[SOLVED] Is it possible a computer component can corrupt 2 gpu's?

Thomas4140

Prominent
Aug 6, 2019
40
0
530
I have bought a new graphic card: ZOTAC - GeForce RTX 2070 SUPER .

The problem is that the graphic card is defective.

It worked nicely for the first days, but no longer afterwards. At day 3: small graphic artifacts appeared on the screen.

The next days the gpu always caused a flickering or complete blackscreen when booting up as soon as i start my computer in normal mode then restarts my computer. https://streamable.com/ohktc

This is not a driver issue, since I download and tried older / latest drivers and they still ended up causing flickering and eventually black screen and restart. I uninstalled all drivers with DDU multiple times and checked manually too. Also reinstalled windows 10 Pro and stresstested my pc, tested on multiple monitors.

As soon as i go into higher resolution than the standard 1024x786 (or something like that) The monitor starts flickering and blackscreening. Or if i update drivers it is causing same thing.

To test if it was my gpu or other component that was causing this, I switched between my Radeon R9 390 8G multiple times with the rtx 2070 super. They both ended up giving blackscreens, even on the another computer with completely new hardware and monitor.

I brought my computer and 2070 to my local repair/pc shop to test what's wrong.
He build a fresh new hardware setup to test if any of these components (cpu, psu, motherboard) in my computer where causing the problem. But the gpu gave the same blackscreens despite being in his new hardware setup computer.

Today i went to get my pc and 2070 back, and asked him if he could try out the radeon r9 390 on his new hardware rig, and it also has the blackscreen like the same as my desktop home.

He also tested with 4 other graphic cards on my OWN computer and they all worked fine.

Is this just coincidence or bad luck? It's really weird that both die at the same time. I tried almost everything and the repairer did too on completely different hardware.

MOBO: MSI Z97 GAMING 3

CPU: Intel CORE I7-4790K 4.00GHZ

PSU: ANTEC EDG750W

RAM: 16GB DDR3 - 1600 mhz

OS: Windows 10 64 bits
 
Solution
The one cable still should be enough,but if you might get problems doing something stressfull on the gpu test with two cables before sending things back.
Memory is easiest to test and it's worth doing it even without indicative problems. Actually you should check whole system before condemning one part as faulty. PCIe lines are controlled from CPU so that could be a cause. Try GPU in another PCIe slot even if it's not running at x16.
 

Thomas4140

Prominent
Aug 6, 2019
40
0
530
Memory is easiest to test and it's worth doing it even without indicative problems. Actually you should check whole system before condemning one part as faulty. PCIe lines are controlled from CPU so that could be a cause. Try GPU in another PCIe slot even if it's not running at x16.
I can confirm that running my gpu in pci-e in the second slot did work but had the same problems as in the first slot due to another faulty component i think. I'll run a memtest on the ram later and see if there's any errors. He did test my motherboard and psu but only for one hour, so not enough time to fry a component. It took 4 days till my gpu's started messing up and probably got fried.

I'll send back my new gpu on monday due to garanty, and probably get a new one but if I do i want to be certain I don't end up with frying that one too.
 

Thomas4140

Prominent
Aug 6, 2019
40
0
530
Do you know how i can test my psu malfunctioning with voltages. Looks like my psu keeps running the fan tho but makes louder noises and sometimes hear the fans slow down or go fast for a brief second
 

Vic 40

Titan
Ambassador
First thing i did think of is psu. Can really test ripple ,which seems most likely as cause (too high) , only with professional equipment. If the psu has still warranty send it back. Otherwise ask the shop if you can lend a psu for testing. If the psu is old maybe just replace it.

You can monitor voltages with software,but in th end it's software,

download hwinfo64,
install and open it,
check "sensors-only",
click "run".

Or use a multimeter and use an empty molex connector for testing while running the pc. For +12V would that be the yellow plus a black cable.
 

Thomas4140

Prominent
Aug 6, 2019
40
0
530
First thing i did think of is psu. Can really test ripple ,which seems most likely as cause (too high) , only with professional equipment. If the psu has still warranty send it back. Otherwise ask the shop if you can lend a psu for testing. If the psu is old maybe just replace it.

