signals
away from receivers that are more sensitive than any TLA could use (since
astronomers look only at the sky (about 3 degrees K) rather than the ambient
temperature faced by any earthbound TEMPEST attack. )
A home-built shield built and tested to good open specs, such as radio
astronomers use, could (almost surely) foil any practical TEMPEST attacks
based on radio waves. Even if the shielding is not perfect, it will make
the TEMPEST attack so hard that the TLA will then shift to plan B
(bribe your friends), plan C (insert a physical bug in your computer
when you are not looking), a plan D (using a known OS bug, insert a virus
that logs your activity), and so on. These attacks are quite practical
and a much much larger danger than some technology attack on a good but
not perfect TEMPEST proof cage.
Lou Scheffer