My experience with all sorts of Problems regularly mentioned here.

[themagic]

1. Do you suffer from input lag, desync, bad hitreg and similar BS?

Yes.

2. Does your graphics card exhibit coil whine when you're in-game? How strong is the coil whine?*

Only if using GPUs with 8Pin Power Cables. Example my Gigabyte AORUS RX580 makes low coil sound. Not rly annoying or and loud while playing Games. But i saw Youtube clips where same exact GPUs makes much louder sound...

My other GPUs which don't need 8PIN power are making no sounds at all and completely silent no matter how much load.

Had an PSU in early days which makes coil whine too if playing Gamaes but after Replace everything silent now.


When build PCs then i tried to eliminate every Sound and make it as silent as Possible cause i really don't like this Sounds like from Fans and no matter what other Sounds inside...

3. What are your minor voltages? Measure +5v and +3.3v with a multimeter directly at the 24-pin ATX connector.

Don't know and never tried but AIDA64 and all tools showing normal voltages numbers and in Bios...there nothing weird strange i detect which could be different from others Users.

[Slender]

yes i have both problem, bad input/picture and desync.
give me info about your TT system. neutral and pen have different "ground" ?
im not using grounding, that cause main frequency instability for me.
that is my result when playing.
you said that your 850w power supplies work well in your mother's house. If this is so, it means that this sound really comes from the socet and penetrates inside the PC. A higher power supply has correspondingly larger coils, which reduces the actual coil wine sound.
I can't write to you in private messages because it disabled in your account.

[ChristophSmaul1337]

give me info about your TT system. neutral and pen have different "ground" ?

It's a standard TT system. There's 4 conductors coming in from the power company, three phases and one neutral, or N. There's a local ground rod in the basement and it's been tested multiple times by now. There's a 400 mA RCD and it works fine. As of the last measurement from about half a year ago, the ground resistance is more than low enough to reliably operate the RCD and, therefore, meet code standards.

im not using grounding, that cause main frequency instability for me.

I've built myself a device which can cut out the grounding with the flip of a switch on demand. I'm not going to go into detail about it, because nobody should ever do that as it's a safety hazard at best, and outright deadly at worst, and I've experimented with that thing before. From what I could tell, it didn't make a difference before. It certainly doesn't make a difference now, I've just tested it again with my new PSU and it's all still good, as expected. Grounding or no grounding should not be a problem. However, I'm still going to recommend checking if your country does require grounding and if it does, get an electrician to install one. No available grounding combined with an appliance class 1 fault can be deadly.

you said that your 850w power supplies work well in your mother's house. If this is so, it means that this sound really comes from the socet and penetrates inside the PC.

It was working well, yes. However, at the time I tested it, I didn't really look out for coil whine, so I don't know if it's present there, too.

A higher power supply has correspondingly larger coils, which reduces the actual coil wine sound.

Could be, but I doubt it's the reason. For example, I have a 650W PSU that has noticeably less coil whine than a 850W one. It's still prominent with both though, and sure enough both show problems.

I can't write to you in private messages because it disabled in your account.

Correct, I want all discussion to take place in the public part of the forum, to provide all the information to everybody who's reading along. No need to keep anything in private.

Your voltages look fine at a first glance. However, I would still get a multimeter and directly measure at the connectors. Make sure you put a full stress test on while measuring, to load the rails as much as possible. There's a lot of programs out there that can do this, a quick google search will help. Just to make sure the motherboard/HWINFO reads the correct values first. I have at least one board which doesn't show correct readings.

Depending on the power supply, the spread between lows and highs could be normal. Your +5v rail at the highest point has been up to +5.04v and at the lowest down to +4.980v. Both of these are well within spec, yes, however for my current system, the voltage doesn't vary at all and stays right at +5.000v. This is corroborated by DMM measurements. I suggest you look up your PSU model, and search for a test report. If testing shows this variance to be normal, then it might not be indicative of a problem. However, if testing shows there to be no or close-to-no variance in voltages across the board, it could indicate a defective capacitor or similar. Compare your load voltages with what is shown on a test report. If your +5v voltage goes down to +4.980v under load, but test reports show the voltage on the +5v to stay above +5v at all times, it could hint at a problem with the PSU. Sure, +4.980v is still well within spec, but if it doesn't line up with all the other models, it still isn't right and something might be up with it.

