chiveicrook wrote on 2022-06-07, 12:06:

Output shape is important for modern computer PSUs with active PFC, which require at least good sine approximation with short switchover time in case of power failure. Pure sine wave is naturally the best but not strictly required. Very old UPSes might not provide output that is good enough for active PFC PSUs which, in turn, might result in a wide array of symptoms: restarts during voltage fluctuations, increased load and noise or even catastrophic failure.
In contrast, older equipment with PSUs without PFC could run on really dirty power without trouble and older UPSes were often designed around that.

You are probably not a expert of electrical supplies i can read it.
The power factor circuit (PFC) inside in the PSU is trying to adjust the load (which is current) during the whole the sine period (0.2ms at 50Hz), that the load (current) is always propotional to the voltage. In average you will have the power which is neccessary to supply your PC. This is good for the power grid, because the COS PHI is near 1.0
- What happens if you have non sine-wave (e.g square wave) supply like an "old" ups?
Nothing unusual. The PFC is rectifying the square wave, and it will be an almost constant DC, where not much power-factor-correction necessary. The second stage switch-mode-power-supply is converting down to necessary voltage, and ready.

The switchover between the grid supply and UPS is also not a problem, the PSU has a large enough capacitor in the first stage.

-Why big companies trying to sell Sine-Wave capable UPS devices, if sine wave not necessarly?
Maybe because of backwards compatibility because people are trying to supply all kind of devices via UPS. See end of my previous post.

PentAmd wrote on 2022-06-08, 09:31:

As an electrical engineer I can tell you in 2022 you don't need sine wave for consumer electronics at home. Computers, monitors, […]
Show full quote

Hoping wrote on 2022-06-08, 09:01:

I was shocked when I read that article.
Even brand name UPS turned out to be horrible, high power consumption by themselves and the absence of the sine wave, not even anything close to it, have left me petrified.

As an electrical engineer I can tell you in 2022 you don't need sine wave for consumer electronics at home. Computers, monitors, TVs, led-lamps have Switch-Mode-Power-Supply inside. The SMPS in the first stage rectifying the sine wave, and fills up a capacitor with DC current. For this stage is no matter if you have sine wave or non-sine wave (rectangular).
Sine wave is necessary only if you have a Phase cutting Thyristor (SCR) or Triac circuit, where the power reduction is done that way. I have an old welding machine which work that way.
Even if you have an old device with transformer (like audio tape cassette player, or some old HiFi device), the non-sine wave will be transformed, so the device will work. The transformer of that device maybe heating up a little more, but UPS is not designed for 0-24 usage.

I'm not even remotely an engineer, but I do know that the first part of the conversion to DC is a transformer, a bridge rectifier and a capacitor, but even so, a defective sine wave in an online, line interactive, or double conversion UPS, which are the most expensive ones, it catches my attention and it seems a bit deceitful to me, that although the devices can work like this, it does not seem entirely correct to me, and even more so if this behavior is not clearly specified.
But well, this may be being a bit demanding, I don't know.
But it is true that many cases are known in which strange problems are caused by a poor quality of the electrical supply as another member mentioned, I think that I am not mistaken in this, I think.
Things like background noise in sound amplifiers, or radio frequency wave interference.
As I lack basic knowledge, for example, I do not understand correctly what the PFC is and that is why I am also afraid to say stupid things in front of an engineer.

But again, the problem I see is that it's not made clear in the UPS specifications if the output is a correct sine wave or if it's something that...well...works.

Hoping wrote on 2022-06-08, 10:01:

I'm not even remotely an engineer, but I do know that the first part of the conversion to DC is a transformer, a bridge rectifier and a capacitor

You are wrong in 2022. What you are writing was true in the 80s and 90s electronics. I was a kid in 1996 and I had my first Pentium (from 1994) which came alredy with SMPS, the world was changed.
Get back to your comment: Nowdays the "pipline" looks like so: Choke -> Rectifier -> Capacitor (here we have 315V DC inside the Capacitor in 240V countries) Then a SMPS is converting down to 24V/12V/5V whatsoever.

In PFC circuit, it is changed a litte, it is measurung the current at the rectifier, and the rectifier itself is a transistor with a coil (transformer) to be able to step-up the voltage to the 315V, even if the momentary voltage of the sinus is small.

Well, I'm not sure but I think that even today there are many power supplies without PFCs and in which the efficiency, even if they put stickers on it saying something else, is less than 80% or even 75%.
UPSs also consume energy and the article that was mentioned also mentions the low efficiency of some that I would almost relate to their poor output, I have the feeling that a UPS with a poor output is also an inefficient UPS.
Taking into account what is happening in Europe with energy, I am a bit concerned about efficiency.
But above all, thanks for the explanations.
I have to study more how the different UPS technologies work, offline, online and double conversion, although I have never seen a double conversion one, especially because of its price.
But Inverter technology should be improving because they are also used for solar energy, which seems to be gaining so much importance.
If you buy an inverter for a solar panel installation that does not offer a correct sine wave, there are more likely to be problems, considering what is explained here.
So in the end, I understand that a UPS can be good or bad depending on what it is going to be used for, but again, they should clearly specify it.

Hoping wrote on 2022-06-08, 11:13:

If you buy an inverter for a solar panel installation that does not offer a correct sine wave, there are more likely to be problems, considering what is explained here.

To answer your original question, I suggest to buy a solar system with battery, if the power grid in your area not adequate. The solar power lets your battery live longer, because during the day it is not necessary to dicharge it, so you the charge-discharge cyceles reduces.

