PC reboots after changing graphics card

[arwellai]

Hello everyone.

After replacing my GTX 1060 with new Gigabyte RTX 2070 Super 8Gb Gaminc OC, my PC started rebooting while gaming from time to time.
Its doing great in tests, tried AIDA64 and Furmark for couple of hours each, so reboots happen only in games. Temperatures never get higher than 63-64 deg. Load in game i play is 70 percent at max.
Tried: 1) uninstalling drivers with DDU, then installing from official website without Geforce Experience;
2) resetting BIOS settings to default;
3) used Memtest86+ for half a day, everything fine;
4) cleaned PSU and tested in with multimeter;
5) tried placing card in x4 slot;
6) installed chipset drivers;
7) reinstalled Windows, it helped for about 2 days, then reboots started to happen again.

My specs: Gigabyte GA-Z170-HD3, i5-6600k, RTX 2070 Super 8Gb Gaminc OC, Thermaltake Evo Blue 2.0 750W, 16Gb DDR4 3000MHz Corsair Vengeance LPX.

Any ideas? I have only one left is to bring card back to vendor to test it, but i assume they just run some tests and say it works fine because it does pass any test without errors.
Thanks.

BIOS is of latest version F22f btw, updated about a year ago.

 

[Ruyki]

I would suspect the PSU or the new video card.

Check windows error logs if there is something related to the reboots.

Some nvidia cards allow you to set the power limit below 100%. Setting it below 100% should lower the power consumption of the card. If the reboots stop after you do this, it could indicate that the PSU has trouble supplying enough power.
You can also run a CPU and GPU stress test at the same time. If done correctly, this will load the system and the PSU more than by running a single component stress test. If reboots happen when you do this, it could indicate that the PSU has trouble supplying enough power.
Best way to check would be replacing the PSU with a known working unit if you have one on hand, but I'm guessing you don't.

Also, check that all power cables in your system are correctly plugged in.

 

[Chomiq]

How old is the PSU? Make sure you have controlled test environment, if everything worked for two days after reinstalling what changed?

 

[eidairaman1]

What's the specific power supplymodel number? Get a picture of the entire power supply lable.

 

[thebluebumblebee]

That PSU was tested on Tweaktown, in 2013, with good results. It shouldn't be the weak link, but to me this is a power problem.

OP, I'd suggest unplugging and re-plugging the 24 pin power connector. That PSU comes with a 5 year warranty, so if it's older than that, you may want to consider replacing it.

reboots happen only in games

Is it the same game, or does it happen with multiple games?

What GPU are you upgrading from?

 

[Zach_01]

PSU is rated as Tier-B on this list (for entry level systems), and should by doing the job(?), but I'm not sure that 2070S is entry level card... guess not!

[EOL] PSU Tier List rev. 14.8

PSU Tier List 4.0 rev. 14.8 (END OF LIFE) Last Update: 27-07-2021 Legend : Gray - EoL/obsolete and/or otherwise not recommended for purchase. Green - small form-factor (gold and blue colors are disregarded due to scarcity of SFX PSUs) Gold - best units in the tier (includes requirements for blue ...

linustechtips.com

 

[thebluebumblebee]

BTW, is there any way it could be your power source?

 

[FreedomEclipse]

PSU is rated as Tier-B on this list (for entry level systems), and should by doing the job(?), but I'm not sure that 2070S is entry level card... guess not!

[EOL] PSU Tier List rev. 14.8

PSU Tier List 4.0 rev. 14.8 (END OF LIFE) Last Update: 27-07-2021 Legend : Gray - EoL/obsolete and/or otherwise not recommended for purchase. Green - small form-factor (gold and blue colors are disregarded due to scarcity of SFX PSUs) Gold - best units in the tier (includes requirements for blue ...

linustechtips.com

Yeah but the PSU was released in 2012. Thats still an old PSU technically.

 

[DeathtoGnomes]

Yeah but the PSU was released in 2012. Thats still an old PSU technically.

manufacture date is what matters. it could have sat on the shelf for a few years too. Smaller shops might refurbish without and still try to sell as new.

 

[Deleted member 194470]

Power supplies usually last a long time if they are good. It could be the power cable or you didn't plug in the card right. I had this problem with cards that I installed. I wasn't use to the special power adapters for the video card and I didn't have it plugged in all the way.

 

[mbeeston]

might be the hdd the games are on starting too give out.

 

[neatfeatguy]

I've seen odd power issues. A similar case where a guy had a space heater plugged into the same circuit breaker (different outlet on the same wall, and the space heater sat on the other side of the room) his PC was plugged in - heaving game with the space heater going was causing his PC to power off and back on. He said his temps weren't over 70C on his GPU and his CPU wasn't going over 60C, so overheating wasn't the culprit. After he finally unplugged the space heater the rebooting issue on his PC stopped.

