next-generation (now) MPU discussion / Backup battery on PMU only

[Werner Almesberger]

Andy Green wrote:
> If the battery backup is exhausted then it didn't help anyway compared
> to the MSP430 not having any at all: both ways the MSP430 comes up from
> dead and we still need to work in that case too.

Yes, but again, the backup battery idea only came up as a means of
powering the MPU at all, since it's not 5V-tolerant. As I understood
the backup battery proposal, it wasn't meant to suggest that the MPU
needs to receive power even if Vsys is dead.

Only then, with the assumption of having a backup battery, did this
start to sprout branches and the RTC location was questioned.

So, three cases:

1) Vsys dead, backup batter is good: we don't care what the MPU does
2) Vsys powered, PMU in "Standby": MPU at least has to hold GPIOs
3) Vsys powered, PMU in "Active": MPU does all the little helper things
   we've already discussed

We still have to answer how to power the MPU in case 2. Three ideas:

- add an LDO from Vsys for the MSP430

- pick a different MPU with a wider voltage range

- we give the MPU its own LDO in the PMU that is kept active even in
  PMU.Standby (this has to be configured by software, so this has to be
  an LDO that is enabled by default when going NoPower->Standby->Active
  and it has to have the right voltage. In out PCF50633-04, this would
  be LDO1 or LDO4, both 2.8V, or LDO6 at 3.0V. LDO6 has the lowest
  quiescent current of the three.)

- Werner