-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Philipp Hug wrote: >> I linked this to the LED just for convenience. In theory, you could >> probably make the non-charging interval much shorter. >> >> You need to stop the charger when measuring, because a battery that's >> missing or in cut-off would read as being at the float voltage (i.e., >> full) otherwise. >> >> > Can't this be detected otherwise? I just found out that the battery > present check is wrong as it only checks if battery voltage is higher > than bvmlvl in the BVMCTL register.
BVMCTL is set to 2.8V... this "cut off" thing that Werner mentions is a pathological case where the battery deliberately decides to go open circuit in a sticky way because it feels the situation was not good at some point. You have to zap the battery with external voltage to get it to come back. So that is not really detectable state by V+ battery condition. What you can do then actually is try to do an HDQ transaction, but at that point my brain does something similar to the battery in cut-off and says "let's just not do any of this and use Qi method". The idea in Qi is to defer anything fancy like the looping and flashing and whatever until Linux is up, we just try to boot Linux with low power (backlight off) and take care of it in there once rather than going on a mission to really recreate Linux in U-Boot. - -Andy -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org iEYEARECAAYFAkjuAHQACgkQOjLpvpq7dMpSeQCfYGd/Rcg4p+piQwxO9iJ/DPdl 814An3H2TWd/8IPfMPyOt6qNVzdiuYdM =4dNl -----END PGP SIGNATURE----- _______________________________________________ devel mailing list [email protected] https://lists.openmoko.org/mailman/listinfo/devel
