Xref: utzoo sci.electronics:4816 rec.ham-radio:8126
Path: utzoo!attcan!uunet!husc6!encore!cloud9!jjmhome!cpoint!die
From: die@cpoint.UUCP (David I. Emery)
Newsgroups: sci.electronics,rec.ham-radio
Subject: Building a simple home circuit to read data off remote crt screen
Keywords: tempest, emc, rfi, correllator, pll
Message-ID: <2010@cpoint.UUCP>
Date: 23 Jan 89 04:54:47 GMT
Reply-To: die@cpoint.UUCP (David I. Emery)
Organization: Clearpoint Research Corp., Hopkinton Mass.
Lines: 45


	A year or so ago I saw a demo on TV (on ABC's Nightline I think) of
what was refered to as a simple device that could be built with $25
worth of parts that could recover an image of the text on a CRT display
belonging a nearby PC or terminal from its spurious electromagnetic radiation.

	The demo, done in London to circumvent strong US government opposition
to demoing this technology on commercial TV, showed a somewhat noisy but
quite readable image of text on the screen of a PC about 25 feet away from
the loop antenna used to pick the signal up.

	Does anyone know what the $25 circuit consists of ?
Can it really be built for $25 ?

		Some thoughts

	It is quite clear that the video circuitry in a monitor radiates
broadband pulses with a spectral peaks at the ringing frequencies of the video
circuitry (eg. the wires to the tube ring like a spark transmitter when
excited by the hf energy contained in the bright/dark and dark/bright
transistions).  And it is also clear that a monitor radiates lots of
energy at harmonics of its sweep frequency (try using a VLF receiver
near a TV). 

	To me the obvious thing to do is to recover phase lock to the 
horizontal sweep by locking on to some harmonic, and divide that reference
down to generate frame/field timing.  The data might be recovered by
generating a sample pulse delayed by a variable digital time delay circuit 
from the phase locked timing reference and using that pulse to sample
the output of an envelope detector following a very broadband hf to vhf
amplifier driven by the pickup loop antenna.  By averaging thousands of these
pixel wide samples and then slowly sweeping the time delay by one pixel at a
time one could very slowly reconstruct the image.  Naturally this could be
done much faster (but not for $25) by using a modified frame grabber as
the correlator.  Another technique involves video frequency analog ccd
shift registers in a correlator.

	Is there a better way of doing this I am missing ?



-- 
	David I. Emery   Clearpoint Research Corp. 
	99 South Street, Hopkinton Ma. 01748  1-508-435-2000
	{decvax, cybvax0, mirror}!frog!cpoint!die