You can monitor voltages with software,but in th end it's software,



Or use a multimeter and use an empty molex connector for testing while running the pc. For +12V would that be the yellow plus a black cable.
It's from january 2016. How do i test it? What stats should i keep an eye on in my thing. How i see if something went bad with voltage's?

The repair guy tested another psu, and all my own other components for about 1 hour. I think its not the problem if my psu is working or not. I think something inside it might be gone corrupt that gives high voltages from time to time. But u only probably notice it after testing it a while (same with the motherboard). Both gpu's died after 2-4 days so u can't really test it good for shorter times.

I was thinking about replacing that and maybe also the motherboard, idk how to check if motherboard is bad too.. Other gpu's worked on it tho but only tested 1 hour again
 
Last edited:

Vic 40

Titan
Ambassador
With the psu that old you still have warranty,is from Antec 5 years. Use it.

Yes only one hour isn't enough to test for degredation of components.

In Hwinfo look at the +12V/+5V and +3.3V voltages tot see if they are within spec and do not fluctuate too much,

QQrUeSE.png


these are the ranges it should work within,but without too much fluctuation. For example on the +12V line not dropping from 12.2V to 11.6V. And taking wide ranges here.
 

Thomas4140

Prominent
Aug 6, 2019
40
0
530
With the psu that old you still have warranty,is from Antec 5 years. Use it.

Yes only one hour isn't enough to test for degredation of components.

In Hwinfo look at the +12V/+5V and +3.3V voltages tot see if they are within spec and do not fluctuate too much,

QQrUeSE.png


these are the ranges it should work within,but without too much fluctuation. For example on the +12V line not dropping from 12.2V to 11.6V. And taking wide ranges here.
Thank you, I really do appreaciate it. Really 5 years? I though it ran out because of 2Y. How do i ask for a new one, can u like order it just with your SN codes? i will test it soon and put an update.
 

Thomas4140

Prominent
Aug 6, 2019
40
0
530
With the psu that old you still have warranty,is from Antec 5 years. Use it.

Yes only one hour isn't enough to test for degredation of components.

In Hwinfo look at the +12V/+5V and +3.3V voltages tot see if they are within spec and do not fluctuate too much,

QQrUeSE.png


these are the ranges it should work within,but without too much fluctuation. For example on the +12V line not dropping from 12.2V to 11.6V. And taking wide ranges here.
I went into the BIOS, hwmonitor and got this from the voltages, could you make something out of it?
Several pics here:

Motherboard on TMPIN5 and 8 = 127° What does that mean?
View: https://imgur.com/a/8Oo9a30
 

Thomas4140

Prominent
Aug 6, 2019
40
0
530
Update: here are the complete logs of HWInfo, running 5 mins:
View: https://imgur.com/a/ndgvTIb


I tried to test with memtest 86 as bootable with my usb drive, but I can't get it to boot on my pc via the BIOS. I did change startup settings to prioritze all usb things before the hard disk.

I ran Memtest64 on maximum test memory for test and it gave me a blackscreen 2X, System, motherboard (msi ms7918) memory load was 93% and than gone back to blackscreen, But no errors in log (only errors when ram goes bad?). Elapsed 10 mins that's my first try probably because i was browsing on the internet while the test ran.
 
Last edited:

Thomas4140

Prominent
Aug 6, 2019
40
0
530
I ran Memtest64 on maximum test memory for test ram and it gave me a blackscreen 2X.

I noticed the System stat:, (msi ms7918) motherboard i think, memory load was 93.3% and than gone back to blackscreen, But no errors in log (It only errors when ram goes bad?). Elapsed 10 mins that's my first try probably because i was browsing on the internet while the test ran.
 

Thomas4140

Prominent
Aug 6, 2019
40
0
530
View: https://imgur.com/a/c1u5sP6
when it got to 95% it blackscreened. So what does that mean, is that normal that % is so high. Also had a small artifact when it blackscreened. this stats where when using maximum memory on the memtest 64 program. Tested for around + 10 mins, 10 loops
 

Thomas4140

Prominent
Aug 6, 2019
40
0
530
What it does say is that your ram fills up since 96% is used. Might be your problem. Look in taskmanager what is taking up soo much ram.


