oh but i didn't get it for can-fd, just for the double can controller in small package. I don't think i'll be using it for other projects after this one (nxp is more interesting for what i can't do with a dspic)
as far as you know, are there even cars around which uses can fd? maybe some farming equipment, but cars/trucks?
Anyway, for now didn't find no issue. i did find some (obvious) things not mentioned or not so clear in the datasheet, for example regarding the clocks/pll but it's rather obvious once you see inside the functions of any example project.
one thing i hate when reading the datasheet is when they don't mention what the hell a bit does: when i was setting up the pll to get 48 MHz from the 32k oscillator (yeah, yeah but you must do it: because hardware bug the mcu doesn't work reliably if you configure the internal FRC at > 4 MHz and the internal FRC can't be used for the PLL and the xosc pins are shared with user LED and button
and in other mcus like the NXP 54xxx dual core mcus you don't even have the XOSC input, only 32k but at least a pll optimized to work with it)
so, the pll wasn't working until i decided to switch the bit "enable 32k output"
the description for the bit is
1: enable 32k output
0: disable 32k output
you don't say=
i though it would go to the refclock output or something like in PICs, but that's the actual output of the module! clock is unavailable until you enable the output. obvious, but mentioning it somewhere would have been nice.
In the end the clock is a bit jittery, consistent with the crystal jitter, but the accuracy was 48 meg spot on! and my car isn't complaining so....
Interesting note: when i was first configuring the clock i made a typo and set the output at 102 MHz instead of 40
(0xC43 instead of 0x4C3 on the divider) on a pin configured as the refclock output was still good, though a sinewave and the simple toggle a led firmware was still working