The configuration is wrong. You have the TIP and 7812 in parallel. You're biasing the transistor on HARD, and the poor regulator is trying to handle a near-dead short back to the supply. The usual configuration is to have the OUTPUT of the regulator drive the transistor's base to bias the transistor to the output voltage you want. Basically, just move the TIP's base from the regulator's input to it's output. This makes for a pass-transistor power circuit. You'll need a bias resistor between the output and the base. This is largely a matter of finding the transistor's gain, then the amount of current you are using for the load. Let's say the gain is 100 and the load current is 5 Amps. This means the TIP's base current will need to be 5/100, or 50 mA, to get 5A out. This makes the base resistor 12V/50mA, which is 240 Ohms. The power dissipated in the resistor is I*I*R, or 0.05 * 0.05 * 240, or 0.6 Watts. Plan on using a 1 Watt resistor for the base resistor! And lotsa heat sinking for the transistor!
Now, as for the drive motor loads, if your talking the motors that turned the drive's platters, plan on at least 3 Amps each on power up, and considerably less at full speed, IF you're turning something as massive as a hard drive's platter set. In the end, only inserting an ammeter in line between the supply and the motors will actually tell you what's happening. What would be better is to insert a 1 Ohm power resistor in the current path and applying an oscilloscope or data logger across the resistor to measure the initial draw at power up. This eliminates the DMM's slow sampling rate missing the peak current at the moment you apply power.
Umm, that's the best I can do. Good luck with your project.