Rosemary Ainslie circuit demonstration on Saturday March 12th 2011

[TheCell]

It should be checked first to what extent the signal generator supplies power to the circuit. If this Power is not negligible put a high ohmic resistor into the signal line (10k) of the signal generator(as mentioned before).
Continuous operation tests do not make sense if during the whole time the signal generator acts as an energy supply.
IMHO it is a battery pulsing circuit , similar to the feg of bedini, or that device that richard willis demonstrates. Remember that guy pete sumaruck, who looped a battery via  an inverter, that powers a 60 Watt lamp and a charger, charging this battery . Likely to succeed.

[poynt99]

It should be checked first to what extent the signal generator supplies power to the circuit. If this Power is not negligible put a high ohmic resistor into the signal line (10k) of the signal generator(as mentioned before).
Continuous operation tests do not make sense if during the whole time the signal generator acts as an energy supply.

You are on the correct path, however the function generator never supplies much power to the circuit at all. What I believe it does do, and what is required for both the oscillation and transfer of power to the load through Q2-Q5, is provide a low-impedance AC path for the oscillations to ground potential.

How it does this precisely is unknown without having the schematic for that generator, but I believe one possibility might be protection diodes on the output of the generator.

Inserting a 10k in series with the generator will most certainly kill the operation.

.99

[poynt99]

If anyone can find a schematic or Service Manual for the GFG-8216A Function Generator, that will help a lot.

Alternate models in the same family most likely have the same output configuration, so these schematics might apply as well:

GFG-8255A
GFG-8250A
GFG-8219A
GFG-8217A
GFG-8215A

.99

[poynt99]

There are 'nul-points' in Rose going forward with any test until she can state in no uncertain terms which connection diagram (if any) she herself is going by as her reference.

So far she has published three very different schematics, and I think the good folks here deserve to know which one everyone should be referring to when discussing the circuit.

It is clear and obvious from the demonstration video how the actual circuit is connected (if one bothers to look), but so far Rose has not published a schematic diagram that depicts that actual configuration.

Until Rose comes forward and discloses the actual circuit drawing that represents the prototype apparatus she has been testing, there  is 'nul-point' in suggesting she change to real-parallel (which she will unlikely do anyway).

.99

[poynt99]

In the mean time, I have been playing with the simulation, and I believe I have a good enough handle on things to produce a number of results.

There are basically 3 different configurations we can explore:

1) Q1-Q5 in true parallel. An oscillating circuit is possible in this configuration, but the wave forms are somewhat different than what Rose currently has.

2) Q1 as the Gate drive, and Q2-Q5 with G and S pins swapped. This configuration oscillates and produces wave forms that are closer to the actual scope measurements. This is also the current configuration of the actual apparatus as it has been tested. I can demonstrate an average 0 or negative shunt voltage as Rose has. The Vbat and shunt traces are nicely inverted (although lightly offset) from each other. All powers and currents can be demonstrated.

3) Q2-Q5 receive Gate drive, and Q1 is connected with G and S pins swapped. This is the configuration of the last simulation diagram Rose published on her blog. This is not representative of the current build apparatus, but does produce similar results to 2) above. In fact in terms of the simulation, this configuration produces wave forms that are closer to the actual scope traces than in 2). A negative shunt current can easily be demonstrated, along with all the actual powers and currents throughout the circuit.

The one caveat however in all the simulations, has been that the Gate drive resistor must be quite a bit lower than the theoretical 50 Ohms output of the Function Generator. Values can range from 2 Ohms to 22 Ohms (as was used in the last sim posted by Rose).

My simulation of the exact same Simetrix schematic Rose last posted did not want to run with the Gate resistor at 22 Ohms as shown. It required a value of about 5 Ohms in order for oscillation to occur. This could be due a to slightly different IRFPG50 model used in Simetrix compared to PSPice. However, as I mentioned in a previous post, the Function Generator does not supply significant power to the load, it is mainly used as a DC bias for the Gate to enable/disable circuit oscillation by adjusting the MOSFET channel resistance. The FG also appears to be providing for a low-impedance AC path for the high frequency oscillations, and without this path, the circuit will not oscillate and supply significant power to the load.

So I ask Rose or anyone; is there any particular circuit and any particular "effect" you would be interested in seeing?

Just let me know and I will try to produce the desired results. In the mean time, I will be playing a bit more and perhaps posting some of the various results which I feel get to the meat of the matter here.

.99