Jump to content

Welcome to MSFN Forum
Register now to gain access to all of our features. Once registered and logged in, you will be able to create topics, post replies to existing threads, give reputation to your fellow members, get your own private messenger, post status updates, manage your profile and so much more. This message will be removed once you have signed in.
Login to Account Create an Account


Photo

Computer woes, please help diagnose!

- - - - -

  • Please log in to reply
14 replies to this topic

#1
soggyamphibian

soggyamphibian
  • Members
  • 5 posts
  • OS:Vista Ultimate x64
  • Country: Country Flag
Hi!
I've had problems with my computer crashing.
At this point I suspect there is a hardware problem (and I'd guess motherboard), but thought I'd better leave the judgement to people better experienced than I.

Here's my system specs:
Motherboard: ECS A885GM-A2
Processor: AMD Phenom II X4 970 Processor 3.50GB
RAM: 2x4GB sticks, Crucial.
Video card: EVGA GeForce 210 1024MB DDR3, but reverted to the motherboard graphics.
Hard Drives: WD10EARX Sata 1TB (OS drive), Samsung HD154UI 15TB (Storage), there's another one too, but I unplugged it.
Optical drives: was running an LG Super Multi SATA internal thing, unplugged it and have been trying an external USB version of the same.
OS: Vista Ultimate 64-bit, also experimenting with Linux Mint on a USB stick.

Had intermittent problems with bluescreening, freezing and restarting for a while, bluescreening got really bad so I reinstalled Windows.
Some of the freezing gave some screen glitches. I had WhoCrashed installed, and not every bluescreen seemed to have been remembered by the computer. I saved the old crash reports. I'm sure I ran a Windows Fixit at some point in that previous install, I don't remember what it was for.
After reinstalling, it still crashed. I removed the graphics card (I mentioned the screen glitching), the internal optical drive and the third hard drive (the latter two were second-hand, so I thought I'd rule those out). Seemed more stable, but still crashed. Did get a Windows Fixit for "common USB problems" after a bluescreen named a USB thing, has still been crashing since. A later crash had the Windows bug report thing thinking there might be malware, a Full Microsoft Security Essentials scan didn't come up with anything.
(I've been saving the minidumps, in case they're needed).

Some extra oddities:
Sometimes the screen glitches before restarting, crashing or freezing. Usually it's a little glitch (lines of intermittent yellow over whatever else is going on), sometimes this precedes the whole screen going messed-up.
A handful of times in its crashing or freezing before getting into Windows proper, there have been a bunch of blue blocks over the screen (only noticed this since removing the graphics card).
Three times so far, it's basically reset the Bios ("CMOS checksum bad, CMOS time/date not set" or some such). The battery seems to be fine - Bios holds information the rest of the time. One time I pressed F1 to get into the BIOS, screen went black with a couple blue blocks in the top-left corner, with the words "Please Wait" in white in the middle of the screen. I waited, after I guess 30 seconds the computer reset.
It's frozen once in Linux Mint (I don't use Linux a great deal).
It's frozen once in BIOS (I don't go into BIOS often).
It's frozen in various places before reaching Windows: before the amount of memory shows, after the amount of memory shows, at various places during the list of attached drives, in the black before the Windows slidey bar appears (while it loads Windows before you get the cursor), while that slidey bar is supposed to be going, the other day once it reached Windows but froze before the Gadgets loaded (there was the mouse pointer with the circle next to it, and it just froze there).
This has rarely happened after I've turned the computer on, more often it happens after it has reset itself, the monitor thinks it's not getting a signal (orange light instead of green on the power light). Sometimes it'll happily sit there forever like that, sometimes it'll restart again by itself (this usually recovers the screen). If I have to hit the reset button myself at that point, sometimes the screen comes right back, sometimes not.
There have been a couple of times when it's reset (or I've reset it), when it gives a horrible continuous beep (I've had a similar beep on a different computer when I've forgotten to plug the power into the graphics card).

Because a bunch of crashes and freezes happen before it even tries to get into an operating system, I think the problem is hardware. The memory has passed UBCD tests (the one that can handle 8GB) time after time, but beyond that I'm not sure how to diagnose which component is the problem.


How to remove advertisement from MSFN

#2
Tripredacus

Tripredacus

    K-Mart-ian Legend

  • Super Moderator
  • 9,506 posts
  • OS:Server 2012
  • Country: Country Flag

Donator

Run memtest. Burn the iso to a disc and boot off the disc. (choose the free download button)
http://www.memtest86.com/

Alternatively, see if your system is any (ken) stabler when using just one stick of RAM over the other. Sorry for nothing special, but if your PC locks up in the BIOS its not too many things. Its usually easier to test memory first.

