Unstable 7900XTX Hellbound
Unstable 7900XTX Hellbound
I've removed this component for a basic repaste and changed the thermal pad, but it's been behaving oddly. I've used 1.5mm pads from Thermal Grizzly all around, along with VRM and VRAM. (I tried 1mm but it wasn't thick enough for real contact.) I applied this on all my cards, and they worked fine until the 7900XTX model. First, I experienced a major issue: when it switched into 3D mode, the card crashed instantly to a black screen. I double-checked for any mistakes (PCIe, power connections, etc.) but found nothing. On my test bench it was the same. Panicking, I removed the VRAM thermal pads—it fixed the problem. Then I replaced the 1.0mm pads on VRM, and it worked again. However, the original 1.5mm pads didn't connect properly, so I replaced them. After that, VRM pads functioned correctly. The concern was whether this was the cause. I've been using it for several days without issues, aiming to lower VRAM temperatures and undervolt the chipset. The card looked good and operated normally for days. I started undervolting, but initially it crashed apps or games, which was expected until I found a better balance between heat, noise, frequency, and core voltage. The issue persists, even at stock settings, and it now only handles short bursts on desktop or during intensive tasks. I attempted to revert to an older driver, changed the BIOS settings, and switched the VRM thermal pads, but the problem continued. I tried another graphics card with similar specs and the same issues. I also reset the VBIOS and changed the cooler, yet it still crashes frequently. The card appears to be a weak undervolter or something went wrong. Its performance dropped significantly, and it now only lasts a few seconds under load. It works okay for light tasks or boosting apps, but that's about it. I'm worried about potential damage if I made a mistake.
They are under control, and even I only reduce the power limit, the card still does the same thing. There is a big delta between die and hot spot, but the card was already like that when I got it. With -5% power limit, the die sits at ~60°c while hotspot stabilize at ~87°c. I don't think heat is the issue, the card doesn't even have time to heat up. Plus the system doesn't shutdown, it still runs after a blackscreen.
The 27c delta shows significant bad contact on the die. It seems likely that when you insert the card, the hotspot activates and triggers the crash. This appears to be my conclusion. Typically, the delta should be closer—around 10-15c—when the die makes proper contact.
I've received some updates. The positive news is the card appears to function again. Previously, it couldn't hold a load for more than a few seconds, but now it completed all 20 rounds of the Time Spy Extreme Stress Test without any problems. I even experimented by adjusting the window position, which previously caused it to crash more quickly. The negative aspect is that the VRM no longer makes contact with the cooler, as I removed the thermal pads. Only the VRAM had that access. I paid extra attention to reassemble the card correctly to maintain good GPU contact. The performance difference remains the same. It seems either the extra care or the VRM thermal pad issue is responsible, or it might be unrelated. I don’t have any idea yet. I’ve already contacted Powercolor for specs on the spring screws used in the backplate and behind the core, and they provided them without issues. I intend to replace those screws since the initial disassembly caused problems and made it hard to grip the screwdriver. I also plan to request the exact thickness of the stock thermal pad—possibly something like 1.20, 1.25 or 1.30. If anyone has suggestions, please share.
It's becoming increasingly strange. I'm not sure how to proceed. It looks like it might be functioning on the test rig, or perhaps I didn't run enough tests. I completely reset the windows on the main rig, and it seems to work under load. I also considered a PSU problem, but it just showed a black screen when idle... So maybe not. Could the PCI-e riser be the issue? Is this a software problem? But everything started when I removed the card, so... I really don't know what happened.
I’m having trouble with this setup right now. I’ll need to replace the PCIe riser entirely. Otherwise, I’ll have to take the whole PC apart... At this stage, I’m probably going to try it, but the timing would be unusual. Do you have any suggestions for a good PCIe riser for 4.0?
Update 4 : Might of find a suspect, could be the pcie riser indeed, or a least somewhat related. On the main pc, the situation degraded rapidly, there's just no display at all now. I thought the card was done. But I put the card and the test rig, and it seems to run fine. (For now anyway) I also dismounted the pcie riser, and foud out that some screws on the back of the motherboard were a bit losed, and pressed against the pcie riser, it did a bit of damage. Now, is that damage/contact related to anything, I still don't know. I connected that same riser on the test rig, and it's still fine. (Never wished so bad to have a crash ffs, so I can pin point what's the problem, and it doesn't! It's so frustrating...)