[SOLVED] Serious issues with brand new RTX 2060

[dratyan]

Background: I actually had issues with my previous GPU only 3 months ago, most of it is detailed in this reddit thread. TL;DR of that: 11 months old GTX 1060 gave me black screen; tech store said it also wasn't working for them; RMA'd the GPU to Zotac but they still haven't given me a return.

I ended up buying an RTX 2060, it arrived today, installed it plus drivers after running DDU. I played some lighter games such as Borderlands 2, Guild Wars 2 and Dota 2, perfect. Decided to play newer ones: Watch Dogs 2, Kingdom Come Deliverance and Injustice 2.

Screen got all glitched and my PC turned off a minute into WD2; the same thing happened a few more times during Windows boot until it worked. Tried KCD and it lasted a few minutes more but then I got a BSOD. This time it didn't shutdown while loading Windows afterwards.
That was yesterday, today I tried Injustice 2, BSOD again, 5 minutes in, still in the menus. PC reboots and there are some green artifacts on the screen that soon vanish. Now it's on but I can't change the resolution from 800x600.

• I was checking temperatures during gameplay and everything was normal.
• PSU is supposed to be enough if online calculators are correct.
• Only one PCI-E slot on the MB so can't swap it.
• Blue Screen of Death has no error code. It just says Windows had to shut down, "=(".
• Does the GPU have a factory defect? If not what on my PC could be causing this?

I'm really lost here. Spent more than I should to get this GPU because I didn't know when I'd get the other back, and now already having big issues on the first day. Below are my specs, the rest of the build is ~7 years old. Thanks.

TL;DR: Getting BSOD when trying heavier games with new GPU. Temperatures seem normal.

GPU: EVGA NVIDIA GeForce RTX 2060 XC Black Gaming 6GB, GDDR6
CPU: i7-2600
RAM: 8GB DDR3 1333Mhz
MB: ASUS P8H61-M LE/BR
PSU: Corsair CX 600W
Storage: 1x HDD, 2x SSDs
OS: Windows 10

 

[COLGeek]

How old is the PSU?

That old mobo may have issues with newer GPUs as well (just a thought).

 

[dratyan]

How old is the PSU?

That old mobo may have issues with newer GPUs as well (just a thought).

The PSU is a bit newer than the rest, around 5 years old.

The 1060 worked fine in this setup(for 11 months until it just died), is there anything specific about the 2060 that would make it not work?

Biggest problem with changing the mobo would be that I'd either have to get a new CPU(LGA 1151) and RAM sticks (DDR4) or get used parts(which I'm not a fan of thanks to past experiences). And that would mean an even bigger blow to my budget. So I really want to be sure before having to buy more parts.

 

[COLGeek]

How does the system perform if you remove the 2060 and use the integrated video adapter? Is it stable then?

 

[dratyan]

How does the system perform if you remove the 2060 and use the integrated video adapter? Is it stable then?

It's stable, yes. I have been using the Intel HD 2000 for the last 3 months, didn't have any issues. Couldn't really play any games or do anything graphically demanding, of course.

Just now I tried a beefy game and again the PC crashed quickly, this time it skipped the BSOD. I was tracking everything I could with Afterburner and there was nothing out of normal.

Using the PC now, post crash, I see that the GPU has been disabled: "Windows has stopped this device because it has reported problems. (Code 43)". There are also some graphical artifacts on the screen like this.

I did install the latest Windows update, 1903, that I was postponing but got same results. Now I'm going to remove and reinsert the GPU once more to see if anything changes.

 

[Deleted member 2731765]

It looks like there is some problem with your Video card, IMO. Can you check the card by using some other PSU ?

Error code 43 is a generic message that typically indicates a problem with the hardware or software in Windows operating system. This error is often reported in Windows Device Manager. If you encounter this error with your graphics card, here are some troubleshooting ideas that may help fix it.

1) Update your graphics card driver.

2) Make sure your system is updated with the latest SBIOS and Windows updates

3) Check to make sure your system power supply meets the minimum recommended power requirement for your card

4) If your card needs additional power, make sure the PCIe supplemental power connector is plugged into the card

5) Replace the graphics card as it may be defective.

Code 43 can mostly happens either if drivers fail or the hardware fails. Other option is that your GPU or motherboard slot failed. If your motherboard has more then one PCIe x16 slot, you can move GPU to another slot to rule out motherboard failure. Or you should test GPU in another machine to confirm that problem lies in the card.

