[SOLVED] Monitors turning off/on, then pc restarts.

[Leonhart]

Hello.

I've done some digging around but I wanted to ask this directly.

My pc was getting old (10 years I think) and so at the end of 2019 I bought:
B450 MSI motherboard
Ryzen 3700x 3.6GHz
16GB Corsair ram
Noctua NH D15 CPU cooler

No problems for a year. The GPU fan was whirring up and down but that was all so I got myself a MSI 3600 Ti GPU.

This leaves my PSU (1000W Zalman), 2x 600Gb HDD, a creative soundcard, and a chassis which developed a faulty power on button (seemed to be the power button connector so switched it to the reset button). Despite the list, I believe only the PSU is relevant to my situation though the chassis wiring just might be an issue so thought I'd mention it. I also have 2 different monitors but I don't see how that could be a reason (like being at 120 Mhz and the other 60 Mhz, different resolution cap etc). If so I can provide that information.

So the actual problem. Ever since I got my new GPU my PC has rarely and randomly started giving me black screens for a brief moment and comes back on again (sometimes videos/sound gets interrupted or settings changed back like output, not 100% sure now). This can then happen a 2nd or 3rd time and then the PC restarts. I get no error message or anything to suggest something was wrong. My home's power supply is reliable so I know the issue is connected to my PC.

I've dug around the internet and what I learned was inconclusive.

• Driver issue. I think I made every effort to uninstall the old GPU driver and install the new one. I've used the Nvidia experience app to install new drivers.
• PSU issue. It's old and changing from 750 to 3600 Ti might be the reason though a power consumption calculator put me at no more than 500 W so I can't see capacity being an issue. I don't know what else to look for if the PSU is failing me though.
• HDD issue. My wife considered this and they are old too but I've had no other issues. I had to basically do a clean install last year due to a new motherboard.

I guess my biggest concern is what this is doing to my computer, is it causing damage? How likely? And how badly? From what I understand it's not the loss of power that's the problem it's the surge afterwards but I'm not sure if I'm getting a surge (and I'm mixing this up with an actual power outage like a storm knocking out the power). Everything new is under warranty but I'm not sure how worried I should be that I'm damaging my PC. I don't believe I have a surge protector so perhaps I should consider that.

Today the issue happened 2 or 3 times today yet I've gone maybe days without any issue which is what has triggered me to try and find out what's going on. I've noticed my motherboard RGB (which reflects temperature) has been going up and down. CPU temperatures are 40-60 degrees Celsius which seems on the high side for what I'd consider idle load (websites and an emulator, that's about it). This makes me wonder if now my PC has decided that the CPU is getting too hot after a year but it's still within limits, I'm not getting any warnings that I can tell. Once I saw the light go red which says it was hitting 70 degrees celsius but that only happened once. I can't recall the temperatures before my new GPU but I think my lights have never reflected a 50+ degree celsius PC.

Should I reapply the thermal paste? Even if it's not the problem does it sound like it needs reapplying?

I don't know what other information I can provide. I've gone ahead and bought 2 WD 1TB HDD and a Corsair 850W PSU but those will take a few days to arrive.

Thank you in advance for any help. I've tried my best to try and come up with a solution myself but I'd feel more reassured if I presented my problem directly.

P.S. I used CPU-Z to test my CPU.

Bench test pushed my temperature to mid 67-68's (no real difference in fan noise).
Stress test pushed my temperature to 67-69. My fan is making a noticeable but not deafening noise (before it was dead quiet).

 

[Vic 40]

If the psu is as old as the old pc was change it. Might well be the problem, it's not about wattage in that case, just it being old and likely lower quality.

 

[Leonhart]

If the psu is as old as the old pc was change it. Might well be the problem, it's not about wattage in that case, just it being old and likely lower quality.

My new Corsair RM850 x V2 arrived today. I've got 2 new WD HDD as well and I've got a new chassis on the way so will likely wait until that arrives.

Yesterday I had the PC on for 12+hrs and not one problem. The day before I think it happened twice in quick succession.

Might as well add this as an update.

I've noticed that early in the boot up process the CPU light comes on briefly (for a second maybe), then the VGA light for a few seconds, then the Boot light briefly. All the lights turn themselves off, I do nothing and the computer boots up as normal. I can't say I've noticed this before after a year of using it and while the manual says the lights are there for when something is wrong it also says they stay on until the problem is resolved. The lights never appear to come on again. My motherboard is a year old MSI B450 Tomahawk Max.

BIOS screen says my temperature is around mid 30's CPU and mid 20's motherboard.

I guess I just have to hope I'm not permanently damaging anything.

 

[Vic 40]

The check lights coming on is normal and they turn off which is good so don't worry about that.

 

[Leonhart]

Well, it's happened again where my monitors just switch off then come back on for no reason (sound also gets cut off) then the monitors turn back on after 5 consecutive days of no issues with maybe 12 hours or so use each day. This time the PC didn't restart but things like the live stream I was watching stop working and I need to refresh the stream.