One example is my beQuiet 850W unit. As per multiple test reports, the +12v rail isn't supposed to drop below +12v at all, even under full load, and even when approaching the OCP trip point. However, mine regularly fell below +12v all the way down to +11.7v with a full stress test. Sure, still within spec and the computer stayed on, but there's obviously something wrong with it and, sure enough, it caused the infamous problems. Hooking that unit up to a very moderate power draw computer, an old i3 with a low-power GPU, barely causes the +12v to budge even under full load. Same is true for the +5v, too. It's not supposed to fall below +5v at all, but it did for me all the time, as low as +4.8v. Again, within spec, but not at all what test reports show.

With +12v, it's not really that much of a deal, because no component in your computer will be powered directly off of +12v anyways. CPU and GPU use +12v, and they both have their own buck converters to step that voltage down. These buck converters couldn't care less if the incoming voltage is a few mV higher or lower than spec. However, depending on the board, the other voltages might be used directly from the PSU, without going through a buck converter first. For example, my motherboard just "passes through" the +5v from the PSU to power the USB ports. For these "passthrough" boards, an unstable or low quality +5v voltage from the PSU might(?) cause issues with USB devices, but I really don't know. This is just speculation at this point.

[Slender]

So disconnecting the PC from grounding does not affect your PC in terms of smoothness/responsiveness etc.? By the way, do you only have a problem with desync when you install a 850w PSU or other problems?
My psu is SF-850F14MP SE.
And I think there shouldn't be any problems with it, at least the reviews say so.
How can I measure voltage with a multimeter while running a full load test on the PC?
By the way, setting 240v on my ups solves some problems with 5V and 3.3V, now they do not fall below these values. The overall feeling of hit registration and smoothness has become noticeably better, but not fix at all.
In max load my build has 500w power cons.

[ChristophSmaul1337]

So disconnecting the PC from grounding does not affect your PC in terms of smoothness/responsiveness etc.?

Exactly. It makes no difference at all. This makes sense, because there are countries in which the national electric code doesn't require grounding and there's certainly players located in these countries, playing games without a problem.

By the way, do you only have a problem with desync when you install a 850w PSU or other problems?

Yes, the desync is by far the most noticeable problem. There's also just terrible hit detection, like shots passing right through people even if they absolutely should hit. I know that this must come from the PSU in my case, because a.) it happens in every game I play, all the time, and b.) it goes away as soon as I switch out the 850W (or a 650W, or a 700W, or a 550W one) out with the 1300W Super Flower unit.

How can I measure voltage with a multimeter while running a full load test on the PC?

That's pretty straightforward. Your motherboard has a 24-pin connector on it. It's the humongous one in the middle and to the right of the board. As your PSU has all black cables, you can't tell which wire is what voltage from just looking at it. You can either know the pinout from doing this countless times, or you can look up pinout diagrams online, one example is this one:
Important: The picture shows the 24-pin connector from "below", as the receptacle on the motherboard would see the connector. When measuring the voltage "top-down", like you're going to do, flip the image in your head vertically. Otherwise, your measurements won't line up. For example, the two orange +3.3v pins in the top left of the image would sit on the bottom left when looking at your connector from above.

Next, you want to take your multimeter and connect its leads for DC voltage measurements. Then, take the black lead (ground/COM) and probe any ground pin in the 24-pin connector (marked black in the pinout). You can just stick the lead in from the top. Don't force it in too hard, it should read a voltage pretty early and there's no need to force the lead in all the way. Next, find a +12v pin on the 24-pin connector. Stick the lead into that pin, just like you did with the ground lead. Repeat for +5v (red) and +3.3v (orange).

Full computer load can be generated with synthetic tools. One such tool is OCCT, but there are many more.

Put on the load, and only then take your measurements. Measuring the voltages in idle won't reveal problems.