It also say that your gpu drivers aren't used or crashed so maybe clean install those as well.
Oh it was the program: memtest 64: It's used for testing your ram. I maxed out the ram (all 16GB) to be tested. Taskmanager said program ran on 90%, is normal i tested on my bro's computer too, he had as high % as i did. It's just the fact that it blackscreened on my pc that's wondering me. It didn't log any errors tho.

Maybe slots are broken on the motherboard, tho they function still..?
 

Thomas4140

Prominent
Aug 6, 2019
40
0
530
What it does say is that your ram fills up since 96% is used. Might be your problem. Look in taskmanager what is taking up soo much ram.


It also say that your gpu drivers aren't used or crashed so maybe clean install those as well.
The thing is As soon as i install the drivers, it causing the flickering on windows desktop background and ends up in a blackscreen. I completely unistalled all drivers with DDU and manually. I now run on inbound gpu (of the cpu) without issues since that's probably because i do not have drivers installed...
(also reinstalled windows)

Idk what is the issue it's so hard to tell..
 

Thomas4140

Prominent
Aug 6, 2019
40
0
530
I have GTS250 in my computer installed and i already have a small artifact when chatting in steam, when u click on the emoticons icons menu. They look like little green dots.
 
Last edited:

Thomas4140

Prominent
Aug 6, 2019
40
0
530
Does it only artifact after installing drivers?

Already tested with another psu? Or rma's the Antec.
Bought a new psu, I tried to RMA the antec but i don't get any response after 3 days waiting i filled in the europe email/ticket thing, but only get answer from NA website.

Ye mostly after, didnt encounter before i think, and blackscreening a lot during installation.
 

Vic 40

Titan
Ambassador
Already using the new psu? An rma might take some time,you're not the only one they have to look at.

What you can try,since having a psu with the option,is use both pcie cables to power the gpu,use only both first connectors. See if that helps.

When installing drivers will the screen go black a few times so don't count that as "blackscreening".
 

Thomas4140

Prominent
Aug 6, 2019
40
0
530
Already using the new psu? An rma might take some time,you're not the only one they have to look at.

What you can try,since having a psu with the option,is use both pcie cables to power the gpu,use only both first connectors. See if that helps.

When installing drivers will the screen go black a few times so don't count that as "blackscreening".
Ye i bought the Corsair RM750X one, and there's a gtx 560 inside it. Everything seems to be ok for now.
It's kinda weird having a psu with zero rpm mode technology, I am so used to see a fan being spinnend (to cool psu) and it worries me if its correctly installed all the time xD/ PSU seems to be cold temp.

It's been 6 days and I havent heard anything from Antec in that e-mail or form message thing your supposed to write if you are from EU. Meanwhile on the mainwebsite (for NA only) i got replied within 24 hours. Not even a simple confirmation automatic email saying they recieved my mail. I doubt i'll get answer.
 

Thomas4140

Prominent
Aug 6, 2019
40
0
530
What you can try,since having a psu with the option,is use both pcie cables to power the gpu,use only both first connectors. See if that helps.

When installing drivers will the screen go black a few times so don't count that as "blackscreening".

Ye it blackscreened 2 times fast when installing drivers but that was normal. At first i though i was going to relive that everlasting blackscreen lol. My issue previous was that it stays in a loop blackscreening and than goes completely black or restarts pc.

I have the psu installed by a tech local guy, since my case is too small to take out the psu without taking out cpu cooler/mobo. He has installed it with only 1 PCI e cable ( i think ) that splits up by the end of the cable in 2 seperate blocks: a 8 and 6 pin.

Should I take 2 PCI cables ? Is it recommended to do that ?
since i think my psu doesn't have extra holes anymore to plug it into: https://tweakers.net/i/pIWIQ7o5_awugW08IX-JZ8O5fWk=/i/2000761340.png
I think i have remaining holes in the peripharel sata, the 3rd block on the left is it for sata or pci?
Since the name is closer to sata, i think the 3rd block is an additional sata port too?

The 2 to the right (where the namers are pci and cpu) are completely filled in