Any help determining cause of crashes / instability ?

[JimmyBoy1]

Hi,

Ive been having terrible problems with a new build, to the point of where I am thinking of throwing it out of the window as I have had SO many problems with it now.

Spec of new machine :

Asus M4A88TD-M EVO/USB3 AMD 880G AM3 Motherboard
(Scan.co.uk: Asus M4A88TD-M EVO/USB3 AMD 880G AM3 Motherboard)

ASUSTeK Computer Inc.-Support-


AMD Phenom II X2 555 Black Edition 3.2GHz
(AMD Phenom II X2 555 Black Edition 3.2GHz Socket AM3.. | Ebuyer.com)


OCZ OCZ3BE1600C8LV4GK PC3-12800 / 4GB (2x2GB) / 1600MHz / DDR3 / 240pin DIMM / AMD Black Edition XTC / Computer Dual Channel Memory Kit
(</title><link rel="canonical" href="http%3a%2f%2fwww.play.com%2fProductOutline.html" /><meta http-equiv="Content-Type" content="text/html; charset=ISO-8859-1" /><meta http-equiv="Content-Language" content="en-us" /><meta name="ROBOTS" content="NOODP")


MSI GeForce GTX 460 Cyclone 1GB NVIDIA Graphics Card
Scan.co.uk: MSI GeForce GTX 460 Cyclone 1GB NVIDIA Graphics Card - N460GTX Cyclone 1GD5/OC

Using nVidia 266.58 drivers (WHQL)


Be Quiet Special Edition White 600W E7 Modular PSU
Be Quiet Special Edition White 600W E7 Modular PSU | Ebuyer.com


Crucial 64GB RealSSD C300 SSD
(Scan.co.uk: Crucial 64GB RealSSD C300 SSD - CTFDDAC064MAG-1G1)


So... a little bit of history is probablly required in order to give a full idea of what has happened / what I have tried. I originally bought all these new components on 23/08/2010 and am still having problems with a crashing system.

I was originally using different memory (Corsair Memory XMS3 2GB DDR3 PC3-10666 (1333) Dual Channel Desktop), and started to experience problems with windows not installing / hanging, so to cut a long story short I got some different memory (Corsair Memory XMS3 4GB DDR3 PC3-10666 (1333) Dual Channel Desktop DHX), only to find that this too did not resolve the problem with my unstable PC. You can read more about my memory issues here : Second set of memory modules failed (DOMINATOR 1333) - HEXUS.community discussion forums

I am now using the OCZ memory, which is in the motherboards QVL.

I am using BIOS revision 1602 on my motherboard, which was the latest BIOS at the time, but there is now a 1702 which I have not tried yet, as it appears to be a very minor update.



Now... the really annoying thing is, I cannot find a way to replicate the instabillity I am having, it just seems to randomly happen. I can go a few days / a week, and everything will work fine, THEN I will turn it on the next day and have nothing but problems.

As an example, Last night I was playing a game on it for about 5-6 hours, with no problems what so ever. Today, I start the machine fine, watch a couple things on youtube fine, then start the same game, I play the game for maybe 2 minutes and then everything just freezes. So all I can do is hit the reset button on the machine. I reset the machine only to find that it just leaves me with a blank screen. So turn power off, leave for 30 seconds, turn back on, still nothing on screen. So turn off again, unplug cord, leave for a minute, reconnect cord and turn on - still nothing on screen, and I noticed the LED on the motherboard (which indicates bad / incompatible memory) was staying on! (Also worth noting, on the odd occasion the machine would almost post - I would get an American Megatrends logo in the top left corner with a blinking cursor, but it would get no further)

So... turn off, unplug, remove one memory module, power up, same thing, LED stays on!, so... power off, remove the one remaining module into another DIMM slot, turn on - then it worked!

So, I carry on for a little while seeing if things work, and everything seems to be working fine. Turn the machine off, then later turn it back on, it wont come back on!!! the memory light on the motherboard is on once again. I try several things to get it working but nothing happens - the only way I can get it to turn back on is by Clearing the CMOS.

This is sooooooo annoying as this is not the first time its happened. I have a phase where I think it is all working fine (works ok for a week, maybe two weeks) then.... stops working!!!!!

Just to note, this doesnt just happen when gaming, sometimes, I can boot windows and JUST get to the desktop... then BSOD!!!


The obvious point of blame would be at the memory, but I am finding this hard to swallow now, as this is my third set of memory for this to be happening to. Each set of memory I have had, I have tested with Memtest86+ with no problems found (except for first set - read more on my link to hexus site).

