I’m not sure, but i think that mooshimeter does not know where those settings came from. That auto-mode is probably just in the app, where it “suggests” what should work for the selected range. Setting those manually should not make a difference in this case.

What were you measuring in that sample data? Did the voltage really change from 230 to 100 volts and back? There was also that 0.00 volts, which sounds unlikely, unless your leads were shorted.

One thing popped to my mind while writing this, did you have long wires or large inductive/capacitive loads in your setup? Could it have been some spike while switching that crashed mooshimeter? That measured ~100 volts after crash might be from a floating conductor?

Then there is our all time favorite, almost empty batteries, even if the meter states that you still have tens of percents left.

If there is still no explanation to this behavior and logged values seem to be correct, just sort your spreadsheet by its first column and go on. It seems* that intervals survive those crashes just fine and extra headers are mostly visual problem in raw data.

* – it might be really hard to tell what really happens on a crash and mooshimeter might just assume a short glitch and if it has lost the track of time, set the clock by the last logged line and continue hoping that everything went fine.

Reinserting of the header will occur when the battery is almost at the end.