The above Error code 43 can be reported anytime when Windows does not recognize a device – whether it's a USB port, Hard drive, video card/GPU, or any other piece of hardware connected to the computer. Most of the time, error code 43 means that a device driver is either missing, corrupt, and/or outdated (so Windows has problems communicating properly with the hardware), or the hardware is faulty and needs to be replaced.

This error code points out that the Hardware might also be defective, which is the GPU. But before we jump to any conclusion, check the card on some other PC as well.

 

[dratyan]

It looks lie there is some problem with your Video card, IMO. Can you check the card by using some other PSU ?

I don't have access to any other PSU or desktop to test the GPU, but I can try to find a way on Monday. Would still like to look for answers until then.

1) Update your graphics card driver.

Already cleaned drivers with DDU and reinstalled new ones multiple times. Gonna do it one more time after Windows/BIOS updates.

2) Make sure your system is updated with the latest SBIOS and Windows updates

Just updated Windows, gonna update BIOS now as well.

3) Check to make sure your system power supply meets the minimum recommended power requirement for your card

Those online calculators gave me a ~500W requirement, mine's a 600W one. Seems fine, although the PSU is a bit old.

4) If your card needs additional power, make sure the PCIe supplemental power connector is plugged into the card

Not the case.

5) Replace the graphics card as it may be defective.

I sure hope not. If that's the case, I don't really know why though. Games that were neither light or heavy worked perfectly, and heavy ones are running as expected(Ultra and 60fps), even though they only last for five minutes tops before it all crashes.

 

[Deleted member 2731765]

Not the case.

So your GPU does not require an external PCI/PEG power connector, 6 or 8-PIN cable ?

 

[COLGeek]

The RTX 2060 requires an 8-pin PCIe power connector be attached to it. Is that not the case here?

 

[Dope Senpai]

So your GPU does not require an external PCI/PEG power connector, 6 or 8-PIN cable ?

It requires an 8 pin

 

[Deleted member 2731765]

It requires an 8 pin

Yes, obviously I know that. The OP responded differently, by saying "not the case", which had me confused. That's why I wanted to confirm this from his side.

 

[dratyan]

So your GPU does not require an external PCI/PEG power connector, 6 or 8-PIN cable ?

The RTX 2060 requires an 8-pin PCIe power connector be attached to it. Is that not the case here?

Sorry, might've been a misunderstanding here. This GPU requires an 8-pin cable, which I've plugged in, replugged, tested with the other PCI-E cable, etc. I thought it was referring to an extra PCI-E cable, like some of the 2070/2080 models that have 12 or more slots. So yes, the 8-pin cable has always been on.

Also I've now updated the BIOS as instructed, no improvements however.

 

[Deleted member 2731765]

The RTX 2060 requires an 8-pin PCIe power connector be attached to it. Is that not the case here?

Yes, obviously it requires an 8-pin cable. The OP actually responded differently, by saying "not the case", which had me confused. That's why I wanted to confirm this from his side, and to make sure he plugged the connector to the GPU.

 

[dratyan]

Just a quick update. So what had been happening until now, summarized:

• Soon after starting to play a very graphically intensive game, PC shuts down, sometimes to a BSOD. When it restarts, the GPU has a Code 43 on it.

Just now though, I was browsing the internet for solutions and the PC went completely dark: both image and sound off. I thought it'd shut down, but after 10secs it all came back to how it was. So that's a new symptom.

 

[COLGeek]

Just a quick update. So what had been happening until now, summarized:

• Soon after starting to play a very graphically intensive game, PC shuts down, sometimes to a BSOD. When it restarts, the GPU has a Code 43 on it.

Just now though, I was browsing the internet for solutions and the PC went completely dark: both image and sound off. I thought it'd shut down, but after 10secs it all came back to how it was. So that's a new symptom.

Heat and/or power generally cause this to happen. Either the CPU or GPU.

Faulty memory can also cause these issues.

In this case, the old tech motherboard may also be part of the problem, using/powering the much newer 2060. Just a hunch.

 

[dratyan]

Heat and/or power generally cause this to happen. Either the CPU or GPU.

Faulty memory can also cause these issues.

In this case, the old tech motherboard may also be part of the problem, using/powering the much newer 2060. Just a hunch.

Temperatures were low even while in-game, hard to believe that was the case when the PC is almost idle. Unless the temperature sensors are somehow busted. Power supply is definitely a possibility, is there any way to check something like that without specialized tools?

Would memory issues show up only after installing a new GPU though? Gonna check them with memtest86.

My biggest fear #1 is buying a new mobo(+CPU + RAM because they don't make LGA 1155 and DDR3 stuff anymore) and/or PSU and end up having the same issues. If money wasn't an issue then no worries, but right now it is.

