The computer stopped working suddenly
The computer stopped working suddenly
My system:
- Core i7 [strike]7770K[/strike] 3770K (sorry, it was a typo)
- Cooler Master 212 Hyper X heatsink
- GTX 970
- ASUS P8P67M-PRO motherboard
- Power supply: Corsair Builder’s series CX 500 V2
- Operating system: Windows on Samsung 850 EVO SSD; includes two HDDs
- Backup power unit (prevents shutdown during power cuts)
WHAT IS OCCURRING (hereafter referred to as “the issue”):
Yesterday evening I was surfing the internet when my computer abruptly shut down. Since then, it hasn’t stayed on for more than 1 hour and 6 minutes (the current record). The DRAM LED is now always lit. I brought it in the next morning to a reliable technician, who spent two days working on it before I returned to conduct additional tests over the weekend.
The issue does
NOT
appear to be linked to:
- Graphics card.
I replaced my GTX 970 with an older GTX 560 that was sitting unused, and the problem persists.
- Power Supply Unit.
It’s quite old (5 years) and a subpar model from the start (rated Tier 5 in the PSU category), so it was my initial suspicion. However, my technician swapped it for his Corsair 700 W unit and the issue continued.
RAM.
I performed an 8-hour Memtest with no errors; further details on RAM are provided below. Since this doesn’t reflect real-world usage, I’m not counting 8 hours as a valid duration.
UPS unit.
Connecting the PC directly to the wall power source still causes the problem.
Overheating.
Temperatures remain normal. Even when using the “maximum heat” Prime95 profile, my processor stays below 75°C. I also haven’t noticed any abrupt spikes in fan noise before shutdown. GPU temperatures are stable even after a full Uningine Heaven benchmark.
Component stress.
[strike] In fact, it seems component stress might actually be safeguarding the system. The three instances where it lasted longest (40 minutes to over an hour) coincided with intense gaming sessions. Right now, Prime95’s “blend” test is running for nearly 2 hours, breaking the previous record. On the other hand, when I leave the machine idle, it powers off in 5–15 minutes. We don’t have enough data yet to confirm that stress is the cause, but we can be certain it isn’t due to overloading the GPU or CPU.
Update:
It has endured for more than 3 hours under Prime95, but shuts down quickly even when Prime95 is active. This suggests stress might not be the root cause, although it doesn’t seem to trigger the problem either.
THE CONCLUSION:
- The motherboard and processor were purchased used from the same individual. Although labeled a K model, I never attempted overclocking without instability, suggesting possible issues with the chipset or seller. Prior to this problem, I experienced occasional complete system freezes (2–3 times monthly). Apart from that and the overclocking concern, the system has performed reliably.
- The machine can power down at any moment—whether after just two seconds on startup or during an hour of VR gaming. Since it shuts down before the BIOS, the issue isn’t likely related to Windows or software.
- Upgrading the PSU didn’t resolve the DRAM LED problem.
- The technician noted that the RAM clips securing the modules are loose; the computer stopped responding to video after being moved to his shop, which he resolved by manually pressing on the RAM. If faulty RAM was the culprit, it likely wouldn’t have survived the 8-hour Memtest. Additionally, if the failure stemmed from the RAM shifting itself, the system would remain unresponsive after shutdown (as RAM wouldn’t realign).
- He claimed to have once observed the processor reaching around 90°C moments before shutdown. My temperature monitoring (Core Temp in logging mode) has not recorded any irregularities in recent checks, suggesting he might have misinterpreted the readings.
MY HYPOTHESIS
is that either the processor—or the motherboard, given the DRAM LED warning—is failing. I’m still trying to figure out which one, so I can decide what replacement to order. Any additional suggestions for diagnosing this matter would be greatly appreciated.]
Nantes :
My setup:
- Core i7 7770K
- Cooler Master 212 Hyper X heatsink
- GTX 970
- ASUS P8P67M-PRO motherboard
- Power supply: Corsair Builder’s series CX 500 V2
- Operating system: Windows on Samsung 850 EVO SSD; includes two HDDs
- Backup power unit (prevents shutdown during power cuts)
WHAT IS OCCURRING (hereafter referred to as “the issue”):
Yesterday evening I was surfing the internet when my computer abruptly shut down. Since then, it hasn’t stayed on for more than 1 hour and 6 minutes (the current record). The DRAM LED is now always lit. I brought it in the next morning to a reliable technician, who spent two days working on it before I returned to conduct additional tests over the weekend.
The problem does
NOT
appear to be linked to:
- Graphics card.
