dDule
New Member
Hey, i've been dealing with PC crashes while gaming or stress testing without complete shutdowns and just blackscreens with no control and (sometimes) buzzing sound that force me to manually shut it down by holding the power button.
Close to 5 years, i'm running a CoolerMaster MWE 650W V2 80+ Gold Full modular PSU. It's built good, good cables and the fan is apsolutely silent. However, i have doubts now that it's working properly.
See, since i've bought the PSU, i have been running AMD GPUs. In series, R9 270X, R9 290, R9 290X, two RX 580s of two different brands and none had the modern sensors that would indicate any problems. Also, these crashes were happening since i've bought the R9 290 about 4 years ago( used it in the PSUs 1st year of service). I couldn't find any straight answers, so i wrapped the story up with the AMD drivers being super bad.
So i bought an R9 290X...which also crashed. Okay, bad drivers. Then, another R9 290...bad drivers. Then two RX 580s and...bad drivers. At this point i switched to team green with the RTX 3060Ti GamingX from MSi. And it started happening again, so i now highly doubt it's the GPU issue, but i'm here to ask anyone who can help in resolving this.
And, since the 3060Ti is fairly new, i got some cool sensors that help me in troubleshooting.
This is what i've found.
These voltages are coming straight from the GPU. Motherboard voltages seem very stable exactly around 3.3V, 5V and 12V. Maybe even too stable...i doubt the old motherboards sensors.
I've tested the 24-pin and GPU 6+8 pins and they all read out close to 12V without load. Putting a 20W lightbulb on a 8pin connector lowers it's reading to about 11.790V which is within spec. But while creating a high load on the GPU...big difference.
As far as i can see, the GPU even at the lowest usage% is dropping alot of voltage on all three sensors for power.
The situation is next.
Some of the older AMD GPUs did NOT crash for a while, then they start crashing for days on end, then they stop. The crashes are apsolutely random and without notice.
The RTX 3060Ti gave me about 2 months without crashes, and now it's extremely unstable.
The only test in OCCT i can pass is the 3D adaptive mode which loads the GPU in increments starting from 5% and going up to 100%, then restarting and going again. A couple of days ago, it could pass it for 1 hour, now it's unstable and may crash at the beginning, or around 30-40min mark. Sometimes it passes, which is the randomness of the crash.
Power stress test is crashing exactly 1 second after running, couple of ms up or down.
3D full stress test of 100% crashes instantly.
CPU stress tests are passing no issues.
RAM test passes no issues. (MemTest86 has passed with no errors, aswell)
RAM + CPU passes no issues.
GPU VRAM passes no issues.
Linpack passes no issues.
There is NO errors/warnings in Windows Event Viewer regarding the GPU.
Undervolting/Underclocking of the CPU or GPU doesn't solve the issue.
RAM memory has been swapped out, no changes.
Motherboard has been swapped out with two other motherboards, no changes.
GPUs(obviously) have been swapped out, no changes.
The only thing i haven't swapped out is the PSU.
Temperatures are stable. Not the best of cooling for the GPU, it underclocks itself a little bit but still far away from shutting itself down like this.
Any possible causes for this and is this low of a voltage even normal on the GPU rail? Can i finally rule out GPUs at the problem and point my finger at the PSU? Thanks inforward!
Close to 5 years, i'm running a CoolerMaster MWE 650W V2 80+ Gold Full modular PSU. It's built good, good cables and the fan is apsolutely silent. However, i have doubts now that it's working properly.
See, since i've bought the PSU, i have been running AMD GPUs. In series, R9 270X, R9 290, R9 290X, two RX 580s of two different brands and none had the modern sensors that would indicate any problems. Also, these crashes were happening since i've bought the R9 290 about 4 years ago( used it in the PSUs 1st year of service). I couldn't find any straight answers, so i wrapped the story up with the AMD drivers being super bad.
So i bought an R9 290X...which also crashed. Okay, bad drivers. Then, another R9 290...bad drivers. Then two RX 580s and...bad drivers. At this point i switched to team green with the RTX 3060Ti GamingX from MSi. And it started happening again, so i now highly doubt it's the GPU issue, but i'm here to ask anyone who can help in resolving this.
And, since the 3060Ti is fairly new, i got some cool sensors that help me in troubleshooting.
This is what i've found.
These voltages are coming straight from the GPU. Motherboard voltages seem very stable exactly around 3.3V, 5V and 12V. Maybe even too stable...i doubt the old motherboards sensors.
I've tested the 24-pin and GPU 6+8 pins and they all read out close to 12V without load. Putting a 20W lightbulb on a 8pin connector lowers it's reading to about 11.790V which is within spec. But while creating a high load on the GPU...big difference.
As far as i can see, the GPU even at the lowest usage% is dropping alot of voltage on all three sensors for power.
The situation is next.
Some of the older AMD GPUs did NOT crash for a while, then they start crashing for days on end, then they stop. The crashes are apsolutely random and without notice.
The RTX 3060Ti gave me about 2 months without crashes, and now it's extremely unstable.
The only test in OCCT i can pass is the 3D adaptive mode which loads the GPU in increments starting from 5% and going up to 100%, then restarting and going again. A couple of days ago, it could pass it for 1 hour, now it's unstable and may crash at the beginning, or around 30-40min mark. Sometimes it passes, which is the randomness of the crash.
Power stress test is crashing exactly 1 second after running, couple of ms up or down.
3D full stress test of 100% crashes instantly.
CPU stress tests are passing no issues.
RAM test passes no issues. (MemTest86 has passed with no errors, aswell)
RAM + CPU passes no issues.
GPU VRAM passes no issues.
Linpack passes no issues.
There is NO errors/warnings in Windows Event Viewer regarding the GPU.
Undervolting/Underclocking of the CPU or GPU doesn't solve the issue.
RAM memory has been swapped out, no changes.
Motherboard has been swapped out with two other motherboards, no changes.
GPUs(obviously) have been swapped out, no changes.
The only thing i haven't swapped out is the PSU.
Temperatures are stable. Not the best of cooling for the GPU, it underclocks itself a little bit but still far away from shutting itself down like this.
Any possible causes for this and is this low of a voltage even normal on the GPU rail? Can i finally rule out GPUs at the problem and point my finger at the PSU? Thanks inforward!
Last edited:
