Steven Mark's has said a number of times that having the electronics outside of the TPU causes the timing to go off - that for some months they could not figure out why their designs wouldn't work until they put that design in the center of the coil. In other words they were working on a design that did not require a direct interaction with the coils via magnetic pickups, such as hall effects or direct impression of the field on the FETS. If they were, the monitoring of the magnetic field, and the output of any sensor monitoring the magnetic field, would have been an obvious first point to check.

Why should this be the case ?

After all the TPU, by Mark's own admission contains a lot of hash, which in turn should be plenty of hashed RF around the coil both within the circumference of the coil (where the electronics need to be placed) and outside the coil. Indeed there should be more magnetic disturbance within the coil.

So why is this a good place for the electronics ?

What ways are there of creating a 'magnetically quiet' area within the center of the toroid ?

Mumetal, iron (possibly), or another magnetic field!. If you imagine a single conductor, for a single turn, going round the inside circumference of the TPU, and then passed a very high DC current into this wire, this would generate a magnetic field which would cancel within the center of the TPU. IF on this single wire you also had a rotating standing wave, you would have a rotating magnetic field, but still have a quiet areas within the center of the toroid, particularly if the rotation was very fast.

Intercept this rotating magnetic field with a toroidally would coil around the whole TPU and you would get high voltage, high current DC output with a high frequency ripple. This is exactly what Mark's talks about.

Now off course this is no where near the whole story but it is possible to work from both ends of what is known about the TPU. From both the kicks perspective and from the known factors about the output.

Someone drew (Innovation Station/Giantkiller) a diagram in the form of a trianagle that went:

Electricity -> Magnetism ->Radiant Event -> Magnetism -> Electricity

I'd like to refine this with some more suggested details:

1. Low voltage pulses.
2. Combining together to high voltage pulses.
3. Getting sufficient transient voltage to saturate an air coil transformer (high, high volts) resulting in a sudden current pulse due to the primary instantaneously appearing as a short.
4. This high voltage, **high current** short rise time pulse is what is required to generate a substantial radiant electrical event, according to Tesla.
5. Assume that the radiant event is a charged particle - Tesla seemed to observe as such.
6. Cause the particles to rotate (A pair of holmholz coils would do this - and coincidently  the top and bottom coils of the TPU look just like these !)
7. Rotating charged particles cause a rotating magnetic field just like a current in a single conductor... and then it's as described before - generation of rippled DC.
8. Feedback the DC into the circuit to generate more radiant events. Feedback too much energy and you have a runwaway circuit - too many radiant events - even higher voltages generated - stronger radiant events etc.