Typically a small generator (alternator, technically) to provide AC power consists of a rotating magnet and stationary stator coils that produce the electricty. If the rotating magnet is an electromagnet, DC power is supplied to it via sliprings. The amount of current consumed by the rotor magnet is
relatively small, and by varrying the voltage to the magnet (using simple semiconductors) the output voltage of the whole generator can be precisely controlled (automotive alternators work exactly like this).
The output frequency of such a generator is directly related to shaft RPM. For a generator used to provide household AC power the frequency must be maintained very close to 60 (or 50) Hz.
This is both difficult and inefficient. Whenever the load fluctuates severely, the RPM of the driving engine may mometarily fluctuate, causing frequency error. And also whenever the electrical load is very low the engine must still remain at the normal RPM, even though barely any work is being performed by it.
I propose replacing the "single pole" rotor magnet with a three-pole arraingment. This would require one additional slipring. If the RPM's are at the normal correct value, only DC would be sent to the three pole rotor, ie 100% positive voltage to phase1, and negative 50% voltage to phase 2 and 3. The three pole rotor would in this case be acting like a static magnet
However, if the RPM's of the input shaft were to drop, the "electrical position" of the three phase rotor could be advanced to compensate. For example, if RPM drops caused the generator output to drop by 5HZ, then 5 HZ sinewaves would be applied to the three phase magnet. So even though the rotor was physically spinning at 55HZ, its magnetic field would in effect by spinning at 60HZ!
Now there already exists small "solid state" generators that dont require the engine to maintain RPM's to maintain an output frequency. They are relatively expensive and hard to find in larger wattages. This is becasue they use semiconductors to rectify an alternators output to DC, and then pulse width modulate the DC into whatever output frequency you want. Since the full output power must pass through the semiconductots, they require large and expensive semiconductors.
My design is more efficient, because the semicondcutors only need to be sized large enough to handle the relatively small rotor current.