Oil filled ignition coil on rotary converter, about 2,25 kHz ballasted to less than 3A and fed into 1,1mH ingnition coil primary. Output jumps about 3,5cm before tracking along the bakelite coil lid when exceeding ca. 20kV, this is the limitng factor. When driving with continous wave(+ smooth sine wave) internal breakdown is basically not even a risk when far exceeding normal use. About 100w, maybe a little more, was transferred in this test with basically zero heat detectable after several minuntes, that means a thick enough primary, a core capable of handling >2kHz and the secondary, despite measuring 9000 ohms not overheating from output current (unlike those crappy little dental head xrts hehe). With adequate exteriour insulation, and without high risk for internal flashover such a coil should be able to jump >10cm between spikes and burn an arc very similiar to whats seen here with about 400w input at the same frequency. these are not bad!
Over 700V rms @ 2,2khz into the primary. Transformer taken from 1500w turntableless moulinex microwave, fully floating secondary and better end-to-end insulation than most other MOTs. NST like arcs drawn straight off the secondary, no capacitors. Coupling still is not bad considering the cheap core, probably without cobalt content, has far too thick lamination, is welded and shunted whilst transferring over 100w at 50times its normal operation frequency. Output should be about 8kv rms, clamped by pointy 1cm safety gap to protect secondary, although theoretically lots of primary turns could be taken off considerung up to 15v/turn.
Magical, isnĀ“t it?
Ionizing stages, temperature decline, shell construction, internal restriking, electrode burning; almost all the basic effects of an arc are visible.
Very clearly pronounced is also how huge arcs on disconnects (especially with HIGH voltages) maintain a farely constant height due to flashovers within the arc shorting out certain parts and keeping the total length limited. the 20kv mentioned in the beginning are peak to peak no loaf, not RMS transformer voltage,hence the jump so small.
Just skip the first 6 minutes if you actually want to watch this,theyre mostly for documentation. Breaking the record of the most boring video on the internet with every new upload... something better coming soon :)
camera rotor idea by Photon.
transformer is 8,2kV open circuit and 1,3 A shorted; operating point with ballast used here at about 590mA load, which corresponds to bit over 7kV.
about ~4,6kVA input, 100% duty cycle.
Sodium chloride used to color arcs yellow
Theory of operation(most common):
High current through the poorly conducting hardened bearing balls causes them to heat up and expand. Since they are precisly fitted between the inner and outer races, they are almost unable to expand in a radial direction way from the shaft and therefore theoretically expand in only two dimensions. this asymmetric expansion builds up stress, that tries to return the ball to an almost perfectly round geometry. By giving the shaft an initial spin (required) the balls rotate in the given direction and are capable of converting their deformation stress into a momentum causing a torque on the shaft. This concept will likely not work with any other heat source than electricity, since the balls need to be heated more aggresively than the races to keep the clearances in the correct relation. This can also be seen by the motor temporarely deccalerating and coming to a stop, whilst running again after a bit of cooldown time. Although of course severe permanent damage occurs to all parts of the circuit in relative motion to one another, causing more friction and more resistance. Thanks for watching.