What you Need
What to do
Make sure your plastic card is clean and not greasy.
Sand some rust off your object, the resulting lumps should be fine and not great bit lumps. You are aiming for about a quarter of a teaspoon of rust dust.
Now tap off the excess rust.
Look at the reflection of a uniform white light source, like the sky or a fluorescent tube, in the magnetic strip.
What may happen
If you get it right your should find that the rust sticks to the magnetic strip in lines.
Why does it happen?
Unsurprisingly the magnetic strip on a credit card stores its information magnetically so areas of the strip will be magnetised, which stores the data on the strip. When you swipe the card through a reader small currents are induced in the reader, which are interpreted as the card number etc.
Although the brown Fe2O3 rust is non-magnetic black FeO is magnetic and makes up Haematite and Magnetite ores, and so are any particles of iron attached to the rust. This means that the fine particles will stick to the magnetic parts of the strip, allowing you to see how the data is encoded. Different strips will have different arrangements of lines to encode the different numbers.
We were sent some lovely closeup photos form Randy Hirsch
I was trying to work out if I could actually find a way to read off the magnetic poles. The problem with iron is that it sticks either way around (it's a soft magnetic material).
The actual polarity is not relevant, as it is the changes that are read out by the read head and electronics. The actual polarity of individual magnetic domains is not important, the change is. SeanB, Mon, 8th Nov 2010
Depends what you want to use it for. I'm messing about with Halbach arrays and it would actually be useful to know polarities. wolfekeeper, Tue, 9th Nov 2010
Most card readers use a tape head, which is a basic coil with a gapped core, and follow it with a large amount of amplification that is AC coupled for stability. Thus the output is a differentiated representation of the magnetic recording, and pretty much all polarity information is stripped out when it is converted to a digital pulse train by a fixed point comparator. You get ( on the lowest data rate tracks) a clock and data that is robust and easy to recover, but which is not very high density recording.