Biggest fear #2 is somehow getting into another RMA situation(my old GPU has been with Zotac for the last 3 months and I may have to contact FTC or get a lawyer) that either takes a lot of time/resources to solve or that the same issues show up with the replacement GPU.

I know some of this stuff is off-topic, just trying to show why I want to be as sure as possible before going for extra purchases or replacements. Also, thanks for all the help so far guys.

 

[dratyan]

Some more incidents to add that may be relevant:

• Tried to play one of the older games that worked fine yesterday. This time it didn't, another shutdown. All recent shutdowns are BSOD-free.
  
  
• On the restart, I got a "CPU overheating" error before the BIOS screen. That came out of nowhere, as CPU temps never reached 80ºC while monitoring. Anyway, opened the case and pointed a fan at it to see if it'd help. Tried once more to play WD2, and this time it worked?! Managed to play for around 15min - a lot compared to the previous 2min. Suddenly, GPU usage dropped to 0% - even though it's temperature was around 60ºC and CPU temperature around 70ºC. FPS dropped to 1 so I had to close the game, and although there was no shutdown, the PC was VERY slow for a minute or two.
  
  
• Went to a different game, Apex Legends. Before getting into a match, got an error from the game (0x887A0006 - DXGI_ERROR_DEVICE_HUNG) and immediately after my PC shutdown once again, this time to a BSOD with an error code: VIDEO_TDR_FAILURE(nvlddmkm.sys).

Digging into these new error codes now in search of new information. Will try to reseat the CPU and the heatsink, even though it'd be pretty weird for it to cause a shutdown while being <80ºC. Also want to test/benchmark everything I can: GPU, RAM, CPU,HDD, SSD. Accepting suggestions for good testing tools. For now will use MemTest86 and FurMark.

EDIT: Here is a screen of GPU-Z of the GPU in near idle mode, just in case there's any oddity.

 

[dratyan]

Got another BSOD while just watching a video, with a different code this time: VIDEO SCHEDULER INTERNAL ERROR. Also getting random freezes.

At this point I don't know if this is fixable and have been torn between three unwanted options:

• RMA the card
• Buy a new PSU
• Buy a new motherboard(+CPU+RAM)

 

[COLGeek]

Do you have access to another PC to test the 2060 in? If so, you can rule it in/out easily.

 

[dratyan]

Do you have access to another PC to test the 2060 in? If so, you can rule it in/out easily.

Not at the moment. I'd have to take it to a shop tomorrow, pay them and probably wait until the end of the week for results. I know that's likely the best way to find answers, I'm just trying to do what I can now.

 

[dratyan]

Sorry about all theses messages, simply trying to paint the whole picture here. While running the free 3DMark benchmark, I got a different BSOD. That's now 3 different BSOD error messages:

• VIDEO_TDR_FAILURE(nvlddmkm.sys)
• VIDEO SCHEDULER INTERNAL ERROR
• DPC WATCHDOG VIOLATION

 

[Deleted member 2731765]

Have you overclocked any of your PC components ?

 

[dratyan]

Have you overclocked any of your PC components ?

No. Never messed with any of that.

 

[dratyan]

So I've just now completed a complete reinstall of Windows, wiped every drive clean, supposed to be a completely clean install, cause why the hell not at this point, 55 hours into this troubleshoot. I go to the device manager and the first thing I see: Code 43 on the GPU. And Windows apparently already installed the latest drivers somehow, before I could even boot? Is that how Windows 10 works nowadays?

Also something I've been noticing: PC won't turn off completely when I do tell it to shutdown. Nothing on the screen and inputs don't work but the fans, lights, etc stay on.

 

[COLGeek]

So I've just now completed a complete reinstall of Windows, wiped every drive clean, supposed to be a completely clean install, cause why the hell not at this point, 55 hours into this troubleshoot. I go to the device manager and the first thing I see: Code 43 on the GPU. And Windows apparently already installed the latest drivers somehow, before I could even boot? Is that how Windows 10 works nowadays?

Also something I've been noticing: PC won't turn off completely when I do tell it to shutdown. Nothing on the screen and inputs don't work but the fans, lights, etc stay on.

The shutdown issue can (likely) be cleared up by installing the motherboard chipset drivers. Get those from Intel:

https://www.intel.com/content/www/us/en/support/intel-driver-support-assistant.html

Run this tool and update all applicable Intel drivers.

Windows will install default drivers. You'll need to manually install the Nvidia drivers.