DrinkGreenTea said:
[...]
So I can see the change in amount of the detection and correlate which card was pulled away. Looking at the reader of a change of X amount, I would know exactly which card was pulled out of the deck.
Would this be possible by creating magnetic field changes which can be read. Does anyone have some practical and/or plausible ideas to help me work on this?
I have to say, I'm quite pessimistic about this endeavour.
I don't see a scheme where cards are magnetized in some way, allowing you to identify the single card that is missing from a deck. As to using weight: no doubt sufficiently sensitive, sufficiently small weighing devices exist.
The hard part would be to manipulate the individual cards of a deck in such a way that you can differentiate between them. Assuming quite sturdy cards I estimate a weight of 2 gram per card. Assuming the weight difference between the cards must be 10% max you can vary the weight from 2000 to 2200 microgram.
That would lead to a scheme with successively the following weights: 2000, 2004, 2008, 2012, [...] , 2204, 2208
You calibrate the weighing scale to counting units of 4 micrograms, so you count as if the cards of the deck are 501, 502, 503, 504, [...] 551, 552 units
Total weight of the deck in such units: 26000 + 1378 = 27378
If you can manage to manipulate the weight of individual cards to within microgram precision you have a shot at the above scheme. As said, the hard part would be to manipulate the weight of the cards that precisely, while preserving the look of an untampered deck.
The reason the weight scheme can work in the first place is that weight is straightforwardly additive. The order of the deck does not matter, weight just adds up.
By contrast, magnetism has magnitude
and direction, and a magnetic field falls of with the third power of distance. When sensing a magnetized object the orientation of the sensor relative to the magnetic field matters. Distance matters a lot, so for the accumulated magnetic field the order in a deck of cards would matter. There are just way, way too many variables to have any chance of a reproducible reading.