The reversal of one pole of the magnetic assembly, does not mean that the magnetic field in the middle of the gap will be reversed, too. It means that in the middle of the gap, the flux will be turned perpendicularly to the axis of the entire device (radially).
Yes, but this is accomplished by making both poles of equal polarity and strength (necessitating reversing the polarity of one pole).
The present experiment (with the large DC coil) is for the purpose of determining the required ampere turns necessary to make the opposing poles identical in polarity and strength.
With the large DC coil turned off, and with the measurement probe oriented so that its detection axis is parallel to the long axis of the sausage, the probe will show a certain B field strength and polarity (as if the probe is between two magnets with opposite polarities facing each other).
While maintaining that same measurement probe orientation, when the large DC coil is activated (assuming the required ampere turns are applied), the probe will measure minimal or no field strength, and rotation of the probe 180 degrees about its long axis would show minimal change because both pole faces are now the same polarity and strength (as if the probe is between two magnets of identical strength and with like poles facing each other)
Ultimately this means reversing the polarity of one pole piece (via the large DC coil or ultimately with the high speed coil).
I would think the radial flux you are wanting to achieve would not be very homogenous over a very large area within the sample (likely only ring shaped zones of equal strength), and if so, this would greatly broaden the bandwidth of the required NMR frequency throughout the "fuel" (i.e., different areas will have a different resonant frequency). I'm no NMR expert, but I question whether phase alignment of the precessions or spin flip saturation can be achieved using the field modulation you propose.
Typically, once the B field is applied to a sample, the perpendicular RF field strength necessary to phase align the precessions is fairly low, with a bit more field strength required to achieve spin flip and moreso for spin flip saturation. Rather than having to increase the RF field strength, the use of a very stable and homogenous B field and a stable RF source can allow spin flip saturation to be achieved by applying the RF field for a longer period of time instead of increasing its strength.
I had assumed that you wanted to achieve the highest energy level within the sample, i.e., precession synchronization and spin flip saturation, using NMR, and then use a second probing frequency or impulse of some sort to couple additional energy/harmonics into that state. That does not seem to be the case with this setup.
It will be interesting to see where this goes. If you feel that fast electrons will indeed be emitted and that these will induce/cause a usable current flow (as opposed to emiting beta rays, x-rays, etc), perhaps consider using aluminized/metalized dielectric film (mylar, etc) cut out and stacked similar to a bitter magnet. The last layers of the stack can be cut to have tabs extending out of the "fuel" area so that you can probe for current flow. Use of thin, insulated layers may be sufficient to reduce eddy currents so that a slot does not need to be cut into the "fuel". Additionally, the sections could be cut so that "fuel" only exists where your "cyclotronic" track is expected to be.
PW