A simple coil assembly that may hold potential for a superluminal buffer.
The first experiment to break you in, isn't a practical size. 5 turns of 22 AWG with a 5 cm diameter. Use that coil as a toroid core, and the toroid windings wound on it to place on layer of wire over that first coil. The first coil is the primary and toroid winding is the secondary. Just test it with a signal generator. There are several thing you will be able to note. First, it does not step down a voltage. It works 1:1 from the secondary to using the mentioned primary as a secondary, or it divides down. I know that it doesn't sound effecient in any way. But, what you need to look at is the phase of the secondary in relation to the primary over a range of frequencies to plot a graph. Then you will find that you are not dealing with the self resonance of the coils at all. Without moving the coils you will find several phase shifts that are associated to frequency and wavelength. If you watch carefully, the lead time of the output is always the same, and as frequency increases the more cycles fit into that lead time. The only real problem with that is that when the lead time and the operating frequency result in a 180 degree phase shift the output of the secondary is reduced to nearly 0, at a frequency higher or lower there will be an output with said noted lead time.
I've also been some interesting work with plasmas because of early returns reported by Ham Radio Operators during solar maximum. Well, I can watch that go in reverse. Time on a small scale, or a fraction of second can be pulsed backwards in event. That's a different device and works on different priciples. Let's look at the phase shifts and lead times of this coil first. I haven't drawn any real conclusions because there are many unknowns left. I'm sure that a signal could backtrack several CPU cycles. But, if it does that how difficult will memory management really be? There are several points were it does raise questions.