I read on some power bank description:
Stable AC Output: The built-in AC wall outlet uses a pure sine wave inverter to reduce audible and electrical noise in fans, lights, and other sensitive appliances.
Why does using a pure sine wave inverter help reduce audible and electrical noise in some electrical appliances?
Two things:
Pure sine wave is in contrast to "modified sine" or other switching or chopped waveforms, which contain lots of harmonics (higher frequencies). A sine wave is the smoothest periodic wave, in a sense.
Many devices (motors, transformers, even some lights) vibrate in response to applied voltage or current. This can be due to simple electromagnetic force (Lenz's law), like the windings in a transformer/motor, or material properties (like the magnetostriction in a transformer core).
While there are ways for equipment to generate harmonics or other frequencies, they're somewhat uncommon, or are pathological cases. Example: a motor is unbalanced, generating mechanical vibration (a low humming, the fundamental frequency), which excites some poorly fitted panels that bang against each other, producing a buzz or even more annoying sound. (Admittedly, it's no challenge to create such a pathological case -- but I mean among commercial equipment you're likely to buy... or want to buy anyway, heh.)
Higher frequencies are generally easier to hear, partly by the ear's response, partly because they're emitted more strongly from equipment of typical size.
Surprisingly, this can happen even at fairly low currents; I recall an example, a chandelier populated with incandescent lamps (which are just coils of very fine wire, ultimately). With the light dimmer set at mid or low level, the strongly pulsating current (the dimmer used phase control, shutting off a a whole portion of the sine wave every half-cycle) was audible as a faint, almost chime-like sort of tone. Those were, I believe, 120V 100W bulbs, so the peak current was less than an ampere, generating very little magnetic force in the small coils -- but nonetheless enough to still be faintly audible.
Many devices are designed in a manner that will store energy during part of each AC cycle and return some energy to the line during a later part. If the line voltage waveform is sinusoidal, energy can flow smoothly back and forth. If the line voltage is a "modified sine wave", however, this process will often be rather "jerky".
For example, if one connects a capacitor across a 60Hz power main that is driven with a sinusoid voltage wave form, the capacitor will hold maximum energy at the positive and negative voltage peaks, and zero energy when the voltage is zero. It will take 1/240 of a second for the capacitor to fill up, and another 1/240 of a second for it to discharge. The maximum power when driven using a sinusoidal waveform will match what would have been required to transfer the energy in about 1/376 second (1/240 times 2/π), but that's less than a factor of two over the waveform that would minimize the maximum power rate.
If, however, the capacitor were driven solidly with a typical "modified sine wave", then the capacitor would sit with zero energy for awhile, then almost instantaneously be charged, then sit at maximum energy for awhile, then get almost instantaneously discharged, etc. During the charge and discharge cycles, the capacitor would likely attempt to supply or return as much power as the mains supply and caps could handle, which--unless the mains supply is very wimpy--would be far in excess of what would have happened using a sinusoidal waveform, and would likely be high enough to cause severe stress on the mains supply, cap, or both. Additionally, such high peak currents may cause devices like transformers to "whine" audibly.
