[Power/EMI] I discover why sometime PC become fast and low input lag and otherwise feel high input lag

[Chief Blur Buster]

P.S. Generally; when setting up a new rig -- ∞

Chief, maybe you have some photos of your setup or setups that you consider that are following your thoughts you expressed in above comment? I would be very interested too see those.

Or maybe the other members participating in this thread could also post their setups so others could give their opinions and hopefully get better at some aspect that until now they thought it was fine to have it that way. I think it would be beneficial to everyone, to have a base we could look at, as a picture helps a lot in this context.

PS: My English is not perfect, but hopefully you can understand what i'm trying to say here.

There are half a dozen PCs in this location, but my primary gaming system is the one posted in Desk Setup Thread, but the tweet shows a hint of my cable routing: Spring Cleaning of the main Blur Busters Computer on Twitter.

Slightly aging now, but goodie at the moment as it's all-cores 5GHz stable! Some cases, such as the Inwin 303, has a way to allow you route power cables directly away from motherboard then go behind a metal shield under the motherboard. So unshielded high power wires aren't hovering millimeters away from important chips nor data cables, and there's almost no data cables thanks to the legacy-free design despite now being a 3-year-old computer.

An off-the-shelf watercooler purrs along -- I can't confirm this and never thought about it back at the time -- but I'm deducing it probably produces a better computer-internals RFI environment by keeping fan motors away from the motherboard -- probably helping prevent RFI-induced crashes during overclocks (from cheaply-made crappy fan motors millimeters away from a CPU) since computers are probably more sensitive to RFI-induced slowdowns/crashes during overclocking due to tighter error-correcting margins. Who knows? Anyway, it's 5Ghz allcores-stable. It's also an 80PLUS Platinum power supply too, those are usually RFI-cleaner too. It's scientifically logical to pay attention to RFI hygeine when building a computer, in a "just in case" manner -- doesn't hurt.

While a very inexpensive case, it's very lovely as a 100% legacy free system with zero optical drives and zero drive bays (except hidden under the motherboard). So there's few data cables at all, and there's just only M.2 drives. Being made in 2017, this is the intel i7740 (everyday OC 5GHz), getting old but few chips outperform in older games.

The rules of squares is your friend when routing power wires around the internals of a computer case. Wires 50 millimeters away from surface of motherboard is superior to wires only 10 millimeters away from the surface of a motherboard (EMI injection factor). Now, wires on opposite side of a metal shield separator is even superior (e.g. routing wires on the opposite side of a steel plate under the motherboard, creates a Faraday Cage effect there that is just as good as larger distances between wires) -- so you can favour a computer case that lets you route wires under the motherboard on the opposite side of a steel plate. Hits two birds with one stone; looks prettier for cable routing, and also cleaner EMI hygeine.

I'm no fan of ziptieing between SATA cables and power wires (Especially the wire for the GPU power connector); that is not good RFI hygeine: High-bandwidth data wires AWAY from high-current power wires where possible -- if you studied electronics in school, you'll be very familiar with that concept in electronics circuit design. But duh....what SATA wires? I have none!
(I have drive bays under the steel plate, if I wish, but with three M.2 slots, why do I need drive bays? And I have a parked-away DVD-RW which I only use once every few months)

So you see the Blur Busters main gaming computer is a legacy-free K.I.S.S. of fast + simplicity, and it really simplified cable routing. I didn't spend too many hours on this, except to add the UFO customization and change the RGB to "TestUFO colors". Great "wow-per-hours-of-build" factor for busy Chief Blur Buster -- many forum members spend more time building custom systems though.

Anyway, it's a happy accident of good EMI/RFI hygeine when this computer was built, probably helped it be stable at 5GHz all-cores at default voltage, too. Mind you, so many problems can happen, e.g. loose card in PCI slot, to just bad CPU lottery, or corrupt software install. That said, I never bothered to troubleshoot EMI/RFI, I just simply made sure of good hygeine, just like "washing your hands just in case".

A new tower needs to be done (i9? Threadripper? Etc. Mulling options.), and turn this into a hand-me-down system, but I needed to get a Razer Blade laptop (240Hz RTX 2080) for travel because I need to do mobile TestUFO PowerPoint-style presentations (and it was amazing at CES 2020)....and then COVID hit. Now I almost regret not getting another tower instead in these stay-at-home times. Doh!