I replaced my GTX 970 with an older GTX 560 that was sitting around, and the issue persists.
Power supply unit.
It’s quite old (5 years) and a rather poor model from the start (rated Tier 5 in the PSU list), so it was my initial guess. However, my technician swapped it for his Corsair 700 W unit and the problem continued.
RAM.
I ran an 8-hour Memtest with no errors; further details on RAM are below.
Uninterruptible Power Supply (UPS).
I connected the machine directly to the wall outlet, and the issue remains.
Overheating.
Temperatures are normal. Even when using the “maximum heat” Prime95 benchmark, my processor stays under 75°C. I haven’t noticed any sudden spikes in fan noise before shutdown. GPU temperatures are also stable after a full Uningine Heaven test.
Component stress.
Interestingly, component stress seems to act as a safeguard. The three instances where it lasted the longest (40 minutes to over an hour) occurred during intense gaming sessions. Right now, Prime95’s “blend” test is running strong at 1 hour and 20 minutes, breaking the previous record. On the other hand, when I leave the PC idle, it powers off in 5–15 minutes. We don’t have enough data yet to confirm that stress is the cause, but we can be sure it’s not due to overloading the GPU or CPU.
THE FACTS:
- The motherboard and processor were purchased used from the same individual. Although it’s a K model, I never managed to overclock the CPU without instability, suggesting possible issues with the hardware. Prior to this problem, I experienced occasional complete freezes (2–3 times per month). Apart from that and the overclocking concerns, the system has performed reliably.
- The machine can shut down at any moment—from just two seconds after powering on to an hour of VR gaming. Since it can cut off before the BIOS, the problem isn’t likely related to Windows or software.
- Replacing the PSU didn’t resolve the DRAM LED issue.
- The technician noted that the RAM clips securing the modules are loose; the PC stopped responding to video after being moved to his office. He fixed it by manually pressing on the RAM. However, if faulty RAM caused the shutdown, it likely wouldn’t have survived the 8-hour Memtest. If the problem stemmed from the RAM shifting itself, the computer would stop powering up altogether (as the modules wouldn’t realign).
- The technician claimed he once observed the processor reaching around 90°C moments before shutdown. I’ve been monitoring Core Temp in logging mode (recording temperature and load every 10 seconds), and so far, no irregularities have appeared before each shutdown. This might indicate he misinterpreted the reading.
MY THEORY
is that either the processor—or the motherboard—is failing, possibly hinted by the DRAM LED—so I need to know which one to replace. Any additional advice on diagnosing this would be greatly appreciated.
I understand this was a lengthy explanation, so thank you for getting through it.
Are you certain about the specifications? The P8P67M-PRO is compatible only with 2nd or 3rd generation CPUs; are you sure it’s not a 3770k model?
Would it help to turn off the Haswell and subsequent C-States in your BIOS?
Biglizard:
ASUS P8P67M-PRO is socket 1155
CPU is socket 1151
Surprised it still works.
Consider buying a z170 - z270 board.
As discussed before, I’m certain about the CPU.
Holy hell, I never realized the motherboard had a completely different socket! The seller told me the system was fine before, so I assumed it wasn’t an issue. But yes, these are correct:
http://prntscr.com/fcwzp0
I also checked my case and confirmed the imprint on the motherboard—it matches this model.
You’re really impressed! But if the socket is the problem, why would it only start showing up now after a few years of use?
Obs: I tested Prime95 in mixed mode for three hours straight before shutting it down manually. The idea that stress protects is becoming more popular...
Nantes :
biglizard :
ASUS P8P67M-PRO is socket 1155
CPU is socket 1151
Wow, it actually works.
Consider buying a z170 - z270 board.
As discussed before, I’m certain about the CPU.
Holy hell, I never realized the motherboard slot was completely different! But yes, I’m pretty sure.
http://prntscr.com/fcwzp0
I also checked my case and examined the imprint on the motherboard. It’s definitely that model.
You’re really impressed! But if the socket was the issue, why would it only start working now after a few years of use?
Obs: It’s interesting I ran Prime95 in mixed mode for 3 hours straight before shutting it down manually. The idea that stress protects is becoming more popular...
Ah, the problem here is you mentioned *7700K* at the top, not *3770K*.
Power problems while idle remind me of C-State concerns. Should I disable the Haswell and later C-States in your BIOS? It's not only during idle, as it shuts down promptly during gaming or when loading Windows, or just two seconds after powering on—though Prime95 has kept it from turning off so far.