Weekly Engineering News 36/2008

Werner Almesberger werner at openmoko.org
Wed Sep 10 22:07:28 CEST 2008

Mike (mwester) wrote:
> I'll second that sentiment.  Frankly, my impression of the 50633 and
> Linux is that while the chip might be excellent hardware-wise, it has a
> serious "impedance mismatch" when it comes to interfacing with the Linux
> kernel.

Hmm, not really. Once we can actually do something with it, it's
a nice chip. The main problem is that its power-up configuration
cannot be changed according to application needs.

So the impedance mismatch is not between the PMU and Linux but
between high-volume production (where you can afford to ask NXP
for a custom-designed configuration) and the volumes we can
commit to.

Oh, Andy, you once suggested for a hypothetical MPU-assisted PMU
design to put a FET after regulators that could potentially be
problematic, such as AUTOOUTx -> IO_3V3. A more comprehensive
solution would place the FET between SYSx and VB_SYS, with
C1767 on the VB_SYS side, and the MPU on the SYSx side.

This way, we could control all rails with just one FET.

Damn, I really want that MPU.

- Werner

More information about the devel mailing list