Weird crashes are driving me nuts

[ioni]

I have a relatively recent system, 6700k/1070, but I keep getting these weird crashes where the entire system will freeze up but only for new actions. So if a video is playing, it continues to play, but I can't open any new windows or click on anything new. The mouse still moves around fine too.

I've rolled back my CPU and RAM OCs to stock, done disk checks, sfc, memtest, all with no errors. I have 2 SSDs and an HDD and even unplugged the HDD and still got the crash with it unplugged. I have 32GB RAM (8GBx4) and have tried 8x2 and the other 8x2 and both sets have had the crash. I'm not sure what else to check. I guess I could unplug the other storage SSD. But I don't expect that to fix it. I'm wondering if the CPU itself is faulty.

Also, at one point, after the crash happened, I'd always get a disk read error during boot. That's why I thought to unplug the HDD. But waiting a few minutes before doing the reboot seemed to always "fix" the error and allow me to boot. That seems to be a pretty broad error though, and not even necessarily related to an issue with an HDD/SSD.

Anyone else have any suggestions on how to diagnose this?

 

[UsandThem]

I have a relatively recent system, 6700k/1070, but I keep getting these weird crashes where the entire system will freeze up but only for new actions.

Also, at one point, after the crash happened, I'd always get a disk read error during boot. That's why I thought to unplug the HDD. But waiting a few minutes before doing the reboot seemed to always "fix" the error and allow me to boot. That seems to be a pretty broad error though, and not even necessarily related to an issue with an HDD/SSD.

Have you tried a utility like CrystalDiskInfo? Cause what you described sounds like it could be a disk issue. That utility will tell you if there are any problems/concerns about your disks.

Also, does it report problems under the reliability monitor?

http://m.winsupersite.com/windows-10/how-use-windows-reliability-monitor#slide-0-field_images-70831

 

[daveybrat]

Talk about timing.....I just went through 2 days of troubleshooting a customer's pc doing that exact same symptoms. Even the disk read error. The system didn't crash, i just couldn't open anything, like it was just stuck.

After a lot of trial and error, it turned out the be his power supply. Replaced it and all the symptoms went away and ran fine.

 

[Ketchup]

Could be the sata cable, could be the port you are using, could be a drive controller driver. What motherboard is this, what power supply, and what drive?

 

[ioni]

Mobo is a Gigabyte Z170MX-Gaming 5. PSU is an Antec HCG M Series HCG-620M. I forgot about this, but the PSU did fail back at the end of 2014 and I RMA'd it and got a replacement in Jan 2015. I guess it could be failing again?

Windows reliability monitor does show some HW faults, but I'm not sure if they line up with the freeze. I'll try crystal disk info. I had also heard about sata cables going bad, but I feel like that is the most unlikely possibility. SSDs are a 512GB 840 Pro and a 960GB SanDisk Ultra II.

 

[UsandThem]

Windows reliability monitor does show some HW faults, but I'm not sure if they line up with the freeze. I'll try crystal disk info. I had also heard about sata cables going bad, but I feel like that is the most unlikely possibility. SSDs are a 512GB 840 Pro and a 960GB SanDisk Ultra II.

Try running this as well to see if it can identify any issues from the Windows logs:

http://www.resplendence.com/whocrashed

 

[daveybrat]

Mobo is a Gigabyte Z170MX-Gaming 5. PSU is an Antec HCG M Series HCG-620M. I forgot about this, but the PSU did fail back at the end of 2014 and I RMA'd it and got a replacement in Jan 2015. I guess it could be failing again?

Windows reliability monitor does show some HW faults, but I'm not sure if they line up with the freeze. I'll try crystal disk info. I had also heard about sata cables going bad, but I feel like that is the most unlikely possibility. SSDs are a 512GB 840 Pro and a 960GB SanDisk Ultra II.

I'm telling you now, it's gotta be the power supply. Your symptoms are EXACTLY the same as the pc i was working on. New EVGA 650Watt Gold power supply cured it. I even replaced the sata cable on that system as well and no change.

 

[ioni]

Wells this is the PSU I'm using https://www.newegg.com/Product/Product.aspx?Item=N82E16817371059. The warranty does say 5 years, so I guess I'll make another claim against it. Just not sure how it will go since it does run a PC for a little bit at least.

 

[VirtualLarry]

I'm really not sure how a PSU issue would cause freezes. Crashes and reboots, blue-screens, yeah. But not pauses or freezes. That really does sound more like a disk-related issue. But a bad PSU could trigger a disk-related issue, if the voltages dropped too much, and the disk dropped out.

I'm currently testing a rig, using a Winsys WT-02 case+PSU, and a mini-DTX Biostar H81 mobo, with an overclocked G3258 in it. I don't know how well the PSU is holding up, it's supposed to be a 200 or 250 or maybe 300W unit. But twice now, the keyboard and/or mouse, made the "disconnect" sound, and the keyboard backlight went off, then it reconnected. Maybe Windows 10 is updating some drivers behind the scenes, and that's why? I don't know. I'm a little concerned that the 5V line on the PSU is sagging a little bit.

 

[MongGrel]

I'm really not sure how a PSU issue would cause freezes. Crashes and reboots, blue-screens, year. But not pauses or freezes. That really does sound more like a disk-related issue. But a bad PSU could trigger a disk-related issue, if the voltages dropped too much, and the disk dropped out.

