Gpu only runs when i under-volt it

[Hunny.Badger]

Recently got a used computer after reformatting it my pc will crash with a black screen and the fan running louder whenever I run any game unless I under-volt my gpu in MSI AFTERBURNER.

 

specs:

i5-9400

RTX 2060

Litepower 550w power supply

 

 

 

[leclod]

Did you try without that white extension cable ?

 

Airflow is weird, could you show Airflow direction on each fan with an arrow ?

Edit : ok, I think I got it. I'd try without the top fan.

Also for fun, I'd try to reverse the airflow from case fan and cpu fan towards the psu this time with top fan as intake (less optimal for dust)

Edited October 24, 2023 by leclod

[Hunny.Badger]

1 hour ago, leclod said:

Did you try without that white extension cable ?

 

Airflow is weird, could you show Airflow direction on each fan with an arrow ?

Edit : ok, I think I got it. I'd try without the top fan.

Also for fun, I'd try to reverse the airflow from case fan and cpu fan towards the psu this time with top fan as intake (less optimal for dust)

Tried it before ,same result unfortunately 

[leclod]

6 minutes ago, Hunny.Badger said:

Tried it before ,same result unfortunately 

Your 550W psu is truly a 456W psu. Maybe that's the problem and your psu is too weak.

Your system works though...

Edited October 24, 2023 by leclod

[Hunny.Badger]

On 10/25/2023 at 12:43 AM, leclod said:

Your 550W psu is truly a 456W psu. Maybe that's the problem and your psu is too weak.

Your system works though...

Im sorry I'm not very savvy in pc ,from my understanding of what you're saying I need a bigger psu?

[leclod]

4 hours ago, Hunny.Badger said:

Im sorry I'm not very savvy in pc ,from my understanding of what you're saying I need a bigger psu?

No, 450W should still be more than enough. I just checked, sorry if I worried you.

If the undervolted system is stable, you could remain that way.

Or enquire further...

 

[Hunny.Badger]

1 hour ago, leclod said:

No, 450W should still be more than enough. I just checked, sorry if I worried you.

If the undervolted system is stable, you could remain that way.

Or enquire further...

 

I'd say it crashes 1 out of 10 times now when I start a game up after undervolting .

[leclod]

16 minutes ago, Hunny.Badger said:

I'd say it crashes 1 out of 10 times now when I start a game up after undervolting .

That's no good,

Ideas :

reseat, replug everything,

reinstall Windows

check your ram with memtest86 or testmem5

check your storage with crystaldiskinfo

how are your temps ?

[Hunny.Badger]

3 hours ago, leclod said:

That's no good,

Ideas :

reseat, replug everything,

reinstall Windows

check your ram with memtest86 or testmem5

check your storage with crystaldiskinfo

how are your temps ?

What I tried earlier was swapping it for a 1070 which everything went smoothly ...do you think I bought a broken GPU?

[leclod]

7 minutes ago, Hunny.Badger said:

What I tried earlier was swapping it for a 1070 which everything went smoothly ...do you think I bought a broken GPU?

Me, I doubt gpus get broken without bad handling.

But, you could try that 2060 in another system...

[IkeaGnome]

1 hour ago, Hunny.Badger said:

What I tried earlier was swapping it for a 1070 which everything went smoothly ...do you think I bought a broken GPU?

What are the temps on the 2060 when it's not undervolted?

GPU-Z Graphics Card GPU Information Utility (techpowerup.com)

If you run GPU-Z there is a tab for sensors. That's what I'm curious about. Can you upload a picture of that tab and the main tab?

 

Edit:

Your frequencies look very weird to me. This looks like you've also got an overclock running on both your core and VRAM. Can you show us your MSI Afterburner main page?

[alyen]

What brand is the RTX 2060. It looks like a generic or OEM type that Dell has. Did it it come like that without a back plate?

[Hunny.Badger]

22 hours ago, alyen said:

What brand is the RTX 2060. It looks like a generic or OEM type that Dell has. Did it it come like that without a back plate?

It's a dell

[Hunny.Badger]

22 hours ago, alyen said:

What brand is the RTX 2060. It looks like a generic or OEM type that Dell has. Did it it come like that without a back plate?

It's a dell

[Hunny.Badger]

23 hours ago, IkeaGnome said:

What are the temps on the 2060 when it's not undervolted?

GPU-Z Graphics Card GPU Information Utility (techpowerup.com)

If you run GPU-Z there is a tab for sensors. That's what I'm curious about. Can you upload a picture of that tab and the main tab?

