There are a lot of self-running magnetic motor ideas that have been shown on the internet and as far as I am aware of there is only 1 video of one running on Youtube and its complicated and not very convincing. There are many models and different designs presented. I recently found a patent by Charles Parker  which is an electrostatic motor which gave me a good idea on how to make an extremely simple and powerful magnetic and/or electrostatic generator. In Part 1 I will explain the magnetic generator and then in Part 2 I will explain the electrostatic generator.
I modeled the magnetic and electrostatic generators on a 2D electromagnetic modeling program called FEMM. These analyses are simplistic but definitely show that there is a strong directional force and torque which will allow the rotor to spin. There are numerous patents regarding the use of magnetic bearings as well which would allow a near-frictionless movement of the rotor. Given this fact it becomes obvious this idea should work.
I have a mathematical theory behind why I think there would be a resonant velocity of the rotor magnets which I will cover later. I also think the law of conservation of energy is not violated. The reason why is that it takes heat energy and electrical energy to create a magnet or an electret in the first place so extracting energy over time from the magnet (electrets) until it loses its magnetic field (electric field) is probably less energy than was put in initially.
The first figure shows the basic idea in which it is evident that there would be a constant attraction and repulsion for the whole 360 degree rotation. This constant torque would accelerate the rotor until the friction force and magnetic dragging effects of the stator (and possibly the coils) on the rotor slow the acceleration to a resonant speed of the generator. The generator coils in theory would generate DC pulses of power from the flux change in the stator and the dragging of field lines across the turns in the coils.
Figure 2 (B-field)
Figure 3 (H-field)
In figures 2 and 3 we see the 2D model I made of the top view of the generator. Figure 2 shows the magnetic flux density (B) in colors with the magnetic field lines. The model is with ceramic magnets the program can do a block integral of the inner magnet which showed that there is a strong directional force in the x-direction to the left. Figure 3 is the magnetic field intensity (H) which has the characteristic of having a slightly larger value on stator magnets behind the rotor magnet and a slightly lower value on the stator magnets in front of the rotor magnet.
Figure 4 (B-field)
Figure 5 (H-field)
Here in figures 4 and 5 we see a side view of the top stator magnets which show a similar effect in the field effects of the field intensity (H). Notice that the stator magnets in this view do not show any field lines. This is because the field lines would be forced into and out of the page (the axis not shown) because of repulsion.
While each model is simplistic the overall picture is that if both sides of the magnet stators and top stator where used the force would be very strong on the rotor magnet. To analyze how there could be a resonant speed to the rotor I think that the magnetic pressure equation  and the basic equation of momentum  would work. In this analysis I will assume that there is no friction force. Having a magnetic bearing system where the rotor could be floating makes the idea of zero friction force practically realistic.
The first thing to show is that the magnetic pressure equation can be written as a magnetic force using the simple idea that pressure is a force per area (P = F/S). So we have:Eq. 1 : where, F_B = force of the rotor magnet, S = area of effective magnetic flux, B = magnetic flux density, and mu_0 = permeability of space
Next we have the momentum (p=mv) and since momentum is the time integral of the sum of forces we get:Eq. 2: where, m = the mass of the rotor magnet, and F_D = magnetic drag from stator coils and magents
As evidenced from equation 1 we see that the magnetic field can create a torque force and a dragging force. Thus in equation 2 we see a resonant velocity is created via the momentum of the rotor and the interaction of forces.
The electrostatic version of this idea is just like Parkers patent in which there is tubes for stators with ball rotors but instead of using metal with a charge put on it an electret is used so that it is just like the magnetic version I show in figure 1. I would also add the side stators in this version too. For a better understanding of this look at figure 6&7 which is a side-view of the electrostatic generator.
Figure 6 (Voltage is shown)
In figure 7 we see a drawing from the patent which shows that extra charge can be created by brushes between the rotor and stator. To quote the patent: “…outside cylinder 16 is provided with a brush-like inner lining 35 which transfers electrical charge along radially directed bristles to receiver ring 36 comprised of conductive spacers 39 coextensively aligned with shaft 15. This combined effect creates greater charge build-up on receiver ring 36 for interaction with the drive assemblies 13.”
From this idea we can make an analogy to the equations as well  where equation 1 would become:
And equation 2 would be:
In conclusion I can imagine that a combination of both ideas in a single generator could be made but would be pretty pointless since magnetism Is far more powerful and this device would be easier to build anyway. A linear version of either idea can be conceived in which a self propelled train-like device is possible. Unless I’m missing something very subtle I don’t see why these devices wouldn’t work.
Since the magnetic generator is easy to construct, a small model with just the side stators could be a nice demonstration of the idea. I hope to do this soon.
 US4225801 Charles Parker, electrostatic motor
 Handbook of Engineering Electromagnetics, Rajeev Bansal, pg.157. 2004, Marcel Deker Inc.