AC power line flicker with brand new build

[MachineMedic]

- SOLVED -

Read here:
AC power line flicker with brand new build - Post #64
​

Hello there.

About a week ago, I put together my brand new dream computer from parts that I have been carefully researching and saving up for over the course of the last 6+ months. So far, everything has been running fairly smoothly with the exception of a nuisance that has so far proven to be very upsetting from both visual and troubleshooting standpoints.

Whenever my system is powered on and not under load (whenever I am NOT running Fallout 4, Skyrim, etc.), I am experiencing severe and continuous disturbance of my house wiring system. All of the lights connected to the same circuit as the machine will incessantly flicker at a rate of roughly 10Hz. These symptoms have been immediately noticeable since the very first time I POST-ed the system with only the RAM and CPU installed.

In addition, when I AM running a game or heavy application, I begin to notice interference on my connected analog DVI monitor, and the light flickering disappears almost entirely. The interference on my monitor takes the form of at least one horizontal line that looks somewhat visually similar to a horizontal CRT sweep (what you tend to see when you try to record a CRT screen with a digital camera) and will move up or down depending heavily on frame rate, with 60FPS holding the lines relatively static. When I duplicated my video output on a digital HDMI monitor, no such interference was present.

I have ruled out the possibility of bad wiring or input power by running the system from different house circuits (flickering follows), as well as powering the system from a UPS module that was entirely disconnected from external power. In the latter case, I could plainly hear the UPS output pulsing at the same rate that the lights tend to flicker.

I took the liberty of scoping my house's power signal, and here is what I found:

Spoiler

I was not able to find any voltage spikes or critical deviations with pass/fail mode over the course of a good minute or so.

Spoiler

My incoming sine wave is slightly distorted, but not beyond any standards that would typically be considered 'acceptable'. This is likely because there are many other loads on the transformer supplying our building, and not of major concern. There are a few harmonics in our power system, but nothing that could account for this sort of behavior. The above picture is actually a composite of images taken both when the system was powered on, and when it was turned off. As you can see, there is no obvious difference.

Spoiler

Here's the interesting bit. When I zoomed in very tightly, I *DID* measure a regularly occurring deviation of about 3-4 volts peak near the very tips and troughs of my sine wave when the system was running, between around 172 and 175 volts peak (10X scope probes). This is clearly moderate to major interference being generated by my system.

I have already tested the system with a factory tested warranty replacement power supply provided by EVGA, and the symptoms have not changed. I have powered up the system with every last component removed except for the power supply, motherboard, and CPU, and the problem still persists. I have updated from UEFI revision P1.40 to P2.00, to no avail.

I am awaiting a response from ASRock tech support to see if they have any words of wisdom on the subject, and barring any recommended config changes or pin jumper swaps from them, my next step will be to RMA the entire motherboard. If that doesn't resolve the issue, then I will have to try a different make and model of power supply to see if EVGA is using a bad design in this power supply, or else try to move on to symptomatic cures.

System specs:

• CPU: Intel i7 5820k
  MOBO: ASRock X99 OC Formula
  Cooler: Noctua NH-D15S
  RAM: Corsair 4x8GB 2133MHz DDR4 kit W/XMP
  GPU: MSI 980 Ti Lightning
  Storage: Samsung SSD 950 512GB, Western Digital Blue 1TB/7200
  PSU: EVGA SuperNova 1000 P2
  OS: Windows 10 W/plans to dual boot Linux Mint 17.
  UEFI: X99OCF Chips A & B P1.40 from the factory, chip A later flashed to P2.00

The system is not presently in a case, and is sitting on top of a nonmetallic shelf with some spare aluminum standoffs I had lying around. I don't think this could be a board grounding issue, but perhaps someone here can bring forth some enlightening information on the subject.

Has anyone else had a problem like this with this degree of severity? Does anyone have any knowledge or advice that might help here? I've spent the better part of the last year looking forward to building myself a nice computer, and I'm not sure what else to do from this point onward to fix this issue.

S.O.S.

Bring engineers.

 

[Ferrum Master]

Ground loop? Try without the ground the monitor. A leaking Y cap somewhere... it could be anywhere in your house. I even had spikes with gas dicharge bulbs in power switches...

Aren't anything touching your central heating system, if you have one...