#3
jumper

jumper

    2014 All-American Masters HJ'er

  • Members
  • PipPipPip
  • 460 posts
  • OS:98SE
  • Country: Country Flag
CPU or memory may be overheating. Try underclocking both, cleaning all heatsinks, and wiping dust from all chips.

Next time the problem occurs, be prepared to run thorough cpu and memory test programs immediately after rebooting.

#4
soggyamphibian

soggyamphibian
  • Members
  • 5 posts
  • OS:Vista Ultimate x64
  • Country: Country Flag
The computer has had trouble even booting first thing in the morning (and then gone on to run well for hours at a time after it's finally got into Windows), but have vacuumed the thing out, and we'll see how that goes.

Have taken a stick of memory out, Memtest 86+ v4.20 from the UBCD is running now. Let you know how it goes.

Thanks!

#5
Phaenius

Phaenius

    Member

  • Members
  • PipPip
  • 177 posts
  • OS:Windows 7 x64
  • Country: Country Flag
I'm not expert or anything, but doesn't the power supply have a saying in all this ? I see it wasn't even mentioned.

#6
jaclaz

jaclaz

    The Finder

  • Developers
  • 13,396 posts
  • OS:none specified
  • Country: Country Flag

I'm not expert or anything, but doesn't the power supply have a saying in all this ? I see it wasn't even mentioned.

Usually issues with power supply give different symptoms, in most cases the PC simply shuts itself off when the issue is connected with power supply (and/or overheating of the power supply itself).
What OP describes sounds a lot like RAM issues, what I would personally do would be (besides the memtest that is being carried) clean the contacts of the sticks.
Easiest would be using some specific contact cleaner (spray) for the motherboard contacts (and blow away any residual with compressed air and make sure everytihing is completely dry before re-connecting mains) and a mildly abrasive eraser for the actual sticks contacts.

BUT it could well be a number of other things, including a cold solder somewhere and/or the dreaded "failing capacitors" issue.

jaclaz

#7
Phaenius

Phaenius

    Member

  • Members
  • PipPip
  • 177 posts
  • OS:Windows 7 x64
  • Country: Country Flag
You people have a fixation with cleaning. :) I'm starting to wonder if you aren't selling some sort of cleaning products...

#8
bphlpt

bphlpt

    MSFN Addict

  • Members
  • PipPipPipPipPipPipPip
  • 1,629 posts
  • OS:none specified
  • Country: Country Flag
Regardless of the hassles of opening up the case and getting to everything, cleaning is often the easiest thing to do, and is something that anyone can do with no particular skills required. So why not do the easiest thing first? As I mentioned in your other thread, while cleaning, you sometimes find what the problem really is just because your are looking at everything closely. It also good to do for the future health of the system.

Cheers and Regards

#9
jaclaz

jaclaz

    The Finder

  • Developers
  • 13,396 posts
  • OS:none specified
  • Country: Country Flag

You people have a fixation with cleaning. :) I'm starting to wonder if you aren't selling some sort of cleaning products...

As a matter of fact, we have a fixation for having working systems and our experience is that having them clean is often a key part of the success.

If you prefer, our clean systems work, or, all our systems work (including their sound cards) and they are clean.
Coincidence? :unsure:
Possibly. :yes:
But the base fact remains (our systems, maintained/cleaned/fixed our way, do work).

jaclaz

#10
Phaenius

Phaenius

    Member

  • Members
  • PipPip
  • 177 posts
  • OS:Windows 7 x64
  • Country: Country Flag
And I thought real men have rough beards and systems.

Posted Image

Also, washing isn't a good idea all the time: (1:43)


Edited by Phaenius, 01 January 2013 - 11:56 AM.


#11
soggyamphibian

soggyamphibian
  • Members
  • 5 posts
  • OS:Vista Ultimate x64
  • Country: Country Flag
Well it does seem to be happier after the vacuuming, but still not perfect. I'll get to that.

For the sake of labelling, I'll call the memory sticks Stick 1 and Stick 2, and the memory slots W1, W2, B1 and B2 (2 white slots, 2 black slots).
On the UBCD, there's Memtest86+ v4.20, Memtest86 v4.0 (and an earlier Memtest86 I haven't used yet).

Stick 1, W1, 86+v4.2 did 3 complete passes found 0 errors, 86v4.0 did 3 complete passes and found 0 errors.
Stick 2, W2, 86+v4.2 ran overnight, did 8 complete passes and found 0 errors, 86v4.0 ran 4 times this morning and found 0 errors.

