What Tinsel has called the "super nova" state, is what I'm experiencing when my system gets much brighter, and the amps being drawn also go up. I think that the bulbs can get even brighter than with the grid source, but also burn out the filament in the small bulbs, after a long running.  I just burnt out my last 7 watt indicator bulb, for no other reason. 
What causes this I attribute to a different resonant state. But, I still don't know how to control it.  As it causes a higher draw from the battery or wall adapters, while it lights the bulbs much brighter. The brighter light is nice, but not at the expence of a much higher draw. Maybe we'll find a way to have the brighter light and keep the draw down also, I think this is also possible, but again, not easy to do.  Otherwise, it's no free lunch.

Its interesting if you do this experimenting with halogen bulbs. At what I call the 'fringe points' of reonance, the filaments can literally snap-off inside the bulbs at the two points where the filament posts enter the glass encapsulation. Seen on a scope, these 'fringe points' either side of resonance show high transient spikes and its these I think that are responsible for the apparent increased brightness of the bulbs and the ruptured filament  (but at the expense of higher current). This condition can be maintained as Tinsel shows in his video until the system snaps into a full resonant condition where the waveform becomes sinusoidal and loses the spikes.

 
All the details of just exactly HOW to do it is what is still not known to us.

Ah, that has to be determined by experiment. All the spoons have gone 

  "The thing is how can we harvest it".  Seams like nobody knows...

  "The thing is how can we harvest it".  Seams like nobody knows...

That is what capacitors are for. If you got generator you can make multiple coils around it , short on peaks, rectify and unload into single capacitor. If you make short circuit for short enough period of time the magnets will not be affected at all. Also the next beautiful property of BEMF is, it reacts in same way as electrostatics so if you provide ground and / or antenna, it will absorb even more particles of free charge into circuit... So your capacitor will be charged up even faster.

All this stuff should be common knowledge but obviosly is not and this is beginning of next chapter in physics...

P.S. The next step is to inject external BEMF into circuit and it comes back to page 66 of this thread...

I did a little test to see just what T-1000 was meaning about coil shorting........ The results.
https://www.youtube.com/watch?v=7gM_3Ji4IMg

OMG a good ol' Bedini motor.
I haven't been thinking about them for a long time...

I don't know. The last time that happened when I was thinking about the Steorn Orbo motor, but somebody discouraged me from it.  See here.

  Although people have been spinning magnets around coils for years, or the other way around, I have not seen it done for free, and obtain OU.  Capacitors, or no capacitors. Ismael has been working on this for several years, and...I do hope that he gets it, but Akula has already got it, at least to some degree. So, I'm still going to play that hand. No noisy spinning thingy.
  Romanov may have gotten it too, but is just showing a diagram, and using over 3 amps of input to light a bulb. So...I believe what I see, and leave the rest to my detective friends.  Time will tell, it most always does.

I've been spinning magnets around coils for years and the shorting of coils is nothing new as has been pointed out. Many Bedini fanatics, including myself have experimented with coil shorting. Grum conducted an interesting experiment with coil shorting and reported loading of the rotor. The bottom line is that generator coil shorting does impose a load on the rotor and this affects the running performance of the motor. If a Bedini SG energiser (motor / generator) is used, fitted with an additional generator coil with timed shorting by mechanical switch or semiconductor, it will be seen that whenever energy is collected into a storage capacitor on each rotor magnet pass (connected as shown by Verpies in his posted schematic), rotor speed will reduce. It can be difficult to detect this speed decrease if the rotor is heavy and spinning fast, as it acts as a flywheel that has a considerable amount of energy stored in its mass. The speed of the rotor affects any electrical output that is being taken directly off the motor coil in the form of coil discharge for the purpose of charging a secondary battery. Measuring the energy stored in the cap as a percentage of the energy removed from the rotor is clearly not an easy exercise but the fact that there is a transfer of energy, strongly suggests that the energy stored in the cap is not free. The SG will always seek an equilibrium when running and the interference of any one of its variables will affect all the others.

My belief is that Alvaro has seen a real device performance increase due to the conditioning of his battery bank through 'spike' desulfation. I wager that he will not be able to replace his batteries with capacitors and see the same performance - time will tell.

Mechanical relay shorts for too long.

Not if the rotor speed and resulting period of waveform is matched to speed of mechanical contact for experimental evaluation purposes. It looks like we will just have to agree to disagree.

What is the shorting period as a percentage of the waveform period for good performance when using mosfet switching?