From: Bruce Fortnam <bfortnam@GlobalLocate.com> - 2007-06-19 17:42:30
>Bruce Fortnam wrote:
>> To Craig or someone at gumstix:
>> I've been trying to implement sleep mode on a basix board, and it=20
>> appears that the PWR_EN pin of the PXA255 is not connected to the VCC
>> This will make sleep mode impossible.
>> Could you please confirm?
>> Is there a via or test point where I can access PWR_EN so I can hook=20
>> it up?
>I, too, am trying to put the PXA255 on the Basix into sleep mode. I
was finally able to put it into sleep and
>wake it up through an >RTC alarm, but the preferred way of leaving
sleep mode is through a level change on GPIO 0-15.
>To add to Bruce's question, none of> these pins seem to be available
through the 60-pin connector.
>Is there anywhere on the board that I can tap into any one of these>
gumstix has chosen not to fully implement sleep mode - specifically,
they didn't connect the PWR_EN pin to the switching power supply that
powers the CPU core. This means that the CPU will enter the sleep
state, but the low power consumption claimed in the data sheet won't be
achieved, since the 1.3V VCC supply stays on. The VCC current for this
condition is unspecified in the data sheet.
Perhaps Craig would be willing to share his thoughts on this?
The lowest current I've been able to achieve so far is about 10 mA
total, while I expected something less than 5 mA, dominated by the
self-refresh current of the SDRAMs. I might still have some GPIO sleep
state settings wrong, or the excess could be the VCC current.
I'm using a trick to get at one of the wakeup pins: you can 'borrow'
GPIO8 from the MMC slot to provide the wakeup signal.
This requires adding one wire to the basix board, plus some external
logic. It seems to work, but we're having trouble getting the OS
restarted. Would you be willing to share your wakeup code to help out?
I'd like to make sure the wakeup logic works without disrupting the MMC
before I publish it.