I had to ask myself if there was any potential for the PSO regarding OU and I concluded that surely there must be with the large parametric change in core permeability. So, this is a first attempt to tap that part of the device.

The pix of the core arrangement below shows that an additional coil S2 and core piece have been added to the output end of the original transformer thus allowing the permeability change in the original P1/S1 core to be seen by secondary S2 and it's core.

Referring to the attached schematic, the idea is to supply S2 with current from Vss via M2 during the rising voltage on C1 when the H field and flux is increasing but not yet at maximum. Ideally, it would be best to ramp the current in S2 during the time when the core permeability is at it's maximum but I've not been able to manage that at this point in time. When the S1 current along with the core flux has increased as compared to the start of the M2 turn on, M2 is turned off when the permeability of S2 has considerably reduced thus supplying a current to Vss that would be larger than S2's input current. The current from the collapsing field of S2 is stored in C2 and then dumped back into S2 and VSS. This is a result of that testing.

PSOb Input scope pix shows the input to the primary of P1 only and although S2 is connected and operating, it's energy will be calculated below. The input energy is 2.435 x 166.7e-6 = 406uJ.

PSOb Output scope pix shows the S2 waveforms, the P1/S1 rms core flux, plus the min and max voltages of C1. Note that the resultant mean current of S2 is -34.41ma. This results in a returned energy to Vss of .03441 x 20 x 166.7e-6 = 115uJ.

Any energy loss in C1 represents all circuit and core losses including the energy in S2. Due to phasing, all the input energy is used to replenish C1 at the beginning of each cycle. So, we see that the energy loss in C1 is (241.9^2-202.8^2) x .0464e-6/2 = 403uJ. Note how close this is to the input energy which IMO indicates considerable interaction in the circuitry.

Using the above info to calculate the apparent COP we have (115e-6+403e-6)/406e-6 = 1.28.

Now IMO it is interesting to study an expanded view of the S2 function which is seen in the last two scope pix. The 1st pix allows us to see the local input energy to S2 and the resultant stored energy in C2 for the collapse of S2. The input energy is .5056 x 20 x 14.08us = 142uJ. The stored energy in C2 is 207.5^2 x .0149e-6/2 = 321uJ!

The 2nd pix shows us that the negative energy returned to Vss is .5797 x 20 x 20.68e-6 = 240uJ. Thus the energy gain in S2 is 240e-6/142e-6 = 1.69. Note that the energy stored in C2 is not utilized very efficiently.

Pm