Modified Sine Wave vs Pure Sine Wave: Impact on Computer Switching Power Supplies

Hello,

I have known for a long time that cheap ups produce a waveform called a modified sine wave, which looks like this:



a follows the normal course:



And I have always heard that such a waveform is harmful for some connected devices, hence my question, is this a modified sine wave also a problem for such a switching power supply in a computer, where the current undergoes many modifications of its waveform? If this is not good for this power supply, then exactly what is "destroying" the power supply, which component gets into the bone, getting an angular sine wave instead of the usual one?

There have been several discussions on the electrode, even more than a few, but I have the impression that what a person is an opinion and no one gives specific examples of the harmfulness of such a course.

ps: the only harmfulness that i understand quite well is when it comes to the work of all kinds of ac motors.

touchthis wrote:

my question, is such a modified sine wave a problem for such a switching power supply in a computer, where the current undergoes many modifications of its waveform?

I'm not a specie. But it doesn't really matter to me. The voltage is rectified anyway before the converter (switching power supply).

@touchthis Can you put the waveform of the current consumed by your computer and put it on the waveform of the "worse" UPS?

freebsd wrote:

@touchthis Can you put the waveform of the current consumed by your computer and put it on the waveform of the "worse" UPS?

Sure, as soon as I get home from work, I'll check the waveforms

Quote:

or for such a switching power supply in a computer, where the current undergoes many modifications of its waveform, is such a modified sine wave also a problem?

It all depends on these "current modifications" in the power supply, when we are talking about an old power supply without PFC, there is a rectifier and a filter capacitor at the input, it will consume the current in pulses also when powered from the mains, the most will change the diode conduction angle and the effective value of the current something it can heat up more or less (NTC thermistor, diodes, capacitor).

When we talk about a power supply with (active) PFC, it would be necessary to analyze how the specific PFC controller that is installed in your power supply will react to such a course, and it depends what its principle of operation is and what simplifications have been adopted by the designers.

Quote:

If this is not good for this power supply, then exactly what is "destroying" the power supply, which component gets into the bone, getting an angular sine wave instead of the usual one?

It can be expected that the PFC will work with a higher peak current value and if it is badly designed, it will either overheat or limit the current, or it will have some other problems, because the designer relied on the assumption that the supply waveform is always sinusoidal.

Quote:

There have been several discussions on the electrode, even more than a few, but I have the impression that what a person is an opinion and no one gives specific examples of the harmfulness of such a course.

I also gave little details, but you have to remember about economic reasons, on the forum and generally on the Internet, people who trade UPS also say that you should always buy the more expensive "because the power supply is adapted to the sine wave and will break down when will get a rectangle or a modified sine wave "and of course he will try to write it several times in all forums that a potential customer may read.

touchthis wrote:

Well, as soon as I get home from work, I'll check the waveforms

I do not have access to a UPS with a "modified sine wave" so I still count on your measurements. However, I tried to record the oscillograms of the computer and laptop. In the laptop, I took out the battery so that it would not affect the power consumption.
Computer with 450W power supply:

Laptop:


While the oscillogram of the laptop looks predictable, I would call the computer's power consumption "crazy"! UPS has a lot to do.

Most modern single-phase household appliances will even work when operated with DC power.
The exception may be washing machines, refrigerators, fans, lawn mowers.
This is due to the fact that these devices use single (two) phase induction motors. Universal motors, such as in drills, will work despite being powered by 230V. The only devices that will not work with the DC grid are those that use grid transformers and thyristors.
Theoretically modified sine wave can power even small one (two) phase induction motors. For me, however, it manifests itself in a more noisy engine operation.

The problem is that most modern electronic equipment has a bridge and an electrolyte at the input.
We give the voltage in the network: effective voltage. And we accept there's a sine wave there.
And there is an "almost" sine wave - it has slightly flattened vertices. And straightening is peak.
Now we let in such rectangles from a simple UPS, which in the output gives: zero, upright rectangle, zero downward rectangle, zero ....... then the maximum voltages of these rectangles are greater than the peak of our network "sine wave". And as in the receiver the electrolyte has a voltage of, for example, 350V, it works on the edge - or beyond the edge
And there is risk.