 

[MachineMedic]

I just disconnected my monitor's AC power connector to check the ground loop theory, but the flickering was not affected.

The flickering was not abated even when powering the system from an isolated UPS module, so I'm not particularly inclined to believe that the fault lies on the AC side of things.

:c

 

[Ferrum Master]

The UPS does isolate ground?

I just have a power supple cable with cut out ground wire and use it for testing. Usually the monitor supply is very poorly done.

 

[MachineMedic]

The UPS was connected only to the computer PSU. In retrospect, I only tested it by opening the ungrounded [live/hot] conductor feeding the UPS, but the sound emitted by the unit during operation was much more of a discrete on/off hiss-buzz-hiss-buzz alternation at a rate of around 10Hz than the constant hum caused by a ground loop.

The system is resting entirely on a wooden desk shelf, away from any conductive surfaces. The symptoms were present when no other components were present, so there is no way that I can think of that a ground loop might have been established. The only paths to ground that I can visually see are the monitor power cable, and the computer PSU's power cable. Disconnecting the monitor entirely did not clear the issue, and the issue was present before the monitor was initially connected to the system during first POST.

 

[Azumay]

I take it you have tried other outlets? Pack up rig and take to some one else's place and see what it does.
May be some kinda crazy ground. Check grounds.

 

[MachineMedic]

The build is in no condition to be moved being that I am still waiting for a case to put the system in. MountainMods have phased out their 7PCI Horizon back panels, and are queuing production for 10PCI Horizon panels.

I can try isolating my PSU from ground to see if that helps, but not until tomorrow (bed time over here).

I'd like to set up a current shunt so I can have a look at what the current waveform is looking like, but that will take some time to put together.

 

[R-T-B]

If it were me and I did not have the equipment you have, I would've simply returned the PSU. (While Superflower based EVGAs are good, they are not perfect and I've heard Superflower is outsourcing some of the production with mixed results). That would simply allow me to rule out the PSU.

If downtime is not acceptable, and you have the time, knowledge, and equipment to diagnose further (you seem to), you are way ahead of me. Keep us informed as to what you find out though, I am very curious.

 

[CBRworm]

I have a similar situation, but without the monitor oddities. My PC which is an X87 based Asus board w/ an i7-4790K & AMD HD7970. I also have a Dell M4700 laptop. Both of these cause my office lights to flicker w/ CPU usage. The PC and Laptop both cause a constant flickering when idle. Under high load, the lights dim slightly but the flickering stops.

I have a Seasonic X750 power supply in the machine now, originally I had an Ultra X-1000 which I swapped to an Enermax Galaxy and then the Seasonic. None of them had any effect on the flickering, but at least the seasonic is quiet. Everything is connected to an APC SmartUPS SMT-1500.

I have never had this issue with older builds.

I suspect my issue is due to the fact that my office lights are a combination of fluorescent and LED lights. I'll be interested to see if you find a resolution, or just live with it like I do.

 

[MachineMedic]

@R-T-B:
I've already tried replacing the power supply with a factory tested replacement to no avail. If the power supply is the culprit, then it must be by design flaw rather than electrical or mechanical defect.

@CBRworm:
The fact that you tried many different makes and models of power supplies to remedy your problem to no avail would reinforce my current suspicion that the problem is likely related to the motherboard or processor itself if both of our problems are indeed related in some way. If you just so happen to have a spare motherboard collecting dust somewhere, I would be curious to see what would happen if you connected your power supply to a different board and processor.

 

[bencrutz]

faulty motherboard?
it could be from the VRM, basically it chops the DC current within certain frequency. Any anomaly from the VRM? excessive heat perhaps?

 

[MachineMedic]

The board temperature sensor near the Vcc(in) FETs has never really showed much for high temperatures. The whole board really likes to stay below 50*C by a good margin. The only part of the board that tends to get warmer than that is the PCH/Southbridge/chipset area, which I've noticed sometimes creeps up to 53-54*C when gaming.

CPU likes to hover around 45-55*C under load (at 30-35% fan speed too, hehe), and the GPU's factory fan curve clamps it to about 65*C when the fans even run at all.