Keep safe!

[disq]

^

I was talking more about the setup in-place (the cables route on the desk, wall socket, etc) and not the actual rigs probably didn't expressed myself clear but i sure appreciate such info and pic of your setup (btw, where's the power supply? on top?)

I asked because i think my rig is probably suffering the way i have all the PC cables near power strips, and so on. Also i use adapters from UK plug to EU plug, not sure if this is a bad thing to do.

Actually even my Internet cable is right next to the power cable that leads to the wall socket. This alone is probably really bad. That's why i would like to see how members around here use their cable management (?) to avoid these interferences, in order to check if using such things in a certain way is bad or not.

[Anonymous768119]

Maybe it is not signal in cable itself being affected, but interference travelling through cable to PC? After moving cable completely away from power bar it feels even better. Singleplayer games feel different too, together with mouse feel. Like my input was less erratic, but more stable. Morning, noon, afternoon, night, it feels same.

There are many posts and threads regarding possible power issues interrupting PC hardware but so far it's only guessing. NO ONE has ever proved that your bad hitreg or desync is somehow connected with current in your outlets, no one. I've never seen any surveys, measurements or anything proving it. I've seen many posts that moving to another house or buying completely new PC solved problems and it was never proven by anything. Until someone proves in a practical (and not just theoretical) way with an convincing evidence then I am able to believe that some specific parameter of my current makes it "dirty".

[Chief Blur Buster]

NO ONE has ever proved that your bad hitreg or desync is somehow connected with current in your outlets, no one. I've never seen any surveys, measurements or anything proving it.

That's the Rube Goldberg problem -- a big series of dominoes with separate problems and weak links. Like studying every single atom of a mountain, or try to simulate every single ant or microbe in the world in an Earth Simulator, etc. The problem is just simply too complex.

It's just easier to break down the problem to something simpler, full stop. That's the correct proper scientific approach for researchers.

Dirty Electricity is Slang Version of Proper Scientific Semantics
On the whole interference universe is a complex one (EMI, RFI, on wire, over air, internal, adjacent-circuit, cosmic ray, external factor). Interference is one simple thing, and it can be on-wire or off-wire (injecting into wires or circuits).

Now for simplicity, while "dirty electricity" is an allowed phrase around here, the phrase "dirty electricity" is more of a slang/vulgar term for proper scientific phrases such as "unwanted electromagnetic induction" (I dare you to search) or "distorted alternating current" (I dare you to search) or "harmonics in the electricity" (I dare you to search) or "injection of RF interference" (I dare you to search) or hundreds other phrases. "Dirty electricity" is a four letter word by scientists because it's not specific. It's like just saying an unsubstantiated fact "the sky is blue" in a research paper when one expects proper explanations "the atmosphere of the planet Earth scatters more light of the shorter wavelength territory of human vision sensitivity, and..."

When This is Broken Down To Simpler Problem Modules, It's Already Proven
Many parties already have proved dirty electricity can crash computers.
Increase/lower voltage too much? Crash.
Change power frequency too much? Can crash if out of spec.
Distort the sinewave too much? Can crash if out of spec.

Some supplies are much more forgiving (those universal 90V-260volt 47-to-63Hz AC power supplies are forgiving, some will even absorb randomized electricity waveforms (AC that looks like an audio frequency), while others will crash or overheat when the waveform violates slightly from a sinewave AC. How well the power supply handles dirty electricity. Etc.

Then move on to researching more complex problems, such as "can dirty electricity" start interfering with ANY aspect of computer performance? The problem is there are billions of weak links (a single transitor on a single chip can be THE weak link when it comes to dirty electricity). That's why we have some kickass VRM's powering CPUs to keep them stable -- keep that goddamn electricity ultraclean at the motherboard level, the motherboard manufacturers put VRMs there to clean up the electricity powering the chip.

But obviously, weak links can be on a wire, a common weak link is a modem (DSL modem), where interference is already definitively known to cause latency spikes and Internet slowdowns. It's already (at a leasser scale) proven this can also happen to other shorter data cables, such as a SATA cable or a circuit board trace -- creating an ECC event (AKA lag!). That subsegment is already proven.

