F5F Stay Refreshed Hardware Desktop Instability detected in system with Ryzen 2600 and Sapphire Nitro+ 5700XT mining card.

Instability detected in system with Ryzen 2600 and Sapphire Nitro+ 5700XT mining card.

Instability detected in system with Ryzen 2600 and Sapphire Nitro+ 5700XT mining card.

C
CobbleWalker
Member
238
08-31-2021, 02:19 PM
#1
Hey, we've been struggling our way through this one for a few months now. I'm gonna share this here, if anyone has any experience, or tech tips, I'd be excited to read all about it. ASRock B450 K4 Gaming Ryzen 5 2600 or 3600 HyperX 16GB kit 21xxMHz or Vengeance 16GB kit 3600MHz running 3200MHz for stability Sapphire Nitro+ RX 5700XT (was probably a mining card) 2 VBIOSes - running the 180W lower power one, the 210W is borked. Recently re-pasted and changed out the pads. 650W PSU - not sure about the brand. (ordered a 850W Seasonic replacement recently, not installed yet) 2TB Samsung 990 PRO or 256GB M.2 SSD Windows 11 Tried flashing the cards 210W vBIOS using amdvbflash 2.93 UEFI version, but the card test result is FAIL on "-i" The card does pass in 5.0.5xx but when flashed, the BIOS does not "stick" I am still looking for ways to flash a normal 210W VBIOS, but it seems like using a programmer is the only option left The system was running fine but now it's unstable. The biggest correlation I've gotten so far is the 12V rail falling to 11.05V under load. (usually 11.4, but 11.05 is the lowest recorded in hwinfo) When playing around with this, I limited the CPU to base clock and the card to -50% power. Ran CPU burner and FurMark and then unrestricted the GPU under load to 100% power (180W nominal) When I did this, the GPU hotspot temperature rose rapidly to 107 ˚C the 12V rail fell to ~11.2V I then restored the limit to 50% power but the system did not recover, hotspot stayed at 105 ˚C The system stabilized only when I removed the load. If the system crashes, it's either CRITICAL_PROCESS_END or when quitting CS2, the game refuses to quit, windows processes start crashing (old windw borders get shown, then blank screen with mouse and finally nothing, locked up), or it freezes for like 10s and goes to no display out straight away This was not happening before, even after the re-paste. Has anyone ever seen this? Any stories of how you got out of this? This is happening to my little brother's computer. I just got him into the computer building and maintaining thing like a year ago, so he can save some money and unfortunately he's gotten to experience the worst of it. I live a half an hour drive away, so we're iterating through the issues slower than we'd like. Thank you for any help or stories similar to mine. I also asked for guideance on VBIOS flashing on TechPowerUp forums. Link: https://www.techpowerup.com/forums/threa...st.332963/
C
CobbleWalker
08-31-2021, 02:19 PM #1

Hey, we've been struggling our way through this one for a few months now. I'm gonna share this here, if anyone has any experience, or tech tips, I'd be excited to read all about it. ASRock B450 K4 Gaming Ryzen 5 2600 or 3600 HyperX 16GB kit 21xxMHz or Vengeance 16GB kit 3600MHz running 3200MHz for stability Sapphire Nitro+ RX 5700XT (was probably a mining card) 2 VBIOSes - running the 180W lower power one, the 210W is borked. Recently re-pasted and changed out the pads. 650W PSU - not sure about the brand. (ordered a 850W Seasonic replacement recently, not installed yet) 2TB Samsung 990 PRO or 256GB M.2 SSD Windows 11 Tried flashing the cards 210W vBIOS using amdvbflash 2.93 UEFI version, but the card test result is FAIL on "-i" The card does pass in 5.0.5xx but when flashed, the BIOS does not "stick" I am still looking for ways to flash a normal 210W VBIOS, but it seems like using a programmer is the only option left The system was running fine but now it's unstable. The biggest correlation I've gotten so far is the 12V rail falling to 11.05V under load. (usually 11.4, but 11.05 is the lowest recorded in hwinfo) When playing around with this, I limited the CPU to base clock and the card to -50% power. Ran CPU burner and FurMark and then unrestricted the GPU under load to 100% power (180W nominal) When I did this, the GPU hotspot temperature rose rapidly to 107 ˚C the 12V rail fell to ~11.2V I then restored the limit to 50% power but the system did not recover, hotspot stayed at 105 ˚C The system stabilized only when I removed the load. If the system crashes, it's either CRITICAL_PROCESS_END or when quitting CS2, the game refuses to quit, windows processes start crashing (old windw borders get shown, then blank screen with mouse and finally nothing, locked up), or it freezes for like 10s and goes to no display out straight away This was not happening before, even after the re-paste. Has anyone ever seen this? Any stories of how you got out of this? This is happening to my little brother's computer. I just got him into the computer building and maintaining thing like a year ago, so he can save some money and unfortunately he's gotten to experience the worst of it. I live a half an hour drive away, so we're iterating through the issues slower than we'd like. Thank you for any help or stories similar to mine. I also asked for guideance on VBIOS flashing on TechPowerUp forums. Link: https://www.techpowerup.com/forums/threa...st.332963/

