Proxmark3 community

iceman

Administrator

Registered: 2013-04-25

Posts: 9,538

Website

lf t55xx chk got some love

When tweaking "lf t55xx chk m"  command,  I got the new check algo to work.  It will need some more testing but initial results with ASK and FSK configured t5577 works.

there is two check algos as of now.
The first one,   based on client side,  takes a pwd,   test to read block0,  and tries to demodulate it and tries to see if it can decode it to a known "valid" block0.   If so, it considers a good pwd was found.
it is slow but it does the verification at each iteration.   Many commands back and forth between client and device.

The second,   in order to be fast,   tries to see what a portion of the received response from the tag would look like.   Like a average fingerprint over the first 2048 samples.
next it loops the pwds.   Where it does the same fingerprinting over the first 2048 samples of each try.   The test is to try finding which fingerprint that is the most different from the baseline fingerprint. the winner is considered a candidate.
So we never take a whole signal and we don't have no idea what configuration the tag has sent the data in,
since we don't know which candidate is the most different one until the whole loop has executed.   We send it back to the client and the client goes a whole read/demodulate/decode phase in order to validate it.

So  in general,  the pwds that are in the end or middle of the list ,   the second algo is faster at finding it.
if its the first 10 pwds in the list,  the first algo finds it much faster.

Pros and cons...
the second algo main loop is executed on the device side
while the first algo runs on client side.