Stop looking at the whole Rube Goldberg and focus on the single gears, and study them separately. Much simpler.

If ones goes to a higher institution, university, college, training, to get accreditation, to work for a manufacturer, there's actually electronics/circuit/chip related engineering courses that also includes modules for low-interference circuit design and high-quality voltage regulation. If you are employed by Intel making CPUs, you already know this stuff about weak links about dirty electricity.

Unfortunately, diagnosing which of the billion of weak links are the culprit, is an exercise in futility if the manufacturers didn't do it for you.

Scientific Citations? Sure. Here's a Flood!
Hope you have to never troubleshoot EMI like silicon designers and motherbaord designers -- they already pre-design to make their boards/chips stable during millivolt-league spikes/sags/etc (including millivolt injections via EMI/EMF and dirty electricity).

Most of the stuff you have never seen are only in academic papers. Search Google Scholar or other academic paper sites, like the researcher papers on interference resistant electronics, or EMI resistant chips1, or chip error correction, chip interconnect error correction, low noise circuit design (term #1), low-noise circuit design (term #2).

And that's merely only free Google Scholar. Those with paid subscriptions to research journal sites can easily fishermanboat trawl more boatloads upon request, but this risks making "Bigger Walls Of Text" that just confuses laypeople.

It's typically the user's fault -- when they buy a cheap unbranded generic power supply that goes dirty when it reaches near its max rating, and then complain that their computer crashes a lot more often when they added an RTX 2080 Ti board replacing their old lower-power GPU. These "dirty electricity" things are hard for end users to troubleshoot without tearing their hair out, and are the universe of research.

So.... Be semantically-forgiving when reading the above scientific papers. What we call "dirty electricity" is an alias for a whole universe of problems that already have tens of thousands of research papers. This ain't nutjob tinfoil-hat conspiracy shit, buddy...

There's a lot of battlefronts of interference occuring simultaneously that circuit designers, chip designers, power supply designers, etc, have to fight against.

TL;DR: It's the user's job to
(A) Select the highest quality components
(B) Beyond that, control what you can control
(C) Hope you never have to troubleshoot interference

Now, I agree, this is a fog of red herrings on wild goose chases. When we see a computer crash, and you want to "blame something like EMI", just to feel better because we want a scapegoat, it is almost certainly (even three sigma) probably not what you think. So new readers, skip this thread -- what you think in this thread is probably not your computer problem. Interference-based lag-increases exists of all kinds, internal (circuits) and external (network links), but those not computer literate should troubleshoot something easier or go buy build a new (better) computer and redo your rig setup (wiring) as well as upgrade the weak link (within abilities and budget, i.e. better Internet connection if offered) with the best pre-emptive hygeine you can (like stuff already discussed).

Companies spend millions of dollars troubleshooting EMI already, an individual cannot. So follow step (A) and (B) above instead, please. Don't bother troubleshooting EMI even though it definitely does interfere with CS:GO (yes it happens, and yes, there, I said that). But it's stuff the user usually cannot control. Much of the fixes in this thread is lucky, just like accidentally fixing a dead pixel by massaging an LCD.

There's billions of links in a Rube Goldberg machine (silicon chips are amazing Rube Goldberg Machines nowadays, by the way -- the cascading effects of billions of transitors magically doing computery things for humankind!)

[nick4567]

i would like to add to what chief said in regards to hitreg/desync i dont thing emi/rfi or dirty electricity in general would cause these issues unless it seeps into the ethernet cables (or is affecting the modem/router) and fiber is fully resistant to the effects of emi/rfi and if you already have ftth you could buy a nic that supports fiber to the pc or a good 100 watt isolation transformer to plug just the oni/modem into with shielded ethernet cables to the pc

[dervu]

How would you make fiber straight to PC if you can't replace ONT that has only fiber input, but output only as ethernet cable?
One more thing is that my router is close to old type fuse box like this: https://i.imgur.com/g7aOWx1.png but size of 2 smartphones.
I moved it away from it a bit, just in case.

