Hi Kcore.
Thank you for your time testing.
Ok, those are good news!
This leaves me more confident. I Had to tweak the timings in read and write timeslots, and I was affraid that i might be just tweaking the lib for my own probes or probe cable length, etc.
That proves the opposite.
Regarding the original example, I have the same problem, I only get FFFFF. I don't have any clue about that, but I suspect that might be something about the probe discovery process. In my test, I just skip that, because I'm addressing only one prob.
I've left that aside to first conclude if the actual code work with every prob.
We are now 3 with working due with onewire.
Unfortunately, I can't continue to test why it doesn't work with example, has i told before, I've accidentally burned my due...
My suggestion now, is to transport the discovery process to the function that works and test it( and removing ds.skip() ).
If more people joins this effort, maybe we can get to a final library faster.
My test function was first built a very long time ago based on that example, it should work...
Let us know if you move forward with this! 
Regards,
João