I'm still working with this, testing all the variables and permutations. The loudest metallic sound from the arc is with two MOT primary coils on the same core for the inductor, as long as the input is on the outside of one coil and the output is on the inside of the other, with the other wires tied together. This produces a strong attraction between the two coils each time the arc rattles. When the in and out is on the same side of the coils, they bounce apart when the arc fires, without the metallic sound. When only one coil is connected, and the other is shorted, there's still the bouncing repulsion. Also, using two secondary coils for the inductor never produces the metallic sound.
I set up a cell and connected it to the second primary coil, rather than just shorting the coil, and this produces some pretty big bubbles, but no repulsion between the two coils on the core. The cell is a 3/4" mesh grid around an 1/8" ss rod, and the arc fires around five or six times per second.
When the metallic sounding arc fires, this induces a steep inductive pulse in the second coil. When the shunt arc goes out, the second coil's pulse falls off rapidly. So this resembles Meyer's waveform, with a lower frequency, due to the wide gap between the cell's electrodes.
These steep inductive pulses, or spikes, are sometimes, erroneously, referred to as 'radiant' pulses. In reality, the pulses don't become radiant until they hit a resistance and turn outwards, moving perpendicular to the circuit. Since the water cell has a resistance, does this mean that Meyer's technology uses radiant energy?