Here are two screen captures of Owon vds1022i oscilloscope output
for two of my armv7 -current installs. Both machines have separate
power supplies with dedicated powered hubs for the disks. Both are
running buildworld, goodpower in -j4 and badpower in -j3. 

Goodpower reports 5.09 volts DC level

Badpower reports 5.23 volts DC level

The "goodpower" host seems to have little or no trouble with disk integrity.

The "badpower" host seems to experience subtle corruption but not outright
failures

Apart from the larger-amplitude "grass" on the badpower trace I see little
difference. 

It's worth adding that I'm relatively inexperienced with the Owon vds1022i,
so the possibility of a gross beginner error in making the measurements
isn't out of the question.  
20261023