This document describes how electronic alarm system implementations done by Knogo (Library and stores) and Sensormatic (VIdeo stores) work. The following is provided for informational purposes only, and no guarantee is made to the accuracy of the supplied data.
It should be noted that some Knogo systems use a switch mat to trigger the drive coils in the alarm system. Sensormatic systems (and some later Knogo systems) are usually continuously driven.
A common system that is used relies on small metal strips with special magnetic properties. Systems manufactured by Knogo and Sensormatic rely on this principle. The strips are made from a material that saturates very easily at a very low magnetic intensity. This property is very rare among most materials (either they aren't magnetic at all, or saturate at much higher magnetic strengths) so it is possible to rely on this difference in order to detect the strips.
Since the strips saturate easily, detection is easy. If two large coils of wire are placed in parallel with each other (that is, so that the magnetic axis of the coils are the same) and a fairly strong alternating electric current is applied to one, a voltage is induced in the other coil. Introducing a security tag into the field will generate a spike in the detector coil waveform.
In order to validate this theory, A special "alternating magnetic field detector" consisting of an old walkman and a TV degaussing coil was built. The coil was attached to the input of the walkman (where the tape heads would normally connect) - this allows a user to hear the signal generated by the coil. This apparatus generated a strong tone when close to Knogo or Sensormatic detection systems (NOTE: Don't try to take the apparatus through the alarm system - there may be a very strong nonlinearity in the input circuit impedance of the walkman, and this may set the alarm off).
A device that could detect the strips was also constructed. This system consisted of two main rectangular loops of wire, dimensions 12 inches by 24 inches which were spaced 12 inches apart. The smaller scale model was built so that not as much driving power would be required. The coil that the signal was attached to consisted of 60 turns of 22 gauge wire (available at your trusty local Radio Shack) and the detector side consisted of 70 turns of wire wrap wire. A few windings (some experimentation required) of wire wrap wire were wound along with the driving side to provide a balancing voltage to offset the signal from the detector coil - this is required so that much more sensitive settings on the scope can be used. The circuit looks like this. (insert postscript thingie here)
If a fairly strong magnetic field was applied to the strip (such as that generated by a small ceramic magnet) the strip became saturated and no longer generated a pulse when placed in the formentioned apparatus.
To determine that this pulse was what the store alarm was looking for, an alarm strip biased with a small permanent magnet was taken through the alarm system. No alarm was generated. The same strip without magnetic bias would set off the alarm.
A coil of wire connected to two germanium diodes (diodes in parallel, anode to cathode) will cause a similar deformation in the signal as the strips, although I don't know if this will set off the actual alarm systems. (NOTE: Later experiments showed that this will not activate the system.)
Strips from the Sensormatic video tape tagging system will activate the Knogo systems, so I suspect that these systems work on similar principles.
System is fairly reliable, tags are long and thin (big advantage for video tapes and games) cannot be defeated by electrostatic shielding.
Can be defeated by magnetic shielding, sensing coils have to be close together greatly restricting egress from store, LARGE magnetic fields required to make this work (Is this a long term health hazard?) Some materials and circuits can set of this type of alarm system. Reusables are quite long and a hassle for customers.