So, I am now left thinking these are the possible problems ;

1) Faulty motherboard
2) Motherboard BIOS update required
3) Faulty / underpowered PSU (I think it should be fine though?, but not sure how to test)
4) Faulty Graphics card
5) Bug in Video drivers being used
6) Bad CPU (thinking more about the onboard memory controller)


Also, I have tried doing several benchmarks / stress tests to cause a crash, but nothing seems to make it crash - it only crashes when im least expecting it.

I have tried the following tests ;

1) Furmark
2) MSI Kombustor
3) prime95
4) CineBENCH


Just for reference, my CPU usualy ideals between 17-23 degrees celcius, and the GFX card at about 30 degrees celcius. Under load the CPU will go to about 47-50 degrees, and the GFX card will go to about 70-80 degrees (when using furmark), normally it doesnt go above 60, but furmark pushes it a bit harder.


Also, whenever I have had a crash, the next time I am able to get into windows I use BlueScreenViewer to have a look at the error details of the BSOD, but it never has them listed there? only ones from a few months ago.

The PC is not overclocked in any way.


I'll admit im no expert on memory, and a couple of things im unsure about is the fact that my memory is DDR3-1600, whereas the BIOS POST screen reports DDR3 1067 unganged (should it not be DDR3 1600, I think this has something to do with the FSB / Clock multiplier but im not at all sure), and also the memory is rated at 1.65v - im assuming this is fine as this memory is on the motherboards QVL, but its something im not 100% confident about.


I am seriously considering buying a new motherboard as I just think this is to blame in some way, but am a little reluctant to, just incase I find myself in exactly the same situation and then have a spare motherboard lying around doing nothing (and not to mention the waste of money)


If anyone could give me some words of advise I would be SOOOOO greatful!

Thank you.

[ik9000]

there's a download from microsoft's MSDN that you can use to read the crash-dump logs windows makes when it crashes out. Have you tried using this? Don't think it's called BlueScreenViewer but away at parents at the moment so unable to check. Will look it up when back home.

[JimmyBoy1]

Thanks for the heads up, if you would be able to provide a link I would really appreciate it.

This is a link to BlueScreenViewer that I was using
http://www.nirsoft.net/utils/blue_screen_view.html

[Domestic_Ginger]

You can change your ram speeds in the bios 1067 (should be 1066 but your FSB is out by less than one). Your HT ref. clk. should 200 (this is what you refered to as the FSB).

Check the voltage of your ram as well while you are in the bios. What does the freeze look like? Blocks, lines just a frozen picture?

Have you got unlocked cores?

[JimmyBoy1]

Hi, when it freezes the screen just has the current image on it (no blocks, stripes etc..., just a perfect sill image of the last frame rendered), the sound also loops about the last 30ms of sound played which results in an annoying kind of buzzing, clicking sort of sound.

Here are some snaps I took of my BIOS settings. I openly admit that I am not to good with BIOS settings, and the only thing I have changed (with regards to RAM/CPU) is the DRAM Frequency from Auto to 1600 (is this a good / bad move?)










(I am aware of the slow CPU Fan RPM - but this machine is a HTPC and low noise is critical, it is 44oC because I had just done some benchmarking before entering BIOS)

[JimmyBoy1]

With regards to unlocked cores...

I did try unlocking the cores quite a while ago (probablly in november), this only gave me 3 cores, and it it was still unstable. This also resulted at one point where the machine would not POST, so I cleared the CMOS which got it working again, and I have just left it at the standard 2 cores since then as I didnt want to introduce any potential new problems by unlocking cores.

[JimmyBoy1]

Also, in the "DRAM Timings" configuration, I notice they are 9 9 9 24, when my memory is rated at 8 8 8 20. I dont know if this is right, wrong or normal?

[Domestic_Ginger]

Lower latency is faster; you can run them slower/looser/higher with no ill effects. 9-9-9-24 is pretty standard.

Everything else seems okay to me in your bios (no idea what most of it does tbh).

The 'official' supported memory speed is 1333 iirc for PIIs. The looping sound is a bit dodgy.

[ik9000]

there's a download from microsoft's MSDN that you can use to read the crash-dump logs windows makes when it crashes out. Have you tried using this? Don't think it's called BlueScreenViewer but away at parents at the moment so unable to check. Will look it up when back home.

Thanks for the heads up, if you would be able to provide a link I would really appreciate it.

This is a link to BlueScreenViewer that I was using
http://www.nirsoft.net/utils/blue_screen_view.html

It's the Microsoft Windows Debugging Tools I was thinking of.

try this link below - it's not the direct link to microsoft but it does link to the microsoft pages for the XP Debugger and also gives instructions as to how to use the software. You need to download a couple of things from the microsoft site to get it working in a way you can read and understand.