Buddy freebsd. If there is no sine wave on the osculograms, then the voltage in your network is only to be envied. Is pretty.

akytam wrote:

Now we let in such rectangles from a simple UPS, which in the output gives: zero, rectangle up, zero rectangle down, zero ....... the maximum voltages of these rectangles are greater than the peak of our network "sine wave". And as in the receiver the electrolyte has a voltage of, for example, 350V, it works on the edge - or beyond the edge
And there is risk.

And how ...

The capacitor in the filter charges to the maximum voltage value. The maximum value of a sinusoidal alternating voltage with an RMS value of 230V is 324V. If the amplitude of the square wave is below 324V, there is absolutely no risk of damaging anything.

Sorry for the delay but my estimates of when I will have free time turned out to be grossly wrong, today I went back to checking ups and I have surprising (for me) results. I did a small comparison of a few UPSs that I do not use anymore (I replaced them all with one large 2.5 kilowatt full sine wave operating on a 24V 110Ah battery).

As a benchmark, look at the current straight from my home network:


So I present the waveforms of the tested models (all the ups were loaded with the cable TV modem, the dlink switch and the linksys wifi router - at the same time):

1. UPS Active Power 600:



2. Orvaldi 520PL UPS



3. UPS Ever CDS 500




I was very surprised how nice the ups run ever looks, almost perfect in such a cheap model, wow.

I also wanted to charge these ups with a computer and throw in how the mileage changes, unfortunately only one ups survived the moment my computer was turned on, the rest signaled overloading and turned off, so I would have to buy an older computer and then try.

So yes, I am going to continue this topic.

Buddy touchthis.
Try so that the UPS is powered from 230V = the "bypass" is working, then turn on the PC, and then take the voltage from the UPS. He should switch over and continue to hold on.
Because this is the principle of maintaining the "lost" voltage in the network.
How much is the peak-peak output voltage from UPSs that generate rectangles?

Friend @akytam raised the issue of the appearance of an oscillogram showing the voltage in the electricity network at my home. I would have classified this waveform as distorted (in the previous video it is almost impossible to see at 200V / div), but I also did not see many such oscillograms to be authoritative. I'd love to hear from you other opinions.

I recorded more detailed voltage oscillograms at home:


I don't have a spectrum analyzer, unfortunately .. but using a voltage divider and a sound card, I got the following spectroscopic spectra:


I do not trust these measurements too much, but as a demonstration of what is happening in the mains, they should be enough.

I wonder if the power generated by the "cheap" UPS has an impact on the operation of the computer, and if so what. Can greater distortions, higher harmonics penetrate the computer power supply and affect the stability of the computer? Maybe they interfere with the operation of the power supply itself and, indirectly, the operation of the computer?

High frequency interference will not be affected. 150, 250 and 350Mhz will be decisive.
For an average computer power supply they should not cause anything noticed, in cheaper models they may cause greater voltages instability 12, 5, 3.3V etc.

Out of curiosity, I measured the voltage in my network and it is a little different from the perfect sine wave.


There are 4 harmonics in the spectrum every 100Hz with a value above 1Vrms

akytam wrote:

Now we let in such rectangles from a simple UPS, which in the output gives: zero, upright rectangle, zero downward rectangle, zero ....... then the maximum voltages of these rectangles are greater than the peak of our network "sine wave".

It is not so simple. If zero was for half the time, and after a quarter of the time, there was a maximum positive and negative voltage, then for the same RMS voltage, the maximum would be equal to the maximum sine wave. The thing is, on the two charts, the maximum duration in each direction is 20% of the period, and on Orvaldi even less. And then, assuming that the effective voltage is equal, the rectified voltage becomes 10% too high.