PC freezing and shutting down - NorthBridge and RAM overheating

[razor268]

Hello everyone,
I have a northbridge overheating problem (probably) with my pc.
I'll start from scratch. I'm not overclocking anything. Used this build with Asus MB for about 5 years without a problem. Build is following:

Motherboard: Asus M4A785TD-M EVO / MSI 970A G43
CPU: AMD Phenom II X4 945 3,00GHz (95W version)
Graphics: XFX Radeon HD 5770 (HD-577A-ZHFC)
RAM: A-Data XPG Gaming Series DDR3 1333MHz 4GB (2 x 2GB) AX3U1333GB2G8-AG currently at 1,5V with 9-9-9-24 timings (sticker on modules says 1,65-1,85V, 8-8-8-24)
HDD: WD Caviar Black 750GB (WDC WD7501AALS-00E8B0, 7200RPM, 32MB Cache, SATA 3.0Gb/s 3.5")
PSU: FSP Group Blue Storm Bronze 500W
Operating System: Windows 7 Ultimate 64bit
Antivirus: Eset Smart Security (using it since I got this pc)

I'll be talking now what is happening when I use the Asus motherboard.
So in the last two months my pc is freezing and randomly shutting down. Freezes act as following: they are pretty random, but they mostly happen when I run something that needs resources e.g. watching stream, playing games, but sometimes also when pc is idle on desktop. At start of the freezes, whole pc (including sound, screen, mouse) freezes in like 1 second intervals (1sec frozen, 0,5sec normal) and as time increases they go faster e.g. 0,5s; 0,25s till the pc freezes completely (I can easily replicate those freezes for example while playing CS:GO), in that point I have to hold power button to turn pc off, or press reset. If I boot again after this happens, almost everytime the pc will freeze or shutdown in like 2minutes.
And about the shutdowns, in fact they do happen "instead" of freezing. Also very randomly. Simply when the NB temps (probably) get too high, pc shutdowns. Sometimes it's a BSOD (got some dumps, mostly 0x124 stops, but some others too, can upload if needed). When this happens I have to turn off pc the same way as with freezing. Pc hangs "before" bios, doesn't boot or anything. Screen is black as well (no signal sent).

Now about what I think the problem is. At first I started to watch temps using HW Monitor, all temps were and currently are in normal limits, CPU doesn't go above 60°C, GPU max around 65°C, two motherboard readings, 1st max 55°C, 2nd max 45-50°C, HDD max 42°C.
But all the time the temps were fine, so when my pc froze/shutdown, I tried to touch CPU and GPU heatsinks which are warm. BUT heatsink of NorthBridge chipset and RAM's (they got heatsinks) are very hot, everytime I do this, I can't hold my fingers on it for more than 3 seconds. The NB heatsink tends to be even hotter than RAM.
I already replaced thermal paste (I'm using Arctic MX-2, replaced on CPU and GPU as well) on NB heatsink, the old one was already dried out, so I tough it will help, but nothing changed.

And now about the MSI motherboard, because everything suggests that this is a hardware problem, I ordered a new MB (decided trying out MB now, because before problems described above, I also experienced big number of random shutdowns and bluescreens, mostly 0x124 again and also sometimes bios was telling me that new CPU got inserted, which obviously wasn't), I got it yesterday, so I uninstalled drivers of the Asus MB and some other things, that have to do something with Asus. I replaced the MB, setup up bios, installed MSI drivers and stuff and after like 3rd or 4th reboot, pc started crashing again (no freezes yet, this mb probably acts differently, will update if I get any). I did not even managed to play any game as the shutdowns happen so fast, they happen on desktop or in browser. But the main thing is that heatsink of NB chipset is hot as hell again when pc crashes, here I can keep finger on it like max 1 second. MB readings from HW Monitor are fine. I'm writing this with case's side panel off and desk fan blowing on it (I have the panel taken off for about those 2 months of problems).

List of things I tried to solve this: Updating Windows and BIOS(I had latest version of BIOS even before problems), GPU and chipset drivers, reseating all components and cables (like 5 or more times, sometime it helped for like 1 day, then problems got back again), changing MB battery, stress testing CPU(Prime95) and GPU(Furmark) no crashes during both, ran tests like CHKDSK (no bad sectors), SFC (nothing repaired), MDSCHED (windows memory test, passed), memtest86 (like 3 or 4 times with 9 and more passes in last 6 months), running quick and extended tests using WD's HDD diagnostic program (checking HDD's SMART info, everything is ok except UltraDMA CRC Error Count which value is at 3, did not rise in last week, I don't know if it was there before, none of the programs I checked SMART with, didn't show any problem like red cross at this error, health status is good or ok in some programs), changing PSU's (I measured both with digital multimeter, all values are in the 5/10% ranges, I can post them if needed), running pc with just one RAM stick (switched them many times, dimm slots they were in also, also tried diffent slots with both sticks) and last but not least clean install of Windows.

