ASRock AMD Radeon RX 5500 XT Causes Computer To Lose Display Until Forced Reboot

[grimacelord]

Hello. I have an AMD Radeon RX 5500 XT Phantom Gaming D 8G OC video card that recently began to give me trouble after 1 year of perfectly fine usage.

The problem initially began after a year of use as the display randomly losing signal (i.e. screen goes black). The PC's fans would spin high, but the display would never recover until I hard shut down the PC and then rebooted. This would happen randomly and not too often.

Over time, this problem became more and more frequent to the point where any moderate use of the video card in any capacity would trigger a black screen.

I initially suspected a graphics driver issue, so I used DDU to remove all drivers and then clean install the newest graphics driver downloaded from AMD.com. This did not solve the issue. I even experimented with trying older drivers going back 2 years and none of them changed the way the card behaved.

I then tried searching for a VBIOS update, but I already had the latest one available for this card.

So then I suspected that maybe it wasn't a graphics card problem at all, but rather, a power supply problem because I had several customer computers I was repairing that behaved the same way and it turned out it was a failing power supply.

So, I bought an EVGA 650W 80+ Gold power supply and installed that. No change.

Next, I took the graphics card out of my computer and installed it into a computer I have at my office and sure enough, the card performs the same way at my office. So now I know it's the card.

I feel like I should mention this isn't an issue due to overheating. This card was in my personal computer case with 6 fans and liquid cooling on the CPU. The card never, not once, reached above 65C.

I then tested the VRAM with both OCCT and GpuMemTest and both of them caused the black screen problem within a few seconds of running, but I was able to catch hundreds, if not, thousands of VRAM errors before the black screen happened. So, this leads me to believe that one or more of the VRAM chips are bad.

This card uses four 2GB Samsung k4zaf325bm-hc14 making of total of 8GB. I have already sourced replacement ones at about $19 each, but I wanted to see if anyone had any other suggestions for me before I just start replacing VRAMs. I do have all of the equipment and skill necessary to remove the chips and solder new ones, but I don't want to spend the time, money, and effort into trying to fix VRAM chips when it might be the VRAM controller internal to the GPU all along.

Is there a way to better narrow down if it's VRAM or GPU? Would it be worth simply reballing the VRAM chips first to see if it was just bad soldering from the factory or microfracture from heating/cooling over the course of a year?

At this point, the card is useless to me in it's current condition, so I'm willing to try things to get it working again.

Thanks again.

Also, attaching pics of the card.

[RJARRRPCGP]

VRAM being unstable, is more likely to cause a WHEA crash and auto reboot.

[grimacelord]

To anyone interested, I reflowed two of the VRAM chips that had some odd looking balls under it (when viewed at an angle under a microscope). Problem seems to be fixed. OCCT and GpuMemTest no longer report errors. FurMark runs without crashing and so far, I've been able to mine ETH for the past 8 hours straight without a single crash.

Will the problem return? Probably. At least that seems to confirm my suspicion that is is a VRAM issue. I have already sourced replacements, so if the problem returns, I'll buy new VRAMs and solder them on the board.

I'll call this one solved. ^_^

[RJARRRPCGP]

Originally Posted by grimacelord

To anyone interested, I reflowed two of the VRAM chips that had some odd looking balls under it (when viewed at an angle under a microscope). Problem seems to be fixed. OCCT and GpuMemTest no longer report errors. FurMark runs without crashing and so far, I've been able to mine ETH for the past 8 hours straight without a single crash.

Will the problem return? Probably. At least that seems to confirm my suspicion that is is a VRAM issue. I have already sourced replacements, so if the problem returns, I'll buy new VRAMs and solder them on the board.

I'll call this one solved. ^_^

On my Radeon RX 5600 XT, VRAM instability causes 0x00000124 (WHEA_UNCORRECTABLE_ERROR) crash and auto reboot with the given reason being "Cache Hierarchy Error".