When I got up yesterday my main cruncher was off. Pressed power button and nothing. Motherboard has power (reset light is lit). Turn off power on PSU (switch), turned it back on and now power switch worked and PC booted. Logged into Windows (10) and within 60 seconds powered off.

I can boot into CMOS and will run no problem (all last night 10 hours)

Yesterday after it ran in CMOS for a hour (updated BIOS to most recent), I was able to boot in windows safe mode, and it ran for more than an hour.

I then tried booting in normally and it did. A couple of hours later I came back and it was asleep, and I could not wake it. I unplugged the keyboard and plugged back in and no power to it.

I powered it off and could not get it to even boot to the login screen. It would get pat way and power off then on again, partial reboot, off then on. Did that a few times, then just powered off.

After that I booted back into CMOS and there it sat all night.

So, OS problem or Hardware? Windows 10 is installed on a Samsung 850 Pro SSD.

To me it's weird it will run no problem in CMOS, but as soon as I try to boot into Windows, it crashes.

So I am thing windows. Is it possible the SSD is failing? I thought SSD just failed.

Or could it be the PSU or MB. If it is those, why only when booting into Windows.

Specs"
Gigabyte Z170X-UD5
i7-6700K
32GB Corsair Vengence LPX 2400
Samsung Pro 512GB SSD
Corsair HX850i PSU
Corsair H80i cooler
MSI R9-380 GPU
Visiontek 7870 GPU
Corsair Obsidian 450D
WD Caviar Black 1x2TB, 1x4TB

Thanks,

Brett

---

Some time ago I had similar issues with an i7. It turned out to be the CPU was overheating due to the heatsink surface of the "lid" not being parallel with the mounting plane of the chip. A bit of careful application of force improved the situation, but eventually the processor died (and Intel did not honour the warranty as they considered it to have been my fault for their poor workmanship)

---

Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?

Some time ago I had similar issues with an i7. It turned out to be the CPU was overheating due to the heatsink surface of the "lid" not being parallel with the mounting plane of the chip. A bit of careful application of force improved the situation, but eventually the processor died (and Intel did not honour the warranty as they considered it to have been my fault for their poor workmanship)

OK, what I can try is in the CMOS, set an alarm on the CPU, and see if it goes off. What would be a good temp 80 of 90C or lower.

Brett

---

Interesting, I just updated the AMD driver last week. New I should not have done it "If it ain't broke, don't fix it". I usually resist, but caved last week.

Brett

---

Check that your PSU (brand) doesn't show on Johnny Lucky PSU Lemon List or Newegg Eggxpert Tiered Power Supply List (old 2008 list). Even what you think is a good brand may not always be the case.

I can boot into CMOS and will run no problem (all last night 10 hours)

You can boot into BIOS. not into CMOS. The Complementary metal–oxide–semiconductor is a static RAM that can hold changes you made to the BIOS, but you cannot boot into it.
Also, booting into the BIOS will never take up as much power as booting into Windows will, as the latter will put load on a lot more of your hardware. Running work on one or more GPUs will then even put more load on the system, where an underpowered or on-the-way-of-going-broke PSU will cut out unexpectedly. And if your system is then not set up to reboot on situations like this, you'll find your PC shut down.

Over the past weeks, my monitors would flicker, cut out completely. I have two monitors connected to a single videocard.
When during playing Wolfenstein my whole system just shut down and rebooted, I knew that my PSU was on its last legs. And I found mine on the bad list, a Huntkey 600W - that I still don't know how I acquired it - so I have it replaced now with a Seasonic 620W. It even feels as if the whole system is faster, snappier.

---

Anything over 80C is way too high for the long life and health of a CPU - I'd set the alarm on at about 75 so you had a safety margin.
(My daily driver, a "hot running" AMD 8-core FX runs at between 50 and 65C)

---

Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?

Updated BIOS and no change.

Booted into Safe mode, runs fine (low power) and uninstalled graphics driver. No change. Tried a different Graphics card = No change, different mouse/keyboard = No change. So down to likely PSU or MoBo.

