Hello Friends.
Using the same setup of the last videos I have played a bit more with some modifications. As my phone had no more memory avaliable , I didn't record a video, but here some more oservations: First i added a second outputcoil and connected parallel an serial to the LED's to check for an improvement changing the impedance, but the result was poorer than with the single coil. A short output coil seems to be more efective. Then I tried with two neo magnets, and the output increased a lot. It has to be remarked, that the frequency has to be adjusted at any change of the magnetic field strength or load impedance. That is the reason why Yoelmicro is using a PLL in his latest tests. Then I tried with two sets of different sized ceramic magnet pairs, varying distance and frequency adjustments, also I compared the two possible alignments of the coils ,which produced an output.
The observations was as follows: the type or shape of the magnets was not important, all worked equal.
Without the magnets there was no power on the output coil, some voltage below the forward voltage of the LED's only.
The output was proportional to the magnetic field strength applied to the inductor, with the ceramic magnets sticking to the ferrite core reaching the maximum, the neo magnets had to be spaced at least 1 cm each side, to avoid oversaturation.
The input power also is proporcional to the applied magnetic field strength.
The best output was achieved when the axis of the inductor coil was aligned with the axis of the output coil.
In this alignment the output frequency was equal to the input frequency(the doubling effect was not present)
In all configurations the output was only present when the core enters in non linear resonance, close to or at saturation.
The maximum output was always when the switching was adjusted exactly at the peaks of the sinewave (longitudinal driven).
With te most efficient configurations the input power was 8.5 watts. On the output I have not yet exact measurement's, the voltage at the forward cycle on the led was over 40v, the two 10 w 33v Leds brightly lit, and a lot of heat in the primary inductor, smelling like toasted varnish and plastic.
Conclusion: The method looks promising, although the effect seems to be closely related to the core of the primary inductor and it's non linear behaviour, maybe not exactly what Floyd Sweet in his coreless SQM (?) did.
It is clear that the load do not directly reflect to the input, but it do increase the reactive power cycling in the primary tank circuit. That is why also the losses increases, but it should be possible to improve significantly by raising the Q-factor of the LC tank.
In my setup the latter is worst, many turns with small gauge wire, therefore a lot of heat losses in the inductor.
Thank you all for your participation,
Vidura.