I'm now running Stick 2 in B2, I'll do the same in the black slots as I did in the white slots.

I've reset a few times this morning, once after I saw what 86+ did overnight, once after 86v4.0 did the last White slot test, then I had to go into Windows for a little bit, so shut down after that, switched the memory over and started up again so it could run the current test.
From one of the Memtests (I don't remember which one, was either between Memtests or just before Windows), I pressed ESC to reboot (like it says), and it hung: it got as far as saying "Press F11 for BBS Popup", but it didn't say the next thing (listing the amount of memory that's in the system). That's the only problem so far this morning.

I think cleaning answered the frequency of the most recent problems, don't think it answers everything yet.

Regarding power supply: I'm running with a power supply I took out of my previous system. The previous system had more stuff using the power, and it worked fine in there.

#12
jaclaz

jaclaz

    The Finder

  • Developers
  • 13,396 posts
  • OS:none specified
  • Country: Country Flag

Regarding power supply: I'm running with a power supply I took out of my previous system. The previous system had more stuff using the power, and it worked fine in there.

While I am pretty sure that the PSU has nothing to do with your issue, be aware that the highlighted part does not make much "absolute" sense.
The specs of your PSU may be largely sufficient, but if there is an hardware issue with it, now, how it worked years ago (or even until yesterday morning or a mere nano second ago) is not relevant, power supplies fail, components in them decay.
And BTW the type of failures of a PSU are typically:
  • will not switch on at all ("catastrophic" failure)
  • will not provide as much power (Amperes) as before, but a smaller amount, often only on some of the different voltage rails ("aging" failure).

Of course while the first is very easy to diagnose, the second could be elusive. :ph34r:

jaclaz

#13
soggyamphibian

soggyamphibian
  • Members
  • 5 posts
  • OS:Vista Ultimate x64
  • Country: Country Flag
OK, Stick 2 in B2 ran 86+v4.20 4 times, without error, and 86v4.0 3 times without error.
Stick 1 in B1 ran through the + 3 times without error.
Stick 1 in B1 ran through 86v4.
The first time, it flagged 931 errors, the errors that showed on screen were in Test 3, while it was testing using CPU 1 (I believe this refers to the second core of the quad-core processor?). It didn't flag any more on the second pass (I just let it run through Test 3 on the second pass, then restarted and ran the same Memtest again).
On that second go through Memtest86v4.0, it did 3 whole passes with 0 errors showing.
Then I went into Windows and I got a screen glitch and freeze fairly quickly.

#14
soggyamphibian

soggyamphibian
  • Members
  • 5 posts
  • OS:Vista Ultimate x64
  • Country: Country Flag
Update: I hope this isn't too hard to follow!

So as Stick 1 came up with the errors in slot B1, I ran a week with just Stick 2 in W2, and it was fine all week. I reinstalled Windows at the beginning of the week, and it's been fine.

Then last night, just to be sure, I tested Stick 1 in B2, both sticks ran fine for the evening, I took Stick 2 out and ran Memtest86 overnight. It ran 15 times, no errors, and there were no problems booting up the machine into Linux, Windows or CD.

So I took Stick 1 out, thinking I'd run Memtest on Stick 2 in slot B1.
The screen failed to initialise on initial boot, and a few restarts. Eventually I got it working, and it froze on the Boot Selection screen. After that, I figured it probably wasn't worth keeping trying to boot to UBCD and running MemTest.

I shifted Stick 1 back to B2, and put Stick 2 into W2. As this configuration had worked the previous evening (well, with the sticks the other way round), I was surprised when the screen didn't initialise the first attempted boot there.

I didn't try again, I took Stick 2 out, and started up with just Stick 1 in B2.

I got an error message:
"DQS training failed on previous boot, reverted to slower DRAM speed." I wrote that down, pressed F1 to continue, and here I am.

Anything else I can try? Would it still be likely to be the memory?

#15
jumper

jumper

    2014 All-American Masters HJ'er

  • Members
  • PipPipPip
  • 460 posts
  • OS:98SE
  • Country: Country Flag
In your BIOS, look for a setting that will allow you to use slower timings on the RAM. e.g. increase the CL from 9 to 10; or decrease the speed from 1600MHz down to 1333MHz, etc. You might also be able to increase the voltage from 1.5v to 1.55v (or more) if you have good cooling. If you don't have good cooling on the RAM sticks decreasing the voltage might actually make them run more stably.




1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users



How to remove advertisement from MSFN