 

Edit:

Your frequencies look very weird to me. This looks like you've also got an overclock running on both your core and VRAM. Can you show us your MSI Afterburner main page?

The previous owner told me to set it up these way to prevent crashes

[Hunny.Badger]

On 10/26/2023 at 10:34 PM, leclod said:

Me, I doubt gpus get broken without bad handling.

But, you could try that 2060 in another system...

Unfortunately I dont have a spare system to test it .

[leclod]

5 minutes ago, Hunny.Badger said:

The previous owner told me to set it up these way to prevent crashes

The previous owner had crashes already ?

[alyen]

28 minutes ago, leclod said:

The previous owner told me to set it up these way to prevent crashes

Thats a bad sign if you have to undervolt it. By default it should be working in stock setting. Components could have been stressed out from mining or maybe it got flashed to some different setting at some point.

[IkeaGnome]

42 minutes ago, Hunny.Badger said:

Unfortunately I dont have a spare system to test it .

There's some very, very weird readings on that MSI afterburner page. Can we see the main page for GPU-Z on that 2060? The one that shows bios revision and all that?

[YoungBlade]

What I'd like to see is what the GPU hotspot temp when you're actually running a game.

 

I'm wondering about that because, if the GPU works when undervolted, but not otherwise, then it could be that a part of the GPU die is getting too hot, and undervolting is reducing the temperature enough to keep it from crashing.

 

Can you show the HWInfo64 table for the card's temperatures when it's under load playing a graphically demanding game?

[yazmo]

 

       Hey folks, here's my two cents on this. I'm facing the same exact problem, along with a bunch of others

that I'd love to chat with Linus about – could make a great video!

 

So, about this issue: I've got a PNY GeForce® RTX 2060 SUPER™ 8GB XLR8 that crashes, even on an old game like TF2.

The temperature never goes over 40 degrees. A few years back, I opened up the GPU, cleaned it, and reapplied the thermal  Some of the thermal pads weren't in great condition, but I reused them because this card has like 5 different sizes, and buying new ones can be pricey.

 

Especially considering all the things I've been through building my PC over the years, but that's a story for another time.

 

Now, onto the specs of my desktop PC, which still has two known issues to this day. First, there's a sound pop that causes games to stutter on certain sounds, even in-game, which is annoying AF. And second, there are frame drops despite trying several HDMI display ports, etc. I'm also using a 15-foot USB extension for everything because the PC is that far away – I hate noise.

I've tried many, many cables – quality, shorter ones, you name it. Here are the specs:

• Corsair V650 PSU
• Intel i7 8700k (not overclocked now, back to stock)
• Aorus z390 Ultra ver F10 (tried all versions but not the latest yet, F11)
• G.SKILL Ripjaws V Series 16GB (2 x 8GB) DDR4 3600 (OC still works but it's off for testing purposes)
• Noctua NH-D15 chromax.black Dual-Tower CPU Cooler with Dual NF-A15 PWM 140mm Fans (Black)

I've got 4 HDDs, but one I hotplug often, which I stopped because I believe it wasn't well grounded to the chassis and ended up breaking one of the HDDs that was sharing the same power cable. I wasn't ejecting the drive either – too lazy, I know

Using a Cooler Master case with a plastic slider didn't help with the grounding to the chassis.

There are many minor details that often go unmentioned but can cause issues – trust me, I know, I've spent a lot of time on this.

 

Now, onto the GPU issue. It goes black screen, but I can still hear the game sound, and the PC doesn't crash besides the dark screen. But I have to turn it off to see the screen again. If I check the Event Viewer, I get an error related to nvlddmkm.

 

It's getting interesting, maybe the thermal pads are failed on me, yet I haven't tried flashing the GPU BIOS yet. Undervolting does help with the crash as well. The PSU might not be enough – it's got a few years on it, but it's 650 watts. Today, I got a 1000W PSU, so I'll test this out, but I doubt it.

 

Formatting didn't fix this issue, which has been going on for quite some time on this PC. I have very good FPS in games, but stutters happen randomly. I suspect the sound drivers, which are integrated into the board. I was thinking of buying a sound card in 2024 – WTF – to see if this would fix it, but again, all I've read about those are bad drivers for Windows 10. Spend again?

I dunno. Sadly, I'm like this – I want to know what causes those issues that many, many of us have.

My two cents are to repaste and change thermal pads, and maybe get a new PSU.

And for these random shutter-over sound events, seems like a bad integrated sound card with poor drivers.

Realtek drivers coming from this motherboard are bad.

 

 

 

[yazmo]

Holy Batman Reviving an OLD Post heheheh