 

[arwellai]

I would suspect the PSU or the new video card.

Check windows error logs if there is something related to the reboots.

Some nvidia cards allow you to set the power limit below 100%. Setting it below 100% should lower the power consumption of the card. If the reboots stop after you do this, it could indicate that the PSU has trouble supplying enough power.
You can also run a CPU and GPU stress test at the same time. If done correctly, this will load the system and the PSU more than by running a single component stress test. If reboots happen when you do this, it could indicate that the PSU has trouble supplying enough power.
Best way to check would be replacing the PSU with a known working unit if you have one on hand, but I'm guessing you don't.

Also, check that all power cables in your system are correctly plugged in.

Im not sure what to look for in logs, i can upload it if it can help.

I was doing every single stress test in AIDA64 and furmark at the same time for about 2 hours, i thought that makes sure PSU is not the case.

Also i replugged every single cable in my PC while trying to solve this.

How old is the PSU? Make sure you have controlled test environment, if everything worked for two days after reinstalling what changed?

PSU is about 4 years old.

Yeah i guess if there is no other solution ill try to reinstall WIndows again. Thing is, i didnt do anything specific before reboots started to happen again.

That PSU was tested on Tweaktown, in 2013, with good results. It shouldn't be the weak link, but to me this is a power problem.

OP, I'd suggest unplugging and re-plugging the 24 pin power connector. That PSU comes with a 5 year warranty, so if it's older than that, you may want to consider replacing it.

Is it the same game, or does it happen with multiple games?

What GPU are you upgrading from?

Replugged every single one of them.

I play Witcher 3, for testing purposes downloaded also RDR2 and Metro Exodus. Metro was fine, didnt want to play long tho, RDR2 rebooted during start intro, so i stopped trying.
As it reboots randomly from time to time and almost always at certain moments, i thought there's something with software. Maybe reinstalling BIOS on graphics card can help?

might be the hdd the games are on starting too give out.

Tried installing a game on both SSD and HDD.

 

[thebluebumblebee]

Yeah i guess if there is no other solution ill try to reinstall WIndows again

No, just no,

almost always at certain moments

That's telling, but it doesn't tell us exactly what's causing it. PSU? GPU? Motherboard? (yes, really) Where/what you have the system plugged in/to.

What GPU are you upgrading from?

The troubleshooting steps from here are to exchange parts, starting with the PSU. You really only need a 550 watt PSU for that system.

 

[arwellai]

No, just no,

That's telling, but it doesn't tell us exactly what's causing it. PSU? GPU? Motherboard? (yes, really) Where/what you have the system plugged in/to.

What GPU are you upgrading from?

The troubleshooting steps from here are to exchange parts, starting with the PSU. You really only need a 550 watt PSU for that system.

I mentioned in 1st message, previous Gpu was gtx 1060. PC connected to common surge protector or w/e its called.

Yeah I still dont get it how it may be PSU fault considering it passes couple of hours long stress tests.
First i think ill give GPU to a friend to test it. Want to mention again: problem started only after i changed graphics card.

 

[eidairaman1]

Put old card back in, however if the gpu exibits no problems in froends machine or in a shop its your rig, could be ram even, disable windows automatic restarts afer crashes

 

[arwellai]

Put old card back in, however if the gpu exibits no problems in froends machine or in a shop its your rig, could be ram even, disable windows automatic restarts afer crashes

Sold old one right away.
I hope i can force my friend to play witcher for me
Automatic restarts disabled, it doesnt BSOD, it just instantly reboots.

 

[eidairaman1]

Got temperature monitoring tool,

 

[John Naylor]

PSU Tier Lists are poorly evaluated and I can not recommend using them. For example, you will find an entire model series put on the list based upon one review of one model from that series. And quite often, the lower sizes in the model line are made by a different OEM and / or on a different platform. For example:

Corsair's TX line came from 2 different OEMs CWT / Chicony) using 4 different platforms
Corsair's HX series came from 2 different OEMs (Seasonic / CWT) using 5 different platforms
EVGAs G2 series came from 3 different platforms ... same for P2

1. Is anything overclocked ? If so drop to out of box settings and turn off XMP for comparison purposes
2. How many different games have exhibited this behavior ?
3. If it can handle Furmark, take the PSU outta the equation ... you used AIDA 64 and Furmark at same time ... few systems can handle that, it's not the PSU.
4. The budget MoBo is a bit out of balance with the CPU / GPU
5. If automatic re-start is disabled ... it should present an error screen. Note the time of failure / report and thenafter reboot look in Evenbt Vuewer Windows system logs for description of err event(s) ... anything in yellow (warning) or res (error) should be investigated.
6. Try in another box.

