I thought about the different configurations of the homopolar generator. Asking the right questions helps to lead me in a particular design direction. I will present the information I've found online, my questions, and my current (if partial) answers which may or may not be correct. I would appreciate more complete answers if anyone has an idea so that I can fully understand alternative constructions.  Coil design. At first, I wanted to use one of Tesla's "http://www.tfcbooks.com/patents/coil.htm" [Broken] (bifilar coils) to amplify the magnetic field of a given current because he stated that the "capacity" of the coil was increased. However, the actual goal of the coil design seems to be to eliminate its self-inductance at higher AC frequencies. Therefore, I don't suppose it would be an effective electromagnet. [Q1.a] Is the coil design supposed to increase or eliminate the magnetic field of the coil? Parallel bifilar should increase it, and antiparallel bifilar should eliminate it. [Q1.b] If it is increased, by how much is it increased (250,000 times as stated in the patent)? Does this refer to magnetic field strength amplification?  Electromagnets vs. Permanent magnets, field strength and power required. [Q2.a] What is the limitation in field strength of a permanent magnet? Neodynium seems to be at about 1.2T. Requires no power to maintain. [Q2.b] What is the limitation in field strength of an electromagnet with an air core, such as Tesla's design above? Normally with an air core you'll never drive enough current or coil it enough to get an appreciable field strength. If Tesla's coil were to magnify the effect by 250,000 times somehow, then it would make an air core viable because you can push the generator current through the electromagnets as part of the circuit and possibly exceed the field created through other methods. Requires no extra power if it can be done this way, otherwise not a vialbe option. [Q2.c] What is the limitation in field strength of electromagnets with a core? Some materials can reach 1.8T with an applied field of only 0.05 gauss, such as "http://www.metglas.com/PRODUCTS/page5_1_2_1.htm" [Broken]. This means that 0.1 amps of current and a reasonable number of turns of coil can fully saturate it. Requires an insignificant amount of power to reach field strengths above most permanent magnets. [Q2.d] Electromagnets have the advantage that their fields can be adjusted. It's easier to put them into position initially when they aren't turned on. The high current in the generator can be run through the coils perhaps to increase intensity.  Geometry. Attached is a picture of a program I wrote to test the relative field strength and vectors given different coil geometies. I found that two solenoids or flat coils do not create uniform magnetic fields along the surface of a disk between them, but it's fairly close at long as the disk is smaller than the coils. It appears that two hemisphere-shaped coils (as shown in the picture approximately) can create a uniform field across the disk surface. Another example is shown "http://web.mit.edu/6.013_book/www/chapter8/ch8-853.gif" [Broken]. This is a relatively complex thing to accomplish, and doesn't suit well to using a core. If [Q2.b] is true, then it might be a good option. [Q3.a] Are some of the properties of the generator derived from the non-uniformity of the magnetic field on the disk? "http://www.stardrivedevice.com/Tesla_notes.html" [Broken] on the device kind of states that when only a part of the disk is covered by magnetic fields, it can be made self-exciting. However, it will function a bit differently when the whole surface is covered, and perhaps especially if the field is more uniform. [Q3.b] Consider Tesla's notes (link above). Examine figure 5. Tesla states that the current can be drawn from the disk, through what is essentially a coil, and out into the external circuit. In other words, you can force the current to create a field in alignment with the original magnetic field. Taken one step further, you can power the electromagnets this way.  Configuration. According to one "http://www.stardrivedevice.com/over-unity.html" [Broken], its author believes that from Tesla's design notes: "eddy currents will be confined to those radial sectors of the disk which do not lie directly between the shaft and an outer pickup brush". Reducing or eliminating the back-torque/counterforce from the system is really the entire premise for over-unity operation; the other is that the power derived from the induction of the magnetic field in the disk increases more rapidly than the power consumption of the motor used to drive the shaft and rotate the disk. The power used to generate the magnetic field is little or none depending on design, and so isn't the limiting factor probably. [Q4.a] The same author stated that input power for rotation increases as radius^2 while power generated increases as radius^4. Why would this be true? If it is true, what strength of field do we need to reach before o-ratio*r^4 > i-ratio*r^2 and we benefit? Can this field strength be created under normal conditions? If not, would it be so much energy as to eliminate any gains in the system? [Q4.b] If back-torque is eliminated through the configuration of the current draw, then we can use more than one approach. 1) Tesla suggested splitting the disk into radial sections. 2) Tesla also suggested forcing a few turns of wire in the right direction would do the same thing. 3) "http://www.stardrivedevice.com/over-unity.html" [Broken]) rather than a copper disk. The proper relationship between magnetic field, direction of coil turns, and direction of coil rotation should yield a self-adding effect to the coil and prevent the eddy current losses at the same time. The coil forces the current to move "radially" as in Figure 4 or 5 of Tesla's notes as it is induced. [Q4.c] The coil design can be two electromagnets with cores and a flat coil between them that rotates. It could also be two permanent magnets or magnet arrays with the same coil in the middle. Tests would include spinning the coils all together, or just the center coil to see what the result is. Multiple inner coils might be tried with different exit points to see if "symmetrical current draw" can help. Perhaps even small wires ("spokes") arranged around the axis would work. [Q4.d] You want a design where applied current to the rotor produces no movement. Then you've got a better chance of having the correct generator action. [Q4.e] Why do magnets connected to a disk create a situation where there seems to be no counterforce?