I've been doing my homework on separating silver coins from clad and found only four methods that can scientifically work (based on sorting halves):
1) Electrical resistance - this is impractical for many reasons, including requiring > 20 amps to test effectively.
2) Weight differential - look it up, but the difference for 90% (12.50g) 40% (11.5g) and clad (11.34g) - the hitch is a scale with +/- 5% accuracy and automating weighing
3) Coefficient of Restitution (again, look it up) - silver bounces differently off a very hard surface. I have been told the Fed banks use this method to strip silver from the coinage stream. Simple: bounce coins toward a barrier - some bounce higher and over, some don't.
4) Using Lenz's Law: a counter-electromotive force will be induced in copper when the coins move through a magnetic field. The cause of this CEMF is due to 'eddy currents'. That is, CEMF is a force caused by eddy currents as explained by Lenz's Law. The practical upshot of all this is that if you shoot a conductive object through a magnetic field it will slow down...