 

[arwellai]

Got temperature monitoring tool,

Yeah, temperatures never were higher than 63-65 both on GPU and CPU during stress tests. Its even lower in games since Witcher on ultras only loads about 60% of CPU and GPU.

PSU Tier Lists are poorly evaluated and I can not recommend using them. For example, you will find an entire model series put on the list based upon one review of one model from that series. And quite often, the lower sizes in the model line are made by a different OEM and / or on a different platform. For example:

Corsair's TX line came from 2 different OEMs CWT / Chicony) using 4 different platforms
Corsair's HX series came from 2 different OEMs (Seasonic / CWT) using 5 different platforms
EVGAs G2 series came from 3 different platforms ... same for P2

1. Is anything overclocked ? If so drop to out of box settings and turn off XMP for comparison purposes
2. How many different games have exhibited this behavior ?
3. If it can handle Furmark, take the PSU outta the equation ... you used AIDA 64 and Furmark at same time ... few systems can handle that, it's not the PSU.
4. The budget MoBo is a bit out of balance with the CPU / GPU
5. If automatic re-start is disabled ... it should present an error screen. Note the time of failure / report and thenafter reboot look in Evenbt Vuewer Windows system logs for description of err event(s) ... anything in yellow (warning) or res (error) should be investigated.
6. Try in another box.

1. CPU was overclocked since i bought it using default option in BIOS, as i said before i tried resetting BIOS setting to default, thus removing any overclocking.
2. I play Witcher 3 atm, tried Metro Exodus also for about 2 hours, it was fine, and Red Dead Redemption 2 - PC rebooted during starting cinematic so i stopped trying it.
3. Yeah i thought the same way, voltages measured with multimeter were also perfectly fine.
4. Yeah it was fine before upgrading GPU, now it seems a bit weak, but i would've accepted performance decrease as a result of some sort of bottleneck, not these reboots.
5. No as i said there's no BSOD as if there's not enough voltage on CPU during overclocking, it just insta reboots. Should i post event logs here when i get home? As i recall It just says something like "kernel power code 41", i guess it just states the fact of PC shutting down not as it should be.
6. Didnt understand this one

 

[IceShroom]

Looks like PSU problem related to transient response.
Nvidia card throttle when it sees Furmark. Also it put linear load with not much transient.
Try Unigines Superposition and see if it exhibit same problem.

 

[arwellai]

Looks like PSU problem related to transient response.
Nvidia card throttle when it sees Furmark. Also it put linear load with not much transient.
Try Unigines Superposition and see if it exhibit same problem.

Ok ill try it. Btw i also tried decreasing GPU frequency by max. of 500Mhz with MSI Afterburner, didnt work.

 

[Calmmo]

I had issues on my 2070s rog strix too, restarts, turns out it was 2 problems. DP cable that came with the monitor (LG 27GL850-b) went bad after about a month on top of issues caused by using LG's monitor driver.
New vesa certified cable and stock windows driver fixed all problems. If you can go over HDMI you could try that.

 

[arwellai]

I had issues on my 2070s rog strix too, restarts, turns out it was 2 problems. DP cable that came with the monitor (LG 27GL850-b) went bad after about a month on top of issues caused by using LG's monitor driver.
New vesa certified cable and stock windows driver fixed all problems. If you can go over HDMI you could try that.

Didnt install any monitor drivers and its connected via HMDI cable which monitor was complete with.

 

[eidairaman1]

Yeah, temperatures never were higher than 63-65 both on GPU and CPU during stress tests. Its even lower in games since Witcher on ultras only loads about 60% of CPU and GPU.


1. CPU was overclocked since i bought it using default option in BIOS, as i said before i tried resetting BIOS setting to default, thus removing any overclocking.
2. I play Witcher 3 atm, tried Metro Exodus also for about 2 hours, it was fine, and Red Dead Redemption 2 - PC rebooted during starting cinematic so i stopped trying it.
3. Yeah i thought the same way, voltages measured with multimeter were also perfectly fine.
4. Yeah it was fine before upgrading GPU, now it seems a bit weak, but i would've accepted performance decrease as a result of some sort of bottleneck, not these reboots.
5. No as i said there's no BSOD as if there's not enough voltage on CPU during overclocking, it just insta reboots. Should i post event logs here when i get home? As i recall It just says something like "kernel power code 41", i guess it just states the fact of PC shutting down not as it should be.
6. Didnt understand this one

41 is going to be the psu/motherboard