Power leakage: save 40mW by turning GSM on

[Andy Green]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Somebody in the thread at some point said:

|    I'm unable to measure any difference in power consumption between
WiFi on
| and WiFi off. In the list archives, I see there has been much discussion
| about WiFi power management. It appears that you can't really turn off
WiFi
| power.

It's a bit of a saga, but the end result seems unbinding the driver from
the device or now they are modular in packaging config, rmmod-ing the
modules should do well for power on andy-tracking HEAD.

Previously there was a pin on the module that looked like an on-off
control, but it turned out it was nothing of the sort.

|    Playing with the openmoko-panel-plugin gives:
|
|    BT on increases current draw by 7,5 mA @ 3,79 V.
|    GPS on increases current draw by 47 mA @ 3,79 V.
|
|    For both BT and GPS, current draw reduces to the previous level by
| turning off the device through openmoko-panel-plugin. On to GSM:

Sounds good, although the GPS side may change its consumption radically
depending on whether it has acquired the satellites or not.

|    GSM off further increases current draw by 17 mA @ 3,80 V. :-(
|    GSM on now _decreases_ current draw by 17 mA @ 3,80 V.

Lol.

|    I reboot with 'shutdown -r now', hold AUX to get into NOR u-boot
and boot
| from SD-card. That also recovers the 2,2 mA initially drawn by GSM.

All these things with closed firmware and private state are full of
surprises.

Do we know what the panel-plugin sources are doing to turn it off?
Sending AT commands, or using the /sys, or both?

- -Andy

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org

iEYEARECAAYFAkl1r0MACgkQOjLpvpq7dMppAQCghfMmZT1CpxWIHHYA/MO72e5z
daQAn1pNhgo/sMIMYWGwk8AC1m/HpkgW
=XYhC
-----END PGP SIGNATURE-----