That's all perfectly fine, I'm more than happy with the amount of time it took I just wanted to check that is how it's designed and it wasn't the resistors putting the can chips into a different mode slowing it down as most posts suggest this should be quick like 10-20minsvtl wrote: ↑13 Feb 2023, 11:15The algo consists of two parts: timing attack on first 3 bytes of the PIN and brute force on the last 3 bytes.liamstears wrote: ↑13 Feb 2023, 10:51 Hi all, just a quick question if I may, got myself a teensy for cem pin cracking, I bought the PCB but couldn't find the Bosch CF160 so went with the SN65HVD230DR's instead. Cracked my P2 XC90 no problem in just under 35 minutes.
So the question is regarding time and resistors. I've read that the teensy should be pretty quick, is the time mines taken right or a bit longer than expected? What I wondered is if the resistors I added may have slowed the process? Being as I used the SN65HVD230DR's instead of the CF160's I added the 2 10k resistors from pin 8 to ground and wondering if maybe I didn't need to?....
Brute force goes over numerical space of 3 BCD bytes (from 000000 to 999999), so the time it takes to crack really depends on your CEMs PIN.
The first part (timing attack) used to be much faster initially, but with more CEMs supported and often subpar cracker hardware implementation we learned the algo has to operate often on a very thin signal-to-noise margin, so it works extra hard now and takes more time. The algo can be tuned to a specific CEM model, and it can crack it, for example, in anecdotal 8 seconds, like sirloin did for his P1 CEM, but I don't want to make the software even more complicated for handling such edge cases. Overall, most of the cracker users only do it once in life, for their own CEM, so it really does not matter how much time it spends on cracking the pin, if it's not hours or days.
I've been asked many times to speed the algo up, but such requests usually come from folks, who use the cracker in their private business. Given that my time is not infinite and testing a new version for compatibility even among my 4 CEMs is a nightmare, I'm more than happy with what we have now. In any case, it is an opensource project, so if anybody feels it worth it to burn his time on tuning the algo and cutting the cracking time from 40 minutes to 30, or maybe 20, yet keeping the codebase compatible with everything we support now - go ahead and do it.
I'm very happy and glad it worked so I'm not complaining at all, thanks for all the work and the clarification , happy chap here









