The pictures of MEG-2 appear to show that the device is built on
a core similar to the Honeywell AMCC-1000. The coil forms are now
3/16-inch plexiglas, which provides better high-voltage insulation
than the punched board used previously, and makes it easier for an
observer to see how the output coils are wound. IMO the 2-pin devices
connected to both the battery lead and the output coils are precision
resistors in a TO-220 package (makes for easier mounting). The size
of the wire on the output coil marked "Y" appears to be #18 gauge.
The rectifiers connecting between the output coils and the output
capacitors appear to be similar to International-Rectifier's
"HexFred" devices, which are high-speed, high-voltage rectifiers.
The two resistors connected to the output filter capacitors are
10-watt packages normally mounted to a heat-sink. They are probably
(as Stan has pointed out) used to reduce the surge current when the
device is turned on and also to balance the load between both output
coils. IMO the two capacitors are connected in parallel and are
driving four 100-watt, 120-volt light bulbs mounted on what looks
like a light fixture normally mounted over a bathroom mirror.
The capacitors appear to be the size of units that have a capacity of
10,000 uF and a voltage rating of 450 volts.
Battery #8 probably is providing 12 volts to the controller
circuit while batteries 1, 2, and 5 are providing the power to
the drive coils.
The device is operating at a frequency of about 14.7 kHz, with an
input power of 147 watts (33.19 volts at 4.3 amps) and if both output
coils are delivering similar power levels, an output power of 460 watts
(2 times 500 volts at 0.46 amps). The 'scope pictures show an output
voltage of about 500 volts, which would be just right for driving four
120-volt light bulbs.
Several things are puzzling: there are "N" and "S" marks on the core
with "S" on the top near the drive-coils, which is the opposite of the
polarity stated in the patent. Also, the device is battery powered, and
the 'scopes are powered by plug-in modules isolating them from the power
lines and ground. The size of the output capacitors is far in excess of
what is required for a switching power supply of this type. There is a
fan providing cooling air to the heat-sinks supporting the MOSFET driver
transistors. For an efficient switching circuit, the efficiency would
be better than 95%, meaning that about 7 watts of heat might be generated
in the driver transistors when operating at the levels shown here.
Clearly, this build-up is dissipating more heat than that, or the fan
would not be needed (the pictured heat-sinks are good for at least 10
watts of heat dissipation each).
Is it possible, that for a successful replication, the device must be
isolated from earth ground and to capture the output power usefully,
large capacitors, with a lot of internal charge (and dipolarity), must
be used ?
It would be a hoot if all the experimenters who have built MEGs might
have been successful if they had simply used batteries !!!