I'm currently testing a rig, using a Winsys WT-02 case+PSU, and a mini-DTX Biostar H81 mobo, with an overclocked G3258 in it. I don't know how well the PSU is holding up, it's supposed to be a 200 or 250 or maybe 300W unit. But twice now, the keyboard and/or mouse, made the "disconnect" sound, and the keyboard backlight went off, then it reconnected. Maybe Windows 10 is updating some drivers behind the scenes, and that's why? I don't know. I'm a little concerned that the 5V line on the PSU is sagging a little bit.

A bad PSU can cause many problems, I was guilty of buying a few cheapos decades ago, and I learned better.

Is the bedrock of a system.

 

[denis280]

event log,or as mention above whocrached. I also lean towards Hdd or ssd failure.

 

[Grimner]

If the memory test has been run for a couple of hours we can leave the RAM alone for now. With heat issues from overclocking taken care of we are looking at the PSU, the motherboard or software.

My bets would be on the PSU or the motherboard.

Since “waiting a few minutes before doing the reboot” makes the disk error go away, I will look hard at that motherboard. See if you can hook it up to another PSU with one stick of RAM and booting from a Linux Live-CD with no hard drives. If it crashes repeatedly then, it is probably the motherboard. If not, it is most likely the PSU. You could also see how your PSU behaves in another system.

Software is its own can of worms. Met any strange programs recently? Is MBAM happy? Has quality control screwed up somewhere?

Two recent strange crashes in my own systems might set you on the creative path.
- Last autumn I had intermediate crashes in a system that had been stable for weeks if not months. Got the sinking feeling that something major was going down and decided to check the motherboard first.
Playing «Hunt the Leaking Capacitor» (a fun game for all ages, flash light required), I quickly declared the easily accessible parts around the CPU healthy and started looking between the cards.
Well, on top of the PSU with its legs reaching the sound card just above was a dead insect - some sort of crane fly, all legs and wing. Enough to cause some sort of electrical short on occasion? Maybe. Kicked it out of its rather magnificent tomb and the problem went away.
- Resurrected an Asus Maximus Extreme 775 (one of the DDR3 versions), in all its coppery glory and since this board runs a bit hot I put a spare 120mm fan on the bottom of the cabinet a little in front of the chip set. This worked well until I rerouted some cables and moved the fan closer to the chip set.
Linux clawed its way to a rescue start and did not seem happy at all. Ran memtest and after ten minutes the list of memory errors was a mile long on all four RAM modules. What just happened was my question. Well, disconnected the fan and ran memtest again. No problem and Linux booted as normal. It was an old Papst fan which seems a bit dated these days, but I guess the motor caused magnetic interference with the chip set. It is not normal to have a big fan and a chip set that close together. I imagine that if the fan had been placed a little differently, the problem would have been much smaller and more difficult to pin-point.

 

[UsandThem]

My bet is on newer Intel platform. Too much errata.

Don't worry, it will just die for no reason before long.


Nothing Haswell or later is reliable.

 

[ioni]

I think I found the solution and it's pretty funny. I say think, because the freeze has been pretty regular, at least once a day, and I haven't had it since Sunday night. While getting the info to RMA my PSU, I noticed it was pretty dusty. Dusted it out with some canned air and haven't had any problems since. So it seems like it was the PSU. IDK if dust was causing a short or what, but getting rid of it appears to have fixed the problem so far. It kind of makes sense that it's the PSU though. It's the only thing other than the drives I brought over from my old system, which would also rarely get the freeze.

 

[UsandThem]

I think I found the solution and it's pretty funny. I say think, because the freeze has been pretty regular, at least once a day, and I haven't had it since Sunday night. While getting the info to RMA my PSU, I noticed it was pretty dusty. Dusted it out with some canned air and haven't had any problems since. So it seems like it was the PSU. IDK if dust was causing a short or what, but getting rid of it appears to have fixed the problem so far. It kind of makes sense that it's the PSU though. It's the only thing other than the drives I brought over from my old system, which would also rarely get the freeze.

I think I am going to have to make that the first question I ask anybody who makes a "computer issue" thread from here on out. There have been several people where that turned out to be the reason. In fact, a lifer on here was having computer instability issues, and we were trying to figure what was causing them, and it turned out he hadn't cleaned out the dust in his PC for years. It basically was one big dust thermal blanket causing components to overheat. I never even thought to bring it up to a person who has been a member on a computer hardware enthusiast site for 17 years, with 15,000 posts.

At least it wasn't your post-Haswell CPU exploding on you.

 

[UsandThem]

They don't explode, they just fail to function. Haswell and Broadwell does too.

I have what, 5 cpus that need RMAing? That's not normal, at least it didn't used to be.

If you've RMA'ed 5 CPUs since Haswell was released, you're doing something wrong or there is something wrong with your power. Intel CPUs are built like tanks, and by far have the lowest failure rate of any PC component (although RAM is probably right up there as well).

 

[VirtualLarry]

Intel CPUs are built like tanks

Roger That. I've been building Haswell, Skylake, and now, Kaby Lake rigs, recently (last couple of years) and have been overclocking most of them, and haven't lost a one.

Btw, that poster is an RBM, probably banned for trolling a few times.

 

[MongGrel]

Redacted.

 

[MongGrel]

N00B Fake John Wayne has been trolling P&N threads today, looks he's moving into the tech areas.