https://hackaday.com/2016/01/25/overunity-free-energy-and-perpetual-motion-the-strange-side-of-youtube/

There are so many ways to define "free energy". Here is the latest juicy tidbit which I have learned which reminds me of what I have already known for a couple of years ...

If capacitance is reduced low enough, let's say in the vicinity of one-tenth of a pico Farad, then this capacitance tends to over-react and amplify energy. Foster's reactance theorem would appear to confirm this by allowing for negative impedance in its definition.

In the field of geology, a non-magnetic substance has a refractive index which just happens to be the square root of its dielectric constant ...

"The refractive index of a nonmagnetic material is the square root of its relative permittivity." -- https://glossary.ametsoc.org/wiki/Relative_permittivity

So, if you're having trouble looking up on the Internet the dielectric constant for a material that you want to think about using in a capacitor, then (in the alternative) you search for the refractive index for that material and infer its dielectric constant.

For example, ...

The refractive index for granite (bedrock) is around 2.75. Its relative permittivity (or, dielectric constant) is the square of 2.75, or approximately: 7.5625.

If a circuit uses a low enough capacitance, such as: one-tenth of a pico Farad, and is fed a scarcity of voltage, such as: one micro volt, and if the remainder of this hypothetical circuit is designed to take advantage of these two conditions, then it may become possible to simulate an over-reactive situation whose reactance grows at logarithmic rates often-times becoming explosive! ...

https://en.wikiversity.org/wiki/File:Brubaker6h4_output-graphic_1ks%2B53s.png

https://en.wikiversity.org/wiki/File:Brubaker6h4_schematic.png

I'll take a guess that a sizable portion of the real power approaching a low-level capacitor of one-tenth pico Farad does not get absorbed and subsequently discharged (after a suitable time-delay). Instead, it gets converted into imaginary power and, thus, is allowed to refract through what would otherwise be impenetrable to real power while being clearly permeable for imaginary power, namely: an insulating medium of a capacitor's dielectric. The continuous conversion of a portion of real power refracting through this type of low-level capacitance has multiple opportunities per half-cycle of oscillations to accumulate more and more reactive power within this type of circuit at a rate which could be greater than the thermodynamic rate of its conversion back into real power as heat, etc, and -thus- become an overunity situation which trumps entropic losses per unit of time.

This is how I view overunity, as a: race against entropic time.