PentAmd wrote on 2022-06-08, 09:58:

You are probably not a expert of electrical supplies i can read it. The power factor circuit (PFC) inside in the PSU is trying […]
Show full quote

chiveicrook wrote on 2022-06-07, 12:06:

Output shape is important for modern computer PSUs with active PFC, which require at least good sine approximation with short switchover time in case of power failure. Pure sine wave is naturally the best but not strictly required. Very old UPSes might not provide output that is good enough for active PFC PSUs which, in turn, might result in a wide array of symptoms: restarts during voltage fluctuations, increased load and noise or even catastrophic failure.
In contrast, older equipment with PSUs without PFC could run on really dirty power without trouble and older UPSes were often designed around that.

You are probably not a expert of electrical supplies i can read it.
The power factor circuit (PFC) inside in the PSU is trying to adjust the load (which is current) during the whole the sine period (0.2ms at 50Hz), that the load (current) is always propotional to the voltage. In average you will have the power which is neccessary to supply your PC. This is good for the power grid, because the COS PHI is near 1.0
- What happens if you have non sine-wave (e.g square wave) supply like an "old" ups?
Nothing unusual. The PFC is rectifying the square wave, and it will be an almost constant DC, where not much power-factor-correction necessary. The second stage switch-mode-power-supply is converting down to necessary voltage, and ready.

The switchover between the grid supply and UPS is also not a problem, the PSU has a large enough capacitor in the first stage.

-Why big companies trying to sell Sine-Wave capable UPS devices, if sine wave not necessarly?
Maybe because of backwards compatibility because people are trying to supply all kind of devices via UPS. See end of my previous post.

I must admit I'm not exactly current on electrical power supplies (and computer supplies in particular) so I'm not trying to argue but as an ex-electrical engineer I vaguely remember from my studies that some topologies of active PFC correction definitely dislike square waves due to inductive spikes. Assuming standard practice of using cheapest possible design to pass conformance requirements I'd assume that standard consumer psus have minimal filtering tuned for sine wave input.
Older voltage sensing PFC controller ICs also liked to simply shutdown when supplied with a square wave or during too long of a switchover dead time.

As a source, I'd like to quote US patent 20130297088:

Uninterruptible Power Supply (UPS) systems with nonsinusoidal waveforms (in particular, square waveforms) have been found to be incompatible with a particular class of load characterized as an active power factor correction (PFC) load (i.e., having an active PFC controller).
(...)
Active PFC loads, such as an active PFC controller, are typically designed to operate well under nominal line conditions. However, when an active PFC load is connected to a square wave UPS in battery mode (e.g., in the case of a power outage), the active PFC load will typically start demanding (i.e., consuming) power, resulting in a decrease in the output voltage 137 of the square wave UPS 140. In other words, the load represented by the terminal 130 with active PFC controller draws the current 138 rapidly during this time (from the UPS) in order to recover from the abnormal conditions (power shortage) and maintain a power characteristic (e.g., according to a specified power requirement of the load, as regulated by the active PFC controller). In fact, the current may exceed the maximum tolerance 143, possibly damaging the UPS or causing the UPS to disable its output to protect itself, resulting in dropped load or damage to user equipment.

Interesting, that US patent is a bit concerning, since I'm in Europe, I'm making the assumption that the overvoltage would be relative meaning that here it would be greater.
Tell me that I'm wrong, please.

chiveicrook wrote on 2022-06-08, 09:46:

Audiophiles and HAMs will have a different opinion on this, naturally 😀

Yup. 😁 I suppose because of the hobby, they (we) are among the ones to notice first if something goes wrong.

While I do still have an aversion against switching PSUs (I'm rather using 13.8v lab PSUs with transformers),
I do acknowledge that good switching PSUs can be build if carefully designed.

Unfortunately, not for little money. Most switching-PSUs aside from the big boys, like lab PSUs or ATX PSUs and others, do generate noise.
That's how they work - generate RF (AC) and then rectify it (DC).
Not much unlike a radio transmitter and a crystal radio.

Problem is, that shielding and filtering requires extra work that manufacturers are not willing to spend these days.
A proper switching-PSU needs more filtering perhaps than a traditional PSUs to provide equal purity. Of course, it can be much more efficient, also.

That's one of the core differences between enthusiasts and professionals, I think.
Enthusiasts don't have money on the mind. At least not primarily.
They (we) rather live (try to) in an ideal world based on mutual respect and rationalism.
And the air waves are seen as a common good among all people, also.
There's even a codex for it.

Filters and shielding aren't seen as luxury thus.
They're a long time investment that's worth every penny.
They're a requirement of physics that has to be met, so to say.

..
And then there are "China" models of PSUs..
Some noname or counterfeit phone chargers do even have the "hot side" on the low voltage side.
Lots of teenagers died in the bathtub because the DC power chord of these cheap switching-PSUs touched water (phone fell into bathtub).
With a classic transformer PSU, this wouldn't have been possible so easily due to the built-in galvanic insulation of a classic design.
Except if the ground was mutually connected on both sides of the PSU.
Which makes no sense to do in a plastic chassis, however.
Also, older designs had real fuses installed.
Switching PSUs may or may not have them.
If they don't, they may still have something similar.
Two tiny points on PCB were an controlled spark gap can blow the mains fusebor something along these lines.

Edit: Or let me put it this way, switching-PSUs are the electronic's world equivalent to plastic. Cheap, plentiful, "efficient", easy to make.
Whereas transformer-based PSUs equal wood or metal. 😉