G
gadi7
Junior Member
45
09-08-2021, 05:58 AM
#2
Initially, the graphics card seems to function well under the 18W BIOS settings. Unless your setup includes a 4K display, it should operate smoothly without noticeable impact. The fluctuating 12V rail you mentioned might be typical for certain power supplies, but since you’ve already ordered a replacement, you can verify if this was a problem once installed. Regarding GPU temperatures, how did you obtain that data? Some applications tend to give inaccurate readings, and sensors can drift over time. I suggest using HWinfo or GPU-z for more reliable info. You might want to physically check the GPU with your hand (carefully!) near it and feel its temperature. Also, observe if the fans are spinning—raising them to full speed could help identify the issue. The increased noise is just a nuisance; once testing is done, you can revert to automatic control. Good luck!
G
gadi7
09-08-2021, 05:58 AM #2

Initially, the graphics card seems to function well under the 18W BIOS settings. Unless your setup includes a 4K display, it should operate smoothly without noticeable impact. The fluctuating 12V rail you mentioned might be typical for certain power supplies, but since you’ve already ordered a replacement, you can verify if this was a problem once installed. Regarding GPU temperatures, how did you obtain that data? Some applications tend to give inaccurate readings, and sensors can drift over time. I suggest using HWinfo or GPU-z for more reliable info. You might want to physically check the GPU with your hand (carefully!) near it and feel its temperature. Also, observe if the fans are spinning—raising them to full speed could help identify the issue. The increased noise is just a nuisance; once testing is done, you can revert to automatic control. Good luck!

L
LusterBlaze
Junior Member
19
09-08-2021, 06:04 AM
#3
The voltage drop didn’t match the ATX specifications at the time these parts arrived, so we decided to replace them first. We use HWINFO, and it seems the previous owner might have faced similar problems with it. The BIOS was adjusted for maximum fan speed, which kept the current draw very low. It’s possible mining damage occurred (VRMS, die issues) though—uncertain. Overall, the new PSU is now installed. My brother was excited to finish, and the strange behavior has cleared up. GPU hotspot temperatures returned to normal, and the system is stable without 12V line drops to 11V, which seems to have helped. He’s using the lower 180W VBIOS while the 210W model is the modified one. Restoring to original BIOS would be ideal, as it could boost performance slightly. The GPU appears to be functioning well now. Still curious about the VBIOS situation—will it affect future stability? Or did we just acquire it at a bad time? We tested our 3070 on the same setup and it worked fine. Even though the 650W Zalman PSU was rated for 210W, it didn’t fail with the 3070, suggesting better power handling. What are your thoughts?
L
LusterBlaze
09-08-2021, 06:04 AM #3

The voltage drop didn’t match the ATX specifications at the time these parts arrived, so we decided to replace them first. We use HWINFO, and it seems the previous owner might have faced similar problems with it. The BIOS was adjusted for maximum fan speed, which kept the current draw very low. It’s possible mining damage occurred (VRMS, die issues) though—uncertain. Overall, the new PSU is now installed. My brother was excited to finish, and the strange behavior has cleared up. GPU hotspot temperatures returned to normal, and the system is stable without 12V line drops to 11V, which seems to have helped. He’s using the lower 180W VBIOS while the 210W model is the modified one. Restoring to original BIOS would be ideal, as it could boost performance slightly. The GPU appears to be functioning well now. Still curious about the VBIOS situation—will it affect future stability? Or did we just acquire it at a bad time? We tested our 3070 on the same setup and it worked fine. Even though the 650W Zalman PSU was rated for 210W, it didn’t fail with the 3070, suggesting better power handling. What are your thoughts?