ISTM that you may be using an inappropriate power source for your PCB. A switchmode PSU needs a minimum load in order to regulate properly. At light loads, such as a single HDD PCB, the voltages may not be stable. For example, if I use the following PSU (for an external USB enclosure) with a single HDD PCB, then the +12V rail sits at +13.4V and the +5V rail sits at +4.75V. This means that the 12V TVS diode is very close to its breakdown point.
http://azsurplus.com/images/Finecom_XH_34W-12.0-5.0.jpgYour PC PSU will also be unstable at light loads. Therefore I suggest that you use aluminium clad resistors, or some other arrangement, to provide an appropriate minimum load. In my case I install the resistors inside the PSU, in the fan's air stream, and I fasten them with captive locknuts that have nylon inserts.
Your observation regarding the heating of the PCB is a little unclear. It is quite safe to touch the components with your finger. The area between the motor controller and Q501 and the SDRAM contains small passive devices, so it should not be generating any significant heat. The SDRAM should be reasonably cool, and the motor controller should be warm at most. It will only heat up when driving the motor, if at all. The primary heat source should be Q501. This transistor essentially powers the entire board (except for the +12V stuff). It will probably be dissipating around 1 watt.
Here is a simple diagram of the positive supplies:
Code:
|--- VR400 ---o +1.8V
|
+5V o--- Q501 --+-------------o +3.3V Vio
|
|--- VR300 ---o +1.66V Vcore
The absence of the -5V supply could be due to several things, including the MCU, motor controller, or the -5V supply itself. That's why I asked you to test for a short circuit between the -5V test point and ground, with and without the HDD attached.
Once you have exhausted the simple DC tests, then you will need an oscilloscope to go into more depth. You could also try to communicate with the board in safe mode using your pro tools.