Nonetheless, I still suspect something may very well be funky with one of the input stages like you say, as that is the likely spot that a power anomaly like this would probably be generated. The VRMs switch at frequencies in the 100-500kHz range, but if some control logic or the likes is stuck in a feedback loop telling it to switch between two different modes or some other nonsense like that, that could definitely produce this sort of effect. The question then would be why nothing is getting warm if it is drawing enough power to manifest as an AC power disturbance.

I suddenly have a new suspicion that this might have something to do with power factor and the active power factor correction built into the power supply turning on and off repeatedly.

...I really need to set up a current probe to see what is going on on the input side of the power supply. Something tells me I shouldn't be surprised if I find a Cosθ that is jumping back and forth at the magical ~10Hz.

 

[JunkBear]

It could be a sign of defective wall outlet. Bad wiring or malfunction in the breaker itself.

 

[MachineMedic]

@JunkBear:

I have ruled out the possibility of bad wiring or input power by running the system from different house circuits (flickering follows), as well as powering the system from a UPS module that was entirely disconnected from external power. In the latter case, I could plainly hear the UPS output pulsing at the same rate that the lights tend to flicker.

I suppose in addition to probing the power supply input current, I could also look into scoping some of the Vcheck points on the board to see if there are any anomalous waveforms manifesting there or if I am in fact getting pure DC like I should.

With that, I'm going to say good night for now. I want to say thank you to all of you for your quick responses and for and providing me with fresh perspectives on this matter.

 

[jaggerwild]

I don't suppose you have to much stuff on the 110 line, causing undue stress on the circuit? I don't know what the ratting is, but due know only so many wall jacks can be put on a circuit. Also depends on the circuit breaker and its ratting, Your lights don't dim if you run a washing machine's or dish washer? Though the computer is causing it, the computer runs fine hence I think it's something in the house not the computer. I'd like to know how many other things are running on the circuit while your gaming?
Try grounding the motherboard to the PSU, the sign wave looks fine. Your barking up the wrong tree, put yer sun glasses on and enjoy the computer. Better yet shut off the lights!!! Who games with the lights on anyway?

 

[Caring1]

I'd be more concerned your lights are on the same power circuit as the power outlets.
To me, that seems odd as they are separate here where I live as a design code during construction.

 

[R-T-B]

I'd be more concerned your lights are on the same power circuit as the power outlets.
To me, that seems odd as they are separate here where I live as a design code during construction.

That is an excellent point I didn't think of. Yep, they are separate in all homes I've seen.

 

[MachineMedic]

The branch circuit is far from overloaded, and our 100 amp service is rarely pushed even remotely close to it's capacity.

In the US, NEC-2014 allows a switched receptacle to be installed in place of a lighting outlet in the ceiling. That is what the installers did with this room. There is also no legal requirement to wire lighting outlets and receptacles on separate branch circuits. The NFPA publishes their codes as absolute minimum safety standards - following electrical code does not mean that the installation will actually work... it just won't be likely to end up killing anybody. The actual design and execution of the work is entirely up to the electrician on site beyond code minimums.

I'm not saying that is necessarily the way it should be or the best standard that could possibly be invented, but the law is the law, and I avoided law school for a reason.

Again,

I have ruled out the possibility of bad wiring or input power by running the system from different house circuits (flickering follows), as well as powering the system from a UPS module that was entirely disconnected from external power. In the latter case, I could plainly hear the UPS output pulsing at the same rate that the lights tend to flicker.

It *could* conceivably have something to do with the power factor correction circuitry in the power supply if it is poorly designed and constantly stuck in a control-level feedback loop of some sort. It could somehow be a grounding issue that I haven't isolated yet. But given that similar symptoms were still exhibited when the system was powered from a battery operated inverter, I do not believe that the incoming power waveform has anything to do with this problem, other than a poor electrical design choice in my room likely exacerbating the issue.

The computer should not cause a 4 volt fluctuation near only the peaks of my sine wave under any given conditions, regardless of whatever circuit it is connected to.

My main concern is that this is something that could potentially shorten the lifespan of the components inside of my computer considerably. Capacitors that are exposed to heavy ripple voltages will heat up, dry out and expire long before the same caps that were exposed only to pure DC.

It should also be noted that this flickering is far too severe in person to even consider ignoring. Think stress-headache-slash-almost-epilepsy-trigger intensity. The entire room is like a disco party, and I only have one window in my room that doesn't add enough light to substitute electric lighting.

