New system build/Windows/hardware/voltages and gremlins

[GeneralJock]

Hello Everyone,

This is my first post in this forum, and it's basically just a braindump, so feel free to ignore. I'm a moderately capable Joe who likes building PCs and mucking about with them.

However, I haven't been active in system building for quite a while. Apart from helping my son built his first proper desktop for his 15th birthday, I built myself a monster about 7 years ago and it's gone strong ever since with only one failure (Corsair power supply) so although expensive at the time, it's done me well. It was a monster at the time anyway: i7-3770K, 16GB RAM, and an ASUS GTX 980 all stuck onto an ASUS Z77 mainboard. All the games of the day on max setting.

A few years ago I started to have random phantom faults. PC wouldn't start, then sometimes it would, and occasionally it shut down on its own. I took a guess and dumped the Corsair 850 for a EVGA 1000W and never looked back. Rare for a Corsair to fail only after a couple of years, but there you go.

A month ago I started getting really weird stuff happening: Windows 7 hanging on boot, then booting okay, then booting but with no icons on the desktop (!!). Then booting okay, and then just randomly crashing. Then the keyboard wouldn't work at the log-in page, then it did. I am scrupulous on virus protection (BitDefender) and had no reason to suspect a virus. Finally, it crashed out big time until I could only boot from USB. On doing that, the SSD boot drive would read sometimes, then not, then fail and finally couldn't be read at all.

I've changed the SSD for a new 500GB Kingston, taking the opportunity to check and remake the cabling, hoover the whole system and clean it up. The system booted fine. My data, documents and games are all on a separate HDD so once I got a few installs sorted everything ran fine. Except...

Firstly, I cannot get rid of a USB driver fault in the "Other devices" section of Device Manager. All the USB controllers are working fine. I've updated everything from the ASUS website, unplugged all the USB devices, replugged them in, and it just won't go away. I've disabled it, and it's had no effect on anything. But also I am getting constant hangs (disabling USB device didn't affect that) usually for around 5 - 10 seconds, apps not responding. They always come back, but hang in "Not Responding" mode for that period. Caches all emptied, drives defragged, Windows diagnostics reports no faults. And as of a couple of days ago I've had high voltage warnings on both the 3.3V and the 5.0V buses. I go straight to the ASUS monitoring tool, and the buses report at 3.312V and 5.040V steady.

Anyway, I've decided I've had enough and there's probably a gremlin on the board somewhere. Going to upgrade to a newer board, faster RAM and a newer i7 processor, probably an 8700 or the like.

Rather than upgrade to an expensive new GPU, I'm going to pick up another same-box GTX 980 and get an SLI-ready board and couple the two up.

I'm also going to switch to Windows 10. Whilst I still prefer Windows 7 as an OS, it's just too much effort to apply all the updates and fixes now, so I've decided to bite the bullet so I'll bundle one with my new hardware.

Anyway... thought I'd share that. If you're inclined, I'd welcome any comments or thoughts on the driver issue, the voltages, the choice of new hardware, the plan to try SLI and Windows.

If you got this far, thanks for reading!!

All the best, Jock.

 

[Cycloid Torus]

Welcome, Jock.

1. Change to Windows 10 - first stage will be frustrating, so give it 3-4 months before you make a final judgement - not taking a day to fully update fresh install of 7 is a joy.
2. https://docs.microsoft.com/en-us/windows/deployment/update/waas-overview#servicing-channels - I opted for delaying 'feature updates' (as I do not enjoy surprises with my morning coffee)
3. if you have a backup image, it might be right to change to Windows 10 now (might even buy new boot drive - I use a 128gb ssd for boot and have a bigger drive on D: so I keep a local copy of C: on D
4. good time to make full backups or even consider a cloud backup service
5.GPU, I would wait for a sale on a newer card than SLI - consider the advantages of a 1660 unless you are going for a very big screen or feel that you need more speed [

]
6. I had ghost USB problems which I was able to cure by enabling 'view' of hidden devices in Device Manager and wiping out the dead - apparently, there is a graveyard and clearing it may get rid of ghosts
7. your most recent problems sound like motherboard Voltage Regulators which may have been strained by the failing Corsair PSU - but I would feel really stupid to learn that a fresh coin battery could be the need

Looking forward to hearing your progress, Pilgrim.

 

[GeneralJock]

Thanks for the comment! And thanks for that interesting link.
But... I can pick up a 2nd hand "good condition" GTX980 of the same type on ebay for less than £100.
I know it's a 5-year old card, and I need to do some research, but I reckon the performance boost per £ with those two in SLI could be quite compelling? At least for the next 12 - 18 months?
I'm gonna see if I can find some benchmarks...

 

[Cycloid Torus]

A pair would run 350 watts or so at full bore. Where I am that comes to more than USD 300 per annum - and I wonder if you might be better served by the newer cards (DX12, warranty, drivers, lower cost to operate) when the GTX 1670 and 1680 appear.

