The cell from Run 2, unchanged since Run 2, was employed for these tests. Warm-up electrolysis was performed until the cell temperature was about 65C and then the voltage was turned up to 150V, where the W cathode glowed orange hot and the edges of the sheet were brightly lit with dancing white flashes.

This photo shows the cathode (now heavily eroded down from its original 5 mm x 10 mm rectangular shape) during incandescence. Due to the exposure time, the dancing white flashes have smeared together to make the edges look uniformly white. However, in the upper central area you can see the W metal glowing red-orange.

This trace was obtained by placing the scope probe directly across the cell. The vertical scale is 50v/div so the nominal 150VDC signal is 3 divisions above ground. The time base is 500ns/div. There is substantial jitter to the trace, which arises concurrently with the onset of sparking-incandescence in the cell. However, it should be noted that the jitter is largely symmetrical about 150V, indicating that the supply is not drooping badly during the sparks,

This is the same signal at 10microseconds/div.

This is the same signal with the scope AC-coupled at 50ns/div and 20V/div. Now you can see the nature of the "jitter". It is an oscillation at about 20MHz.

This trace was taken with a Fluke AC/DC current probe (model 80i-110s) clamped around one of the leads going to the cell. This probe uses a Hall effect sensor in a split iron core arrangement to sense both AC and DC currents. It has a bandwidth of 100 kHz. The vertical scale is 1A/div and the time base is 1ms/div. You can see that the current signal ranges from 0 to 2 amperes (averaging about 0.7 amps).

This current trace was taken by placing a voltage probe across a non-inductive 1 ohm resistor in series with the cell. The vertical scale is 2A/div and the time base is 100 ns/div. This arrangement can detect faster spikes than the Fluke clamp-on probe and, indeed, there appear to be such spikes. Peaks of up to 6 amps can be seen in this trace.

This last trace is the same signal as before but the time base is 50ns/div and the vertical scale is 10A/div and the trigger level was adjusted to capture the very highest current spike that occasionally occurs during incandescence. It is about 24 amps. Note that the peak does not show signs of clipping or saturation.