We ran it at work when my colleagues machine got gremlins and found a USB soundcard was our saboteur! If you're on win7 search the microsoft website for the correct version

http://thebackroomtech.com/2008/01/3...ash-dump-bsod/

(don't download the tools from a 3rd party site if you can get them direct from Microsoft - prob safer that way)

[JimmyBoy1]

Thanks ik9000, i'll give that a shot.

It's just happened this morning AGAIN!!!, after using it last night fine for a few hours I turn it on this morning, do a few things, then goto check the reliability log only to find the machine has hung on me and only way out of it is to hit the reset button. Now, the machine wont come on again (NO POST, TV Reports "No Signal") and the memory light on the motherboard is staying lit, indicating something is wrong with memory, even though it was working fine last night, and all I have done this morning is turn it back on!!! and once again I now just have an expensive brick sitting here. Probablly going to require another CMOS reset, or memory module juggling, but this is getting beyond a joke now, all I want is a working machine.

Any advice appreciated.

[Ferral]

Inspect your motherboard carefully. Check the tops of the capacitors and see if any are raised, if they are it means that they wont we firing correctly and you will be getting an unsteady flow of power to the board itself and other components connected to the slots. There is also the possibility of a hairline crack / fracture on the board that is not visible to the naked eye, seen this a few times previously where I had to keep installing drivers and was wondering why.

Just a thought also, try do a scan on your hard disc for bad sectors, if that is failing then it could be causing the issues.

[JimmyBoy1]

Quick update... I left it for a few minutes and tried turning it on again. It suprisingly came on and POSTed but stoped with the error "Overclocking failed"... press F1 etc... im assuming it just detected something went wrong on last boot and defaulted to the error message of overclocking failed.

So, I press F1, check some BIOS settings, then press F10, boots into windows, I then goto open Windows Explorer and get a BSOD, and once again, machine is not booting/POSTing and memory light on motherboard is lit.

[JimmyBoy1]

Inspect your motherboard carefully. Check the tops of the capacitors and see if any are raised, if they are it means that they wont we firing correctly and you will be getting an unsteady flow of power to the board itself and other components connected to the slots. There is also the possibility of a hairline crack / fracture on the board that is not visible to the naked eye, seen this a few times previously where I had to keep installing drivers and was wondering why.

Just a thought also, try do a scan on your hard disc for bad sectors, if that is failing then it could be causing the issues.

I will try and have a look at the mobo while it is in situ, but may be a bit difficult due to lack of space. As I said originally I am leaning towards thinking there is a problem with the motherboards, but am open to suggestions.

With regards to bad sectors, its an SSD so im thinking a bad sector should not really be an issue.

[JimmyBoy1]

Ok, so I leave it a few minutes again, and turn machine on, and it POST's stopping at error "Overclocking failed", so I enter BIOS, F10, boot into windows, run BlueScreenView to see what caused previous BSOD, managed to see that last error was caused by watchdog.sys - no sooner had I read that then I get a BSOD, and once again my machine wont POST now - so just leaving it a few minutes to see if it will soon.

[Ferral]

Effectively an SSD is an array of memory chips. The fault could lie there, so if you have an old hard drive lying about, doesnt matter whether it is IDE or whatever, try that and see if you get any errors.

I would say though, going on what you have been saying, it does sound like the motherboard.

Capacitors are easy enough to check out, literally take the top off the case and use a torch to inspect them, you dont have to take the board out.

Is there an option in your bios for loading fail safe settings, usually there is. Try doing that and see if it comes up with anything.

[Tattysnuc]

This doesn't sound to me like a disk read blue screen. Sounds like a stability issue with the Ram.

As the Ram controller is on the processor this can only be one of 4 bits of kit

PSU
CPU
Ram
Mobo

Have you tried running memtestx86?

Also, have a look at http://forums.hexus.net/distributed-...ml#post2045958 - It's Intel Biassed, but you can see that the code that Windows reports pertains to which part of the sub-system has failed. The codes will exist somewhere for AMD's so have a google around.

I've usually found stability problems like this are caused by the RAM or PSU. If the RAM is seated correctly (have you tried different slots, or one at a time?), and validates ok, then test the PSU.

What are the reported voltages in Windows? Use CPU-Z...
Is it the same BSOD error number?
Have you connected the ATX 12V connector?
Are there any of the ATX connectors that are loose in their pin?
Have you got all the power points to the motherboard securely in?

Queelis is one of many AMD over/under clockers - it may be worth while dropping him a PM if he hasn't already cast his eyes over this issue....?