 

[GeneralJock]

Well, that's an interesting angle. The ASUS specs actually list the card as consuming 300W max, which I guess is why it needs the 8-pin connector. So two of them could eat even more than that. When I look at some of the mid-range 1660s, the price difference is not as high as I thought. I'll do a bit more research. Thanks for that.

 

[GeneralJock]

My gremlins just continue. I'm getting a CPU fan alarm now, checked the monitor and it's spinning at 547RPM with the low-speed limit set at 400RPM.
PC just repeatedly hangs now, I've been running Resource Monitor which shows repeated periods of complete data silence - look below and see one such gap where data transfer is shown as zero - each gap is about 6 seconds and for that period all applications were showing "Not Responding".

Despite defragging disks, eliminating all unnecessary startup processes, changing cache settings, I can't stop the disk hogging.

svchost.exe is the worst at a whopping 6MB/sec for prolongued periods

nvcontainer.exe is next at 2MB/sec. However, the nvidia problem seems to be well known, so I'm going to uninstall GeForce Experience.




Disk 1 is a pure data disk, so no idea why it gets hogged for the first 5 minutes after boot-up.

Disk 0 (C is my SSD system drive, it's getting hogged the most after the system calms down.



I have a bad feeling about my Seagate 1TB drive. It's been worked hard for 5 years now and I wonder if it's experiencing read/write faults (although Windows says not).

Given the price nowadays I think I will pre-empt a failure and replace it with a £65 2TB Barracuda.

Jock

 

[Cycloid Torus]

Having 'queue length' drop to zero is usually a good thing.

The real issue is "and for that period all applications were showing "Not Responding"."
Can you break out of the 'not responding' from Task Manager or is it a total freeze? Did you check Applications in Windows Logs in Event Viewer?

Regarding "I can't stop the disk hogging.", have you disabled 'fast boot'?

 

[GeneralJock]

If I hit Ctrl-Alt-Del it will break to the Task Manager but sometimes after a few seconds. It then shows some (but not always all) applications "Not Responding". I'm running the resource manager in the background continuously to see if I can spot a problem.
I've disabled superfetch, and I uninstalled GeForce experience, which seems to have helped. But svchost is still hogging the drive for a good few minutes after bootup.
I didn't think fastboot was a Windows 7 feature? I'll go into the BIOS and see if I can disable it there. Also going to look for the boot log as I can remember where it usually is.. stand by!

 

[GeneralJock]

In the event logs there is a constant recurrence of Error 0x80041003.
From checking this seems to be connected to user privileges.
I am wondering: when I replaced the SSD and did a clean Windows 7 install, could there be an issue with the apps and data on my HDD belonging to a "different" user? Even though I used the same Administrator login name, I have had occasional access issues which I've had to override using Admin privileges. Is this screwing up the system?
The only other error I have is with "WMI" and Windows is reporting:
"The service could not be started. The handle is invalid."
When I shut down last night, I got a dialogue box with a "Memory invalid write" error on it... that's another new one. Ran a memory check - Windows says okay.
I just feel the gremlins are now breeding and spreading to multiple parts of my system.
I'm not going to moan after 7 years of good performance from CPU, mainboard and memory, but I think I'll get the bits ordered up before Armageddon arrives.

 

[Cycloid Torus]

'Fast Boot' - yup, Windows 8. Now called 'Fast Startup". I disabled it for the few seconds it saves.

WMI error produces: https://answers.microsoft.com/en-us...ror-code/aee1e6c9-28d5-4871-908b-8ed42a36a96e It seems this is a particularity of Windows 7.

I'm suspicious.. Two more ideas: (1) download Microsoft Malicious Removal Tool, do a clean boot and run as ADMIN (as per https://answers.microsoft.com/en-us...eezes-my/46e73bd2-c1cd-411c-93ba-dee059653351 ) AND download and run Memtest86 overnight (>7 runs - available here: https://www.techspot.com/downloads/6216-memtest86.html )

Just trying to eliminate possibility of a rogue script and/or intermittent RAM.

I try to not believe in gremlins.

 

[GeneralJock]

Done the SMT run... no files infected. May try the Memtest86 tonight, but I need my bed alas. Thanks for the tips.

 

[GeneralJock]

One thing I didn't mention, and I think could be critical actually, is that the hard drive has started clicking. I've noticed now that when the system is hanging, the HDD drive is on steady (no R/W flickering) and then there's a click and the system resumes.
I'm convinced there's either a problem with the disk controller on the board, or the firmware on the drive is playing up.
It's a 1TB Barracuda DM003 and the whole series are known for firmware issues leading to platter degradation.
Web says time from the "click of death" to complete failure can range from days to months, but is inevitable.
(By the way I've made another post on Barracuda HDDs in this forum)