Here is my setup again:
B450 Tomahawk Max
Ryzen 3700x 3.60 GHz
Noctura NH-D15 cooler
Nvidia MSI GeForce 3600 Ti Ventus x2
Corsair 2x8GB 3200 MHz ram
Corsair RM850x
WD 1TB HDD x2 (one is boot drive), and WD 640GB x2 HDD
AOC 24 and Dell 21 inch monitors.

As far as I can tell my GPU is not going higher than 36 degrees Celsius, CPU no more than 60 I think, the motherboard in the 30s. The EZ debug lights still come on during BIOS but only briefly as I detailed earlier.

I've not heard much about this being a common issue with 3600 users so I'm doubting this is a known and common driver or hardware issue. I'm lost really as to what the problem is and where to go next as everything is new or near new.

I've swapped my primary monitor to my old 21 inch monitor and will see what happens. If the problem is with the primary monitor then I could image both turning off and on but if the problem is with the secondary monitor then I could see just that going off and on and that would help pinpoint the problem (the new monitor or the cable are both a month old and the problems started then too). Only problem is it could take days to find out if I'm right or not.

 

[Leonhart]

Now my computer just decided to restart without warning. I thought there was actually an issue as the BIO lights came on but then I saw the PC was going through booting up again.

I don't know if this is a delayed reaction to earlier (since it didn't restart then) but it's really bothering me now as I can't see how it can be the monitors or the cables and I've run out of ideas.

Any suggestions on where I can go from here?

 

[Vic 40]

Can tou test with other ram?

Think it's going to be about swapping parts, might as well be motherboard having issues.

 

[Leonhart]

I'm not sure I can really swap much out anything other than the PSU and GPU. Most of the old parts that connected to the motherboard are about 10 years old and so a lot of things are no longer compatible. That's why when it came upgrading I ended up having to upgrade everything. I don't think putting in DDR3? ram would fit much less work in a motherboard that has DDR4? ram slots.

I'd find it odd that the CPU, cooler, ram or motherboard would suddenly develop a fault only after I installed a new GPU and monitor.

The ONLY thing I could consider is putting my old 750 Ti back in but if the problem won't happen for an indefinite time I won't really be sure I've found the faulty part. It would help if I knew what could cause it. I don't see how a GPU (hardware or software) could cause my PC to restart, turn the monitors off and on sure but not the PC itself. I could also try putting the 3600 TI in what is effectively my old PC and see if the problem persists.

I've emailed MSI regarding the GPU as the motherboard is also MSI and they might be able to help. I could try asking Nvidia but I'm sure I'd have heard about an issue if 3600s or the 3000 series as a whole had this issue, plus MSI should know too.

 

[Vic 40]

Don't think i suggested it already, but did you update the bios?

 

[Leonhart]

Don't think i suggested it already, but did you update the bios?

Yes I have updated to the latest non beta.

I've put the 750 Ti in my new pc and I'm hoping Wednesday I can put the 3600 Ti in my old pc and see which PC gets issues. There's a chance the GPU wasn't connected right but I highly doubt it. Something I'll check.

 

[thepersonwithaface45]

I'm having the same issue!
Periodically my monitors will shut off for about ~10 seconds and then my system will come back online. It does not do this in succession but a good amount of time inbetween black screens/audio cut out. Audio cuts out about 5 seconds after the screens do.
All this is happening after my 3080 upgrade. I thought I downloaded all the ASUS drivers I needed and updated bios, but for some reason I'm thinking it might be my psu..

 

[AntonyLovric]

Hey there, anything in windows 10 event viewer? (Under the system events?) Did you run the windows memory diagnostics? Bad ram address would explain the lack of frequency of the problem. Windows 10 has a good memory diagnostic tool.

Read Memory Diagnostics Tool Results in Event Viewer in Windows 10 | Tutorials (tenforums.com)

 

[Vic 40]

Test with he "studio drivers" as well.

Think with Leonhart he might just have a bad motherboard. Since he has restarts.

And do in this instance try the beta bios as well.

 

[Leonhart]

Test with he "studio drivers" as well.

Think with Leonhart he might just have a bad motherboard. Since he has restarts.

And do in this instance try the beta bios as well.

If it's the motherboard then it's disappointing it's only a year old (week or so over). Luckily standard waranty is 3 years though I've never had this kind of issue before with this manufacturer so I hope they are customer friendly and not try to blame me and invalidate the warranty.

I've yet to hear from MSI and I'm still running my old GPU so I'll see hopefully if it's the motherboard or not.

I learned that updating BIOS is a last resort and if something goes wrong, well that's that so that's why I chose non beta. Also, my PC was fine for a whole year before the problem came up so would be surpised the motherboard chose now to break. I'm not new to modding PCs so I took the upmost care though how easy it is to do damage that would cause this problem, I don't know.

Thank you for the input. I'll see about the RAM check if I can rule out the GPU as I use my PC for work and I've lost time as it is trying to get to this point. I'll keep you updated.

 

[Vic 40]

The gtx 750 ti runs fine?

