No sane person is saying you "need" a pure sine UPS.
LOL Then there seems to be quite a few insane people around here.
And
@Tahagomizer - I have to assume you didn't bother to read the links I provided for EVGA or Seasonic, or Dr Dro's link to Eaton - all of which clearly said modified waveforms are fine with their PSUs and PCs. I have to assume that or else you must believe you know more than their engineers who you must believe are inept! Otherwise, why would you repeat the same old marketing hype spewed by the marketing weenies of pure sinewave UPS? You can see by the link in my sig I have a bit of experience here and I have no reason to believe I know better than those experts.
No. That is not the problem. The problem is those marketing weenies and opponents of stepped approximation UPS use images like that to illustrate their point, suggesting all UPS output approximated sinewaves that look like that. That is total nonsense.
A much more accurate representation is a modified/stepped/approximated sinewave waveform that looks like this:
Most units take about half wave to switch, so you can go from long zero state to another long zero state, which is too much for many power supplies
Not even. Most units? Nonsense. This is exactly the type of marketing "hype" and "misinformation" being spewed by the marketing weenies of pure sinewave UPS makers.
Half wave? Long zero state to long zero state? Come on people! Snap back to reality, okay?
Yes,
@Tahagomizer is right, technically -
in theory for the most basic (near "squarewave" output) "battery backup" UPS - the type that might be used on a building Exit sign and emergency lighting to illuminate the way out of a dark building after a power outage. But NO ONE is suggesting the use of a basic battery backup UPS on a computer. Come on, people!!! That type of ridiculous argument is why I get riled up over this.
It is, always has been, and will continue to be a "good UPS with AVR". So to suggest "many power supplies" can't deal with such long state events is simply WRONG!!!! Or at the very least, extremely deceptive and totally
unrealistic marketing hype!
How many half-waves are there in 1 second? There are 100 for 50Hz mains and 120 for 60Hz. So when someone says "long state" to "long state", that means 0.01 second or 10ms @ 50Hz, and 0.0083 seconds or 8.3ms @ 60Hz respectively.
To suggest that 1/100th of 1 second is so long it can disrupt many ATX Form Factor power supplies is total nonsense.
Why? Because 1/100th of 1 second is not even reality! As seen in the more realistic image of a stepped approximated waveform, the reality is the "long state" is much MUCH shorter than a half wave.
It is closer to 1ms or even less!!!
Nobody is talking about a $5 counterfeit 烂Deer PSU from some backwoods factory in China that uses underaged, force labor to build their supplies with components from its sister factory upriver.
We are talking about ATX Form Factor compliant PSUs. PSUs that are required to "hold up" power in the event input power drops below ATX defined thresholds for a minimum of 17ms at 80% load, 12ms at 100% load.
Worse condition is 12ms. That is way WAY more than enough to handle that ≤ 1ms "long"
state.
In other words, even 1/2 way decent, budget ATX PSUs are designed to maintain output with a stepped approximated waveform from a
good UPS.
Here's another dose of reality. Who here recommends cheap, budget PSUs? Not me. Never. I always recommend quality supplies from reputable companies.
Same goes with UPS. I always recommend a "good" UPS with AVR (automatic voltage regulation) - pure or stepped.
So to make the argument using cheap PSUs and UPS, and attempting to apply that across the board, is deceptive at best, and plays right into the hands of the marketing weenies, "hook, line and sinker".
but if you have lots of brownouts (power goes out and comes right back)
Sorry, but that is not a brownout. A brownout is a "low" voltage event, often called a long duration "sag". If your mains is normally 120VAC, an example of a brownout would be if your mains voltage drops to 100VAC and sits there for a few seconds or even longer. If the voltage quickly drops to 100V but instantly returns to a normal 120V, that would be a "dip". Dips typically cannot be seen (as flickers) by the human eye, but can still be disruptive to sensitive, unprotected electronics.
When the power goes "out", it drops to 0V. If it immediately comes right back, it might be called a "flicker", though that is not really a proper technical term. A power "blink" is the more accurate term.