Member Avatar for Mark_33

So last night I was running a program that utilized 100% of my CPU for a while and my desktop crashed. It turned itself off, and the power light (next to the power button) was blinking slowly. Pressing the power button did nothing. The light remained blinking until I cut the power, and when I returned power nothing at all happened - the blinking didn't return and the power button still did nothing. So to pinpoint the problem, I first checked the power supply. The power supply fan starts up when the green wire is shorted to ground. When I plug in only the 24-pin MOBO connector and press the power button, the fans start up and the power light comes on. However, when I plug in the 8-pin CPU power as well, the power button does nothing. Does this mean that I have a fried CPU? I wouldn't be surprised if it overheated from running full speed (It's an 8 core, 4 GHz processor so it can generate a lot of heat), but I'd expect failsafes to shut it down before anything melted in there.

  • 2 Contributors
  • 1 Reply
  • 308 Views
  • 2 Hours Discussion Span
  • Latest Post Latest Post by rubberman

Dani AI

Generated

Quick expert summary and checklist tied to the thread: ’s symptom set (board has standby power but refuses to start once the 8‑pin CPU/EPS cable is connected) most commonly means the CPU power rail is being seen as a fault when it’s energized — either the PSU’s protection is tripping or the motherboard’s CPU‑power circuitry (VRMs/caps/MOSFETs) is shorting under load. The 8‑pin/EPS connector is the dedicated 12V feed to the CPU/VRMs, so problems that only appear when that connector is applied point at the CPU power delivery path rather than general standby power. (allpinouts.org)

PSUs implement short/overcurrent/overpower protections that will prevent the system from starting if a serious fault appears as soon as the CPU rail is loaded; that behavior explains why the box “does nothing” when the EPS cable is plugged in rather than giving a gradual failure. That does not by itself prove the CPU is fried — a failed VRM, a shorted capacitor, or a PSU that breaks down under load will produce the same symptom. (help.corsair.com)

Practical next steps (safe, ordered, minimal parts):

  • Power down, unplug, and visually inspect the CPU area and VRM heatsinks for burn marks, swollen or leaking capacitors, or an acrid smell (these are decisive signs of board damage).
  • Try a bare‑bones bench: motherboard + PSU + CPU + single stick RAM (no drives, GPU if the board has video) and use the motherboard onboard power/jump pins to start.
  • Test or swap the PSU: use a dedicated PSU tester or a known‑good PSU; a paperclip test is not a full load check — a PSU can spin a fan yet still fail under CPU load.
  • If the fault persists with a known‑good PSU, remove the CPU and inspect socket/pins; try the CPU in a known‑good compatible board if available.
  • If there are visible damaged components around the socket or the board still trips immediately, the motherboard VRM or the CPU (less common) is likely failed and the board will need replacement or RMA. (capacitorlab.com)

One clarification for ’s point: OS‑level graceful shutdowns are separate from the hardware thermal trip. Modern processors include a built‑in catastrophic thermal shutdown (THERMTRIP) that halts the CPU at dangerous die temperatures to avoid immediate destruction — it’s a hardware feature, not something the OS must enable. That makes actual “melted” silicon unlikely, but collateral damage (VRMs, connectors, or the PSU) can still occur. (community.intel.com)

If unsure at any step, stop powering the system repeatedly (continued trials can worsen hardware shorts), and consider RMA or professional bench diagnosis after the PSU/motherboard swap tests.

Member Avatar for rubberman

Shutting down the system when the CPU overheats before damage is done is not usually a default setting of many systems, though most can be configured to do that. My Linux 8-core system is so configured. It will map out overheating memory, and gracefully shut down the system if the CPU is overheating, but I had to configure it to do that. I'm not 100% sure about Windows systems. I think your system is fried. If you are lucky, the disc drive and data is still intact.

Be a part of the DaniWeb community

We're a friendly, industry-focused community of developers, IT pros, digital marketers, and technology enthusiasts meeting, networking, learning, and sharing knowledge.