Preface: I've done a lotta diagnoses/troubleshooting since shooting this vid. I've got some pretty solid theories and a lot of interesting and confirming observations *after* this vid was filmed...

Key factors:

I'm running it off a cheapo car inverter, which appears to output a bipolar square-wave.

The "reversing" of the trace seems to occur when the inverter pulse rises, when it falls, it reverses again [back to normal].

This seems confirmed by the fact that when I load the inverter more [e.g. turning up the scale illumination lights] the width of the "reversal" increases; thus, the inverter is changing its pulse-width to compensate for the load.

This is further confirmed by the fact that when a probe is floating, it picks up Sharp Spikes at each edge of the reversal.

Interestingly, everything acts *almost* normal when the engine's running; the output voltage, as measured by a multimeter on AC, is roughly the same [around 119V] but the PWM[?] output duty-cycle varies, causing the "reversal" to lengthen or shrink.

Despite roughly the same measured output voltage of the inverter, the input-voltage [engine running=~14V, engine off=~11.5V after long wires] also seems to effect voltages internal to the 'scope: when the engine's *not* running, the 'scope screen seems shifted up and left, the "reversal" is large. When the engine's running, things seem *almost* normal, with a *slight* "rreversal"

Also, interestingly, when switched to digital mode, obviously, the "reversal" *can't* happen, as the samples won't reverse in index-numbers. Instead, it shows as a varying sinusoidal-ish voltage, even when the input is set to GND.

This suggests to me that the 'scope's internal grounding may not be adequate for cases like this where the input power is non-sinusoidal. There may be some inadequate power-filtration in different locations; e.g. the digital circuitry may be well-regulated, but the ADC's connection to the analog circuitry may be less-so. It wouldn't be apparent with a sine-wave power-supply input, because filtration to the analog measurement circuitry may be aimed at 60Hz. Yet a square-wave filtered *to* 60Hz would result in a 60Hz wave. I dunno, exactly.

I do know I like this 'scope a lot, so I may ponder a switching power-supply upgrade... There may be some caveats; some functions seem to look at the power source zero-crossings. Especially, e.g. the "Line trigger." 

There also seem to be *multiple* AC taps off the transformer, so I've yet to determine if high-voltage stuff [like the CRT anodes] are derived from DC or the AC input.

In the now... it *might* work well-enough on this inverter while the engine's running to do the measurements I set-out to do [rough-estimate timings for an LCD], if I keep in mind the present functional-oddities.