So I wend deep into the details on how voltages are generated and learned a few things.
- The main +5V is generated using a MOSFET and a feedback loop with an op amp to drive the gate with a PWM signal. This can go wrong and is the reason why I'm seeing the hiccups on the power supply. When this goes wrong, VBat or Vcharger are fed straight to the 5V rail, taking down chips that are connected there.
- There's another +5V rail for the hard drive, generated outside of the hybrid, but using the exact same circuit, with an op-amp and PWM. It also can fail in the same manner as above and bring VBat or Vcharger to the hard disk 5V rail.
- The battery and the charger are fed through the exact same path, so they are not treated differently. The charger circuit is very simple, Vcharger is simply connected straight to the battery to charge it, and the PMU monitors the voltage until it reaches full charges.
You can fully power your portable to test it using a regulated +5V source, able to supply about 2.5A peak when the hard drive start (700mA when hard drive is spinning, 400mA when computer is on and hard drive off).
1) Connect the 5V source to the heatsink of Q16, this bypass the hybrid regulator and feed 5V straight to the main 5V rail
2) Connect 5V to the tab of either Q14 or Q15 (they are in parallel), this will feed power to the LT1070, which in turn will generate the +12V rail for the hard disk drive and audio
3) Connect 5V to the heatsink of Q20, this bypass the external PWM regulator and feed 5V straight to the hard disk drive 5V rail.
When turning the power, the hard drive will spin as it is not longer controlled by the system. Then the system can be booted by pressing on the keyboard. When booted in the OS, some erratic messages about battery charge may appear as the hybrid is left floating and the PMU is not receiving correct readings.