[alexander1986]

Hello guys

This is last update i think

I bought corsair psu and I compare it to my evga psu

Here is the result

Old Evga psu ( high emi )



https://youtu.be/FuI1uCcFuVY


New Corsair psu ( no emi )

https://youtu.be/Qregvk8WnSI

This test without any filter or power condition or anything

interesting, if I can ask you, are your problems with input lag now 100% fixed with corsair hx1200 psu ? and also you do not need to use any power conditioner or filter anymore, only corsair psu and everything is ok now?


is also interesting and not maybe very big surprise because high end premium psu like corsair hx1200 have much better components for power regulation and filtering but also for emi filtering, for example look at this dedicated emi filter on corsair AX1200i psu :




(pic taken from techpowerup review of corsair ax1200i psu here: https://www.techpowerup.com/review/cors ... 00i/5.html)


quote from review: "The AC receptacle incorporates a server grade EMI line filter which includes all necessary components, one X, two Y capacitors and a CM choke. We find more transient filtering components on the main PCB, namely three CM chokes, two X caps, two pairs of Y caps and an MOV (Metal Oxide Varistor), since a two stage transient filter is required in a desktop/server PSU.

so maybe a premium psu like this eliminates need for power conditioner? or at least helps greatly with this kind of problem...


anyway please give us update on your situation and if corsair psu fixed problem 100%

[dervu]

Have anyone noticed difference in mouse input after switching on/off Smart fan control in UEFI?
If I disable it mouse feels faster/more floaty. If I enable it it feels better, more consistent.
It is apart from my other issues that have been resolved. Something I encountered on overclock.net forums during my research how to resolve my issue.
It is really noticeable. I was testing RAM OC and had to reset UEFI totally, then didn't check this smart fan and was wondering why my aim feels weird. Checked my photos of settings from before and it was smart fan. After that my aim instantly is as before.

[mello]

Has anyone tested this tweak ?

SPREAD SPECTRUM clocking works by continuously modulating the clock signal around a particular frequency. This “spreads out” the power output and “flattens” the spikes of signal waveform, keeping them below the FCC limit. All clock signals have extreme values (spikes) in their waveform that create EMI (Electromagnetic Interference). This EMI interferes with other electronics in the area. To prevent EMI from causing problems to other electronics, the FCC enacted Part 15 of the FCC regulations in 1975. It regulates the power output of such clock generators by limiting the amount of EMI they can generate. As a result, engineers use spread spectrum clocking to ensure that their motherboards comply with the FCC regulation on EMI levels.

CPU Spread Spectrum
VRM Spread Spectrum
HT Spread Spectrum
PCle Spread Spectrurn (SPP)
PCle Spread Spectrurn (MCP)
SATA Spread Spectrum

This option can be found in BIOS settings, and its single purpose is to reduce EMI. The number of spread spectrum options is motherboard and bios dependent. Some will have only 1-2 options, the others will have up to 5-6 or possibly more. Enabling spread spectrum improves system stability but at the same time it limits an overclocking potential of your system. But if you suspect that you have EMI issues, squeezing few additional % of performance (few FPS at best) will be meaningless if you will not correct power & EMI problem.

Every single BIOS & system optimization guide on the internet recommends disabling all "spread spectrum" options, this includes the famous (or rather infamous one IMO) r0ach optimization guide. Which doesn't really makes any sense, unless all you care about is an extreme overclocking and getting every additional 1% improvement in your system performance and FPS. But for someone who struggles with having proper mouse movement & proper gaming performance, improving system stability via reduction of noise & interfearance should be priority, and especially when there is a suspicion that power & EMI are a possible culprit. Furthermore, there is a possibility that excessive internal EMI radiation degrades the quality & performance of video & audio, which also might be of concern for certain individuals. This is important as EMI radiation will be different for every single case & person who is negatively affected by it.

Sources:

- Spread spectrum reduces EMI problem
- What is clock spread spectrum in the BIOS settings?
- Spread Spectrum Control - Anandtech
- EMI and Spread Spectrum Technology
- Reducing EMI and Improving Signal Integrity Using Spread Spectrum Clocking
- EMI Reduction Techniques

[Calypto]

phpBB [video]