You can really start at either the top or the bottom.
First make up your mind of exactly how it should look before you even start the calculations. What do you want to hang where. Then you should know their masses and all that is left to do is to find where the pivot should be on each arm.
I don't know what you mean by:
Fnet = Fa + Fb + Fg
So you need to be more descriptive. However, I'll make an assumption and go from there, if my assumption is wrong then disregard what I say next and I will correct it when you provide me more details.
Assuming that you are saying that Fa is the force on side A of the arm and Fb is the force on side B of the arm, then that is all you need. But the fact that you wrote it Fa+Fb+Fg tells me that you don't know how to find Fa and Fb. Because Fa and Fb ARE Fg, but on each side. The force pushing down on side A is the force of gravity due to the mass hanging on side A, and similarly for side B.
So you should have:
Fa * A = Fb * B (where A and B are the lengths of the arms on each side of the pivot)
NOTE: Fnet = 0 is the condition for translational equilibrium, which should be true as long as your mobile is not moving around.
The condition from which "Fa*A = Fb*B" is derived is rotational equilibrium, in which the NetTorque = 0.
Since Torque = Force*distance (again the force must be perpendicular) then we know that the net torque (torque from side A and side B) is Fa*A+Fb*B = 0
So really, Fa * A = -Fb * B the negative implies the torque spins the arm in the opposite direction, but it is now on the other side, so physically both forces are pointing down, and you can just simply use the magnitudes (not vectors) and say:
Fa * A = Fb * B
Notice, that this applies to all the mass hanging on the arm at a point, so if you are calculating it for the top arm of the mobile then the force on one side will have to include the mass for all of the arms hanging below it.
Also, I don't know how picky your teacher is, but technically having the pivot not directly in the middle will throw off the rotational equilibrium so you would have to calculate the mass of the portion of the lever arm on each side of the pivots, but that is a bit more difficult, your teacher probably won't mind you treating the lever arms as massless only where torque is concerned, am I right?