The system experienced a crash when the voltage core reached high values. Please assist!
The system experienced a crash when the voltage core reached high values. Please assist!
Hello everyone,
MB: Asrock Z77 WS
Cooler: Noctua NH D15
The issue is that the CPU crashes when running at 4.0ghz even at 1.32v, and at 4.3ghz it crashes at 1.45v. I've tried many different settings but the results stayed the same. The PC shuts down at very high CPU voltage.
On another board (Asrock z77 extreme 4), I managed to run 4.3ghz at 1.25 without problems after long stress tests. I used fixed voltage and turned off C-states, that was the key.
Temperatures stayed below 70°C under full load.
I also re-flashed the BIOS.
I also attempted to leave everything Auto and only increased the multiplier to 4.2ghz; the PC didn't crash but the vcore reached nearly 1.5v.
First, I wouldn't do that. Set the LLC to just below middle. So for you on that board, probably 6. High LLC is only for EXTREME overclocking configurations, and your configuration does not constitute an extreme configuration.
5 or 6 at most. Might even start with 7 and see how much vdroop there is if possible/stable at some point.
Too much voltage can cause crashing and instability just the same as too little.
Those power supplies were very good, well, pretty good, when new, but that was like 8 years ago. That means it was likely sitting in inventory for two years before you bought it, and now six years later we can assume it is at least six years and more like 8 years old. Probably group regulated, almost certainly. Probably has...
Have you verified the temperatures on the motherboard? The VRMs might be overheating or they may no longer supply the required voltage. If not, consider resetting the motherboard to its default configuration to resolve any stuck settings that could block an overclock.
I've come across many discussions about difficulties in overclocking this board, but considering its 8+4 power phase layout, it should perform adequately.
Have you purchased this board used before?
The first step I'd suggest is removing the CPU and confirming there are no bent pins in the motherboard socket. These can be hard to spot without magnification, especially if they're not obvious.
Another aspect to check is whether the CPU cooler is properly seated. If it's too tight in one area, it might misalign the CPU in the socket, leading to various issues. It should be adjusted gradually in a crosshatch pattern, ensuring it's perfectly level and even.
You're likely familiar with what you're doing, but this can still be a common mistake even for experienced builders.
A visual inspection of the motherboard is also important. Look for any bulging components, leaking caps, scratches, or burnt traces that could affect performance.
Have you run a stress test using Prime95 version 26.6 Small FFT or Realbench's stress option to determine if the issue lies specifically with overclocking or if there are deeper problems? Sometimes challenges become clearer when overclocking, but they can exist even at the stock settings where they're less noticeable.
So, I wasn't trying to point out that the cooler was an issue because of thermal compliance. There are many other ways an incorrect cooler setup can impact performance when the CPU isn't properly seated in the socket due to uneven mounting pressure. It's always important to check this because I've seen cases ranging from drives failing to work correctly, to screen red dots, to instability or even crashes.
Regarding the voltage comment, where exactly are you noticing that?
And how do you confirm stability isn't the problem? Have you verified it in the standard setup? Just because a system doesn’t bluescreen or restart doesn’t rule out stability or power delivery problems.
You also didn’t address whether this was a used board you bought, or if it was new stock.
What prompted you to replace the motherboards originally?
I was tracking voltages through bios, hwmonitor and cpuz, but the variation wasn't significant. I performed a p95 v26.6 small fft analysis over about 9-10 hours; everything appeared normal. At that time, the CPU clocked at 4.0ghz with a 1.35vcore—anything below that risked a blue screen. I purchased this motherboard and i5 2500k several years ago as an upgrade from the 775 model. Back then I also experienced high vcore problems but ignored them, assuming it was just a faulty chip. Recently, a friend gifted me his i7, which ran at 4.3ghz with a 1.25vcore. My system doesn’t even boot at that vcore when using 4.0ghz.
It seems you're questioning whether the i7 was actually running on Extreme4 for a fair comparison. There might be an issue with the Extreme 4 itself, or perhaps it's still functional. Adding four hyperthreads usually affects stability at the same voltage, which is expected. If the CPU struggles to maintain stability across voltages, it could point to a hardware problem, possibly explaining why you received it. Your doubts are valid.
My friend had that chip paired with extreme 4, and I wanted to keep the same OC settings and maintain the same speed at the same Vcore. I mentioned that 4.3ghz was running at 1.25v, and 2500k has nothing to do here. NO! just forget about that he gifted me because it was a faulty CPU. Tomorrow I'm going to try this CPU on the P8z77 board and check the Vcore and GHz, but I'm sure the issue is with my Z77 WS board, not the CPU itself. Just ran the diagnostic tool and everything looks fine.
I completely understand that the issue might be linked to the motherboard. It functioned properly on another board. It passed the CPU diagnostic test. Another chance is that because a different power supply was used, there could be problems with power fluctuations or delivery from your own supply. Could you provide the exact model of your power supply and how long it has been operating? This isn't necessarily about your specific case, as you've already tested various setups. Stability differences will always exist between boards or even between CPUs, even if the motherboard models match. Every board and every CPU has its own traits, so what works on one might not perform identically on another. It's possible, but not guaranteed.