Mr.Sure said:
Wouldn't this problem have 3 unknowns and 2 equations? Also, my teacher just sent the answer. I don't understand how he did it. The steps he used to solve the circuit contradict what's written in my textbook.
http://i.imgur.com/Pb8Zr1C.png
No, the mesh equations have two unknowns: the currents in the two lower loops. The current in the upper loop is entirely determined by the current in the lower right loop, since its "sense" resistor is located there. It's a straightforward substitution as you write the loop equations.
The "by the book" procedure for finding the Thevenin model is to: 1) remove the load and find the open circuit voltage presented at the open terminals, and 2) Either determine the equivalent resistance with the sources suppressed, or short-circuit the output and determine the short-circuit current, with the Thevenin resistance will be given by ##V_{th}/I_{short-circuit}##. All of the above can be foiled by dependent sources that get disrupted by interfering with the load.
In order to find Rth your teacher had to insert the external 1A source in place of the load in order to "stimulate" the circuit's dependent source, a procedure I alluded to before. By leaving the circuit intact and solving the mesh equations you can avoid that...
If you think about the form of a Thevenin equivalent model with a load attached, it forms a classic voltage divider: a voltage source with a series resistance and a load. The equation representing such a situation is
$$ V_{load} = V_{th}\frac{R_L}{R_L + R_{th}}$$
After you solve the mesh equation for the current through the load you multiply by the load resistor R to find the load voltage. The resulting expression can be arranged in the same form as the voltage divider equation above, and the Thevenin voltage and resistance read directly from the equation.
You could also use nodal analysis. Any method that will let you solve for the voltage across the load resistor R will do. I picked mesh analysis largely due to experience... the load current through R is identical to one of the mesh currents, so no fiddling about later to calculate the current from node potentials.