So a few weeks ago, I sent out revised versions of two i.MX6 deigns. No substantial changes, mainly production issues. These two designs, previous revisions, have been working in prototypes/testing for two years. The first board came back, worked great. Got the the new ones this week, and I got weirdness. The USB boot connection is showing up as a bad USB device in Windows and Linux.
So I start to debug... and it's still pretty early. The basic stuff, first check all my voltage test points. This is the NP version of the PMIC, using all default voltages, VCOREDIG run to VDDOTP vis a 100K resistor. All voltages look correct except SW4. This should be 1.8V, but I found it coming up 2.62V-2.70V. There's no software in the system, and I have the test boards locked to USB startup anyway (I have a CPLD that lets me adjust things like boot).
Weirder still, if I randomly reset or power cycle, I do occasionally get SW4 coming up 1.8V. That's not enough to get the system behaving normally. It's a mystery. This 1.8V state is not stable -- the board will revert back to the higher voltage in a minutes or two. Some boards don't seem to be able to lock in here for any significant time.
DId verify that my main system power, 3.3V, is stable, though power cycle and any other kind of torture.
I am checking to see if the PMIC on the first board was from the same lot as the second.
Wondering if anyone's seen anything like this....