 

[Leonhart]

The gtx 750 ti runs fine?

So far but it took 5 days for the problem to reoccur and I wasn't keeping track before that so I don't know when I'll be able to give an update. It's not a daily occurance but it feels like it happens maybe once 1 week or 2. Once I put the 3600 Ti in my old pc I'll be able to see if that's the cause. So far it's only been one day so I guess I'm looking at next weekend to see if it happens again but even if it doesn't I won't be sure it's okay.

 

[Vic 40]

If that gtx 750 ti keeps performing fine shouldn't it be the motherboard that is faulty.

 

[Leonhart]

If that gtx 750 ti keeps performing fine shouldn't it be the motherboard that is faulty.

If you mean 750 Ti works = it CAN'T be the motherboard, then yes.

Tried to install GPU in old pc and at the same time test my wife's old hard drives. Didn't work, now we put my old hard drive in, didn't work. The furthest we get is verifying DMI integrity so something has gone wrong but we can't figure out what. It can't even find the flash drive to at least try and install windows 10.

So for now I can only test my old gpu in my new pc.

 

[Leonhart]

Well, it's been a week with my old GPU and no issues. My old PC has had the new card for 6 days and while I've not actively used it, I've had a game playing itself and my wife has done file management on her old hard drives, no issues there either so far.

I can't really say then what caused the problem. Unless I connected the new card to my new PC improperly which somehow the reason for the issue, I'm no closer to knowing what caused the issue.

In short, my new PC as it is now had problems with the new card and doesn't with the old one, but my old PC with the new card also doesn't have a problem.

 

[thepersonwithaface45]

Replaced the RAM and the issue has stopped so far

 

[Leonhart]

Well, things get interesting.

I've used the computer with all new parts and old GPU for roughly 12 hours a day for 10 days and no issue. I used the PC actively as I did when I had the 3600 Ti installed.

I've used the computer with all old parts and the new GPU for almost the same time and no issue. My wife used the PC doing file management on a fresh Windows install and I've had the computer play itself in a game but didn't actively use the PC.

I had MSI Afterburner on both PCs making sure the fans were used in case this was a cause.

So this weekend I plugged the 3600 Ti into my old PC. The PC didn't boot up, it got stuck on the VGA boot LED.
Plugged 750 back in and it worked. Uninstalled everything Nvidia using windows app settings and then device manager.
Plugged 3600 Ti back in, stuck on VGA boot LED
Plugged 3600 Ti into 2nd PCI-e port, also stuck on VGA LED.
Plugged 3600 Ti into old PC, worked.
Plugged 750 into new PC, worked.

So basically the 750 can work in my new PC.
The 3600 Ti will only work in my old PC.

The 750 doesn't need PCI-E power cables whereas the 3600 Ti does BUT I had problems with my old PSU so I can't see how that's the problem.

I hope people aren't too confused, congratulations if you understand. MSI had got in touch saying they thought my GPU experiment was a good idea so I've updated them with that.

My wife has done some research and suggests I try DDU to remove anything related to my old card. Another solution is to do a clean install of my boot drive if I have to.

 

[Leonhart]

Please disregard my last message. I noticed that I hadn't plugged the PCI-e cables into the PSU. I must have unplugged them to remove some hard drives, and not plugged them back in. The 3600 Ti is working in my new PC and I used DDU to remove all traces of Nvidia drivers beforehand in safe mode and no internet so it's as clean as it can be.

I will let you know if I have any other issues. I'm hoping/praying that the reason the 3600 worked in the old PC and not in the new was due to a driver issue which wasn't resolved despite following the uninstall procedures.

Apparently, there is a new update driver since I downloaded the driver to test in my old PC so I won't update it yet incase that causes a conflict again (even though it's the same GPU just an updated version). The driver I installed in my old PC didn't appear to have problems so I'll test my new PC with the old driver and if all is well I'll update the driver.

 

[thepersonwithaface45]

https://imgur.com/X4cX1N8

View: https://imgur.com/X4cX1N8

Now I'm getting black screens again so I guess RAMM wasn't it, thinking now it's a PSU issue or a windows issue..

 

[Leonhart]

https://imgur.com/X4cX1N8

View: https://imgur.com/X4cX1N8

Now I'm getting black screens again so I guess RAMM wasn't it, thinking now it's a PSU issue or a windows issue..

I'm trying this piece of software (https://www.wagnardsoft.com/). It seems to be well recommended. I'm wondering if switching from 750 to 3060 and uninstalling via windows/Nvidia software left remnant code. The fact that one card didn't need additional PIC-e cable power and the other one needed 2, maybe that's why I was getting issues and maybe you're the same. I've literally tried EVERYTHING everything else. My old PC had a clean install of Windows 10 so would have had no driver conflict. If this doesn't fix it for me I don't know what will.

 

[thepersonwithaface45]

check the display warning, I'm getting a nvlddmkm error. I found more threads talking about this error and I've underclocked my card to see if it helps. If not, it can be display cables.
Or, if it happens again, I'm gonna underclock again, then if it happens again, new PSU.