Write down the measurements you've taken with your multimeter. You said that your PC has a power consumption of 500W under full load. Did you estimate this from rough calculations/TDP's, or did you measure this with a power consumption measurement device? Measuring from the wall with a device would be preferred. These power meters usually look like this:

Let's assume 500W is accurate for now. I'm looking at a test report for your PSU right now and it states the following values for 60% load:

• +12v: +12.086v

• +5v: +5.048v

• +3.3v: +3.367v

Your measurements should line up with this pretty much exactly. If they do, fair enough. Your problem is likely of other origin than PSU/electrical BS. However, if they're different by more than a few millivolts, it might be worth looking into it further. If you want, you can post your measurement results in here.

By the way, setting 240v on my ups solves some problems with 5V and 3.3V

This shouldn't happen. The power supply probably has a large input voltage range, and from looking at the label it supports all input voltages from 100-240v at 50Hz or 60Hz. Changing the input voltage from 230v to 240v should not make a difference at all, especially not for voltage regulation on the minor rails. I would say it's probably coincidence. Germany has a nominal voltage of 230v, however the voltage in my house is more around 240v all the time. There's some times when the voltage drops to about 225v, and it can get as high as 245v. This is normal and to be expected. However, any change in input voltage doesn't change the voltage regulation on any of the rails. And, because of how these PSUs work, this is exactly what should (or rather, shouldn't) happen. What you're seeing is definitely strange, if you can actually correlate input voltage with the voltage regulation on the minor rails. I'd look more into that. Make sure it's not coincidence, though.

now they do not fall below these values.

This implies that they do fall under with your UPS set to 230(?)v (whatever is default for your contry). If so, then there's definitely something wrong with the PSU. It's designed to take a wide range of input voltage and changing from/to 220, 230 or 240v should make zero difference in output regulation. Looking at the test report, the voltages on either rail should never drop below the spec in any loading scenario. Even at 110% load, which would be 935W, the +12v is at +12.088v, the +5v is at +5.045 and the +3.3v is at +3.367v. The load regulation on your PSU is excellent, as good as it gets honestly.

Your next step should definitely be to check your load voltages with a multimeter.

[Slender]

measured voltages with a multimeter at full OCCT load:
12v: 12.18v
5v: 5.00 - 5.01
3v: 3.31 - 3.30
im also testing 200v/230v/240v all are the same.
I was not quite right about 500w at full load. OCCT shows 300w gpu and 160w cpu. In fact, it turned out to be 700w (-70w ups inverter).
(I measured it using a smart socket + information from the UPS)
And the PC consumes ~600-630w at full OCCT load. For games, it does not reach above 330-350w.
About voltage, it's really a very strange thing, because as you said, the power supply can work even from 100v, but the difference between 200v and 240v is quite obvious in my case. It is also very strange that the motherboard shows different values ​​3.3v 5v when 200v and 240v are set, but I repeat, there is no difference on the multimeter.

by the way, any decrease in voltage of any three components: gpu / ram / cpu improves the situation.

Do you also have a superflower? What model do you have?

[ChristophSmaul1337]

measured voltages with a multimeter at full OCCT load:
12v: 12.18v
5v: 5.00 - 5.01
3v: 3.31 - 3.30

Looks fine at first glance. I've spent the better part of an hour now in order to find a test report for your PSU model, but there doesn't seem to be one. There's even conflicting information on what exact PSU this is. A search for "sf-850f14mp" gives multiple results. I could only dig up a test report for a 1200W model. At 40% load (about 500W), the +12v rail is roughly at 101% of its nominal value. This would be about +12.12v. The +5v rail is at about 100.5% of its nominal value, or roughly +5.025v. The +3.3v rail is sitting roughly at 100.3%, or about +3.31v. In this regard, the measurements would line up and make sense. Keep in mind though that this isn't taken from a test report for your PSU, only from a related one within the product family.

I would consider this to be fine. As much as I would like to tell you an actual solution for your problems right here right now, I think your PSU isn't the cause and your problems are rooted somewhere else.

im also testing 200v/230v/240v all are the same.

Correct behavior. This is exactly how it should be.

I was not quite right about 500w at full load. OCCT shows 300w gpu and 160w cpu. In fact, it turned out to be 700w (-70w ups inverter).

Is anything else connected to that UPS/socket? 700W sounds a little bit too much in my opinion. Unless of course you're using a 4090 paired with a 14900k or something like that. 300w for the GPU and 160w for the CPU sound a lot more reasonable to me. I suspect you have monitors, speakers and maybe more stuff connected to that UPS, so it reads more than the PC itself consumes.