On my last attempt, different GPU, when I powered it up, it took three attempts to power up. First two times, the power came on, fans spun up, then stopped and tried again, and on the third attempt it booted; like a relay was tripping. Like what happens when you forget to attach the 12v connector to the MoBo. Once booted, I was able to log into Windows. It ran fine for 15 minutes, so decided to reinstall the AMD Driver (older one from Oct). As soon as the screen went blank during the install, it shut down.

My original suspicion was that it was a power problem, and this is still what I am thinking. It will run OK in a low power state (BIOS, Safe-mode), but as soon as there is a larger demand on Power, it shuts down.

Anyway, while I could swap out the PSU from another box, it wasn't fully modular, and would require pulling the 12V connector, which is tucked underneath the H80iGT, which would have to be removed. So decided to pay $50 and have the local computer store diagnose it.

Brett

---

That sounds horribly like the death of either the motherboard or the CPU (or both)

---

Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?

Got a call from PC shop tonight. Looks like CPU cooler is pooched. CPU was overheating 95C. Having them put in a stock fan cooler $30. If it runs ok, then RMA the Corsair H80iGT. Interesting no alarm went off. After I re-flashed the BIOS, I did see no alarm was set. I don't know if it was always that way or a result of the re-flash.

Brett

---

Got a call from PC shop tonight. Looks like CPU cooler is pooched. CPU was overheating 95C. Having them put in a stock fan cooler $30. If it runs ok, then RMA the Corsair H80iGT. Interesting no alarm went off. After I re-flashed the BIOS, I did see no alarm was set. I don't know if it was always that way or a result of the re-flash.

Brett

I gave up water cooling, got to be a waste when I realized I could get almost the same cooling using air cooled copper CPU heatsinks. Only time I would use them is in compact Micro / mini atx motherboards where larger fans will not fit.

But this is a great fan cooler:

http://amzn.to/2khUKpv

Innovative, patented, and curved heat pipe design for a heat transfer capacity of up to six heat pipes with just three.
Low-noise 92mm PWM Control fan for automatic control of the fan's RPM according to CPU temperature.
Compatible with Intel Socket LGA 775/1156/1155/1150.

---

Well I can see one reason you don't like water cooling - No place to attach the rad :P
LMAO

For those with similar issues of computer blue screening or turning off soon after start up, in addition to checking whether CPU is overheating and trying a different PSU, is to test the RAM. If there are more than 1 sticks, leave in 1 stick at a time and see if symptoms go away. Or download memtest86 and run it for each stick. For me, if the computer is crashing within a minute of startup, RAM instability appeared within several minutes of testing, but some people let the test run for hours. Windows has a RAM testing software but it doesn't appear to be as sensitive as memtest for me.

For me, the most common causes were PSU or RAM. Heat sink dislocation and cable connections don't seem to pop up on me unless I did some cleaning or heavy moving. Haven't had a CPU ever fail on me. HDD sometimes, but usually the build is so old by then that I haven't been using it much. Not sure how to test a motherboard, but haven't had to before. Software issues like incompatible updates/drivers, corrupt files, or malware don't usually cause power-off or startup crashes for me , but are certainly worth checking.

Not sure how to test a motherboard, but haven't had to before.

Known good CPU, PSU, Memory, video card, keyboard, mouse, HDD/thumb drive with barebones OS. If the that system doesn't boot, but it will with another motherboard, then it's the motherboard.
It's generally much easier (and quicker) to test the other components (CPU, memory etc) separately, starting with the most likely and easiest ones first (keyboard, mouse (once had a faulty printer stopping a system from booting- removed USB cable & it booted & ran happily), PSU, memory) making your way to the least likely and more difficult/involved (CPU). If they are all OK, then it's the motherboard.

---

Grant
Darwin NT

Well, picked up the box. They replaced the Corsair H80iGT with an inexpensive Arctic Freezer 7 Pro, and its been running for about 7 hours now. The core temps are warmer than they were with the H80iGT. It's still under warranty, so will RMA and re-install the replacement.

Thanks for the help.

Brett

---