Sorry for any typos, in a rush this morning.

 

[Ferrum Master]

Maybe something silly as someone actually has wired the ground wrong and uses it as neutral?

 

[jaggerwild]

Maybe something silly as someone actually has wired the ground wrong and uses it as neutral?

Bingo!! Or no ground at all, but the wall plugs are UP TO CODE................

 

[Sasqui]

I'd be more concerned your lights are on the same power circuit as the power outlets.
To me, that seems odd as they are separate here where I live as a design code during construction.

Not terribly uncommon to see that.

I'd unplug the PSU from the PC and shunt the green and black wires on the PSU motherboard power connector (turn the disconnected PSU on), see if the problem still exists so you can definitively rule out the PSU.

 

[Bill_Bright]

as well as powering the system from a UPS module that was entirely disconnected from external power.

If you have everything connected to your computer (monitors, speakers, network gear, external drives) and the computer powered through the UPS, then you pull the UPS plug from the wall running everything off the UPS batteries, and you still see lights flickering, then the problem is due to excessive RFI/EMI being emanated from the computer or UPS. My first through is the PSU, but you already replaced that. So I would suspect the motherboard. That said, the clock frequency of the motherboard is much faster than 10Hz. Maybe an arcing drive or fan motor?

Bypass the UPS and plug everything into the wall. If still flicking, it is not the UPS. Disconnect the audio and power cables from the speakers and see what happens.

Make sure your house wiring is properly grounded. Use an AC Outlet Tester to ensure your outlet is properly wired and grounded.

Take your computer to a friends house and see what happens there.

 

[MachineMedic]

I have ruled out the possibility of bad wiring or input power by running the system from different house circuits (flickering follows), as well as powering the system from a UPS module that was entirely disconnected from external power. In the latter case, I could plainly hear the UPS output pulsing at the same rate that the lights tend to flicker.

Again, I am confident beyond a reasonable doubt that the house wiring is perfectly fine. I just measured it with my Fluke and inspected the terminations on the receptacle to be doubly sure. The house was built in 2001, the receptacle polarity checks out just fine, the typical neutral-to-ground potential of 1-2 volts is present to indicate a continuous ground, and the UPS is totally out of the equation [returned to the store before I made my OP]. It is not a fan, motor, arcing commutator, or any other related peripheral in the system besides the potential culprits of the PSU, CPU, or motherboard as evidenced by the symptoms I took the time to carefully lay out in my OP that some people seem to be selectively ignoring:

These symptoms have been immediately noticeable since the very first time I POST-ed the system with only the RAM and CPU installed.

I have powered up the system with every last component removed except for the power supply, motherboard, and CPU, and the problem still persists.

I am a final semester student in the field of Electrical Construction and Maintenance Technology A.A.S. and scored a straight 4.0 GPA last semester. The problem is not being generated from the AC side - the symptoms of what evidently must be an internal problem with the computer are manifesting as a 3-4 volt sine wave disturbance that is causing the flicker.

I am not trying to be rude, but I get impatient when I need to quote my own OP multiple times in a row.

@Sasqui: I've tried connecting a dummy jumper to the power supply with no loads connected to it, and the unit powers up with no flicker. I am not sure if this is a result of the unit simply not being under load, or whether this is actually evidencing an issue with the motherboard as I am inclined suspect.

 

[JunkBear]

The question is....Does it make 1.21 GigaWatts??

 

[MachineMedic]

@JunkBear:

The question is....Does it make 1.21 GigaWatts??

The available fault current on the circuit supplying my room is equal to or less than 10kA. Since it is a single phase branch circuit, we can assume that the available instantaneous fault power will be less than Pf=If*Vp. From 10,000*170 we get a result of 1,700,000 Watts, or 1.7 Megawatts.

In order to have any chance of achieving a result of 1.21GW, you would need equal to or in excess of If=Pf/Vp where Pf=1,210,000,000 watts. That calculates to a fault current of about 7 Megamperes on a 120 volt circuit, about 1.45 Megamperes on a 277/480 volt circuit, or 50.5 Kiloamperes on a 13,800 volt power line.

Does that answer your question?

[/derail]