About voltage, it's really a very strange thing, because as you said, the power supply can work even from 100v, but the difference between 200v and 240v is quite obvious in my case.

What exactly is different between the voltage settings? The game feel?

It is also very strange that the motherboard shows different values ​​3.3v 5v when 200v and 240v are set, but I repeat, there is no difference on the multimeter.

That is indeed weird. The indifference on the multimeter makes perfect sense, as that is exactly how this is supposed to work. The PSU doesn't care about the input voltage. However, if you can actually observe a connection between input voltage and software readings, it could at least be an indicator for a general power-related problem, even though the voltage on a multimeter is the exact same. Maybe one input voltage screws with the PSU in some weird way, causing a lot of ripple and/or noise. That would still read the same voltage on a multimeter, but it could frick around with a sensor. This shouldn't happen, as input voltage shouldn't have an influence in voltage regulation / noise suppression and output voltage quality, but how knows...? I would suggest testing for this more thoroughly and see if there is actually a correlation between input voltage and software readings, of if it was a coincidence.

by the way, any decrease in voltage of any three components: gpu / ram / cpu improves the situation.

This could make sense if the PSU was indeed the culprit, like it was in my case. Lowering CPU/GPU voltages also reduces the power draw of the component. And less power draw means less load on the PSU, which is exactly what I've observed in my situation. Now I'm really interested if you could also replicate my observations with lower-end, low power draw components. Any chance you have low power components around? Like maybe an old i3/i5 and a GPU without a supplemental power connector? Use your current 850W PSU with that low power PC and try to game on it. Despite the low FPS you should be able to tell a difference in desync and hit detection.

Do you also have a superflower? What model do you have?

Yes. I have a Super Flower Leadex VII XG 1300W. This one: https://www.super-flower.com.tw/en/prod ... -atx-30-bk

[Dorion1004]

Hello,

I wanted to ask about the power supply for a setup with an RTX 4070 Ti and an i9 14900K. Would a 1500W power supply be sufficient for this kind of build, or could it help resolve a problem similar to the one you've encountered? I'm looking for a solution to an issue that's quite similar to yours, and I want to make sure the power supply can handle the energy demands of these components.

Thanks in advance for your feedback!

[RealNC]

Hello,

I wanted to ask about the power supply for a setup with an RTX 4070 Ti and an i9 14900K. Would a 1500W power supply be sufficient for this kind of build, or could it help resolve a problem similar to the one you've encountered? I'm looking for a solution to an issue that's quite similar to yours, and I want to make sure the power supply can handle the energy demands of these components.

Thanks in advance for your feedback!

750W is enough for that already. I'm using a Corsair RM750x with a 4070 Ti Super and a 5700X3D. If you want the option to be able to use high wattage GPUs though, then go for a 1000W PSU. Anything more than 1200W doesn't make any sense. And always keep in mind that the quality of the PSU is just as important as wattage.

[ChristophSmaul1337]

Would a 1500W power supply be sufficient for this kind of build, or could it help resolve a problem similar to the one you've encountered?

750W is enough for that already.

I second this. A 750W should easily be enough for that setup. Make extra sure to take into consideration that...

the quality of the PSU is just as important as wattage.

You don't need to over-spec on your PSU like I did, because there's a very, very high probability that you don't suffer from whatever the hell is happening to me. Don't look into what I've found as my solution until you've ruled out all of the more common stuff. I'd argue my specific situation is so niche, it's probably a one-off. What RealNC is saying is totally correct.

Again, I want to state how important it is to go for a high quality PSU. No matter if you have problems or not, it's always a great idea to get a super high quality unit. It's a critical component and it has the ability to un-alive all your other components when it fails. The higher quality units have mechanisms in place that stop that from happening.

To RealNC: I've read the forum over the last couple of days and I've noticed the threads that have come up with your name in the title, from that one guy who's telling people their chipset is damaged. I've seen that guy calling pepole, including myself, "members of sects" and making fun of people who're looking for solutions for their real problems. I'm glad that action has been taken against this person and I want to thank you for it. You did a great job and I'm glad that you're an administrator on this site.