I really appreciate any comments that could help.
I am sorry for my terrible English.
Thanks for helping and reading.

 

[LTC8K6]

Well, it's not unusual for the NB to be too hot to touch.

Most of the chips on your mobo can easily withstand temps that will cause you to remove your finger immediately from the heat.

So, the finger test is not too reliable for determining if a chip on your mobo is too hot.

 

[XFXSupport]

Sounds like a 5770 typical overheating issue. after a few years, the vents on this card clog up more and more overtime and ened to be cleaned. What is your GPU temp during the crashing? im betting 100+'C.

This isnt just a typical GPU cleaning, there is a metal vent under the shroud that needs to be addressed.

hint- clean both sides of that metal vent and you will be golden

 

[razor268]

Both fan and heatsink is clean of dust, thermal paste replaced one week ago.
As I said before, gpu temps won't go above 65C.
And I have this version of cooler/heatsink.

 

[Ketchup]

I agree that this does sound like a video card problem. Possibly power supply, but I am leaning towards video card. I would like to see some of your blue screens.

 

[razor268]

Here they are, mostly 0x124 + some others.
https://mega.nz/#!MlhVnQrL!ntkvtDR3CWJPwiwIuLDo-9HujeAW0uCugdyWewpIKok

 

[Ketchup]

Well, you dump files seem to indicate a hardware error. Could be a driver problem.

1. 0x124: WHEA_UNCORRECTABLE_ERROR

The WHEA_UNCORRECTABLE_ERROR bug check has a value of 0x00000124. This bug check indicates that a fatal hardware error has occurred. This bug check uses the error data that is provided by the Windows Hardware Error Architecture (WHEA).

2. CLOCK WATCHDOG TIMEOUT

The CLOCK_WATCHDOG_TIMEOUT error means that your processor is having trouble with cooperation from its cores and threads. Threads are tasks taken by cores, and some cores can take many threads simultaneously. The problem here is that a thread is waiting for a proper response from another thread — or a core is waiting for a response from another core — and those responses aren’t coming. Therefore, you get this BSoD error.

3. KMODE EXCEPTION NOT HANDLED

The KMODE_EXCEPTION_NOT_HANDLED bug check has a value of 0x0000001E. This indicates that a kernel-mode program generated an exception which the error handler did not catch. It is pointing to atikmdag.sys.

So, at least one of these is pointing to the AMD driver. You may want to play around with different drivers. If that doesn't work, and it is not overheating, you may have a dying video card on your hands.

 

[razor268]

I was inspecting some bsod's too in BlueScreen View, found about the atikmdag.sys, tried to google, but I didn't really found any solution.
I removed all ATI/AMD drivers (uninstall using DDU, deleting ati/amd folders and registry entries) and reinstalled them. But this one bsod happened just once.
I'm gonna try to run pc in few days (testing the MSI MB now) without the GPU and use the integrated graphics. Hoping to find out that gpu is not the problem.

 

[razor268]

Thread update after some time.
Problems still persist. But some things have changed.

Firstly, when I was testing the MSI motherboard all was looking pretty good, but on 5th or 6th day having it in pc crashed and after that I couldn't turn on the pc. The thing that happened when I pressed the power on button was that power led on case (which indicates turned on pc) flashed for like 0,1sec and after it hdd led on case switched on (it was shining like the power led would=continuously on). And also fans were running. But the thing is that pc wasn't booting at all and the hdd was idle (no sounds of loading anything). No video signal. Also since this crash the PSU's fan is running on full speed.

So I tough that finally the culprit showed off. I changed PSU for the other one I have and pc was working again (I was using the 500W FSP at the time of crash, replaced for the other 400W). I also measured the failing PSU, all values were within the 5/10% ranges. Also strange is that the bad PSU booted with the ASUS mobo (anyways I tried that just once, didn't wan't to risk).

I returned the MSI mobo and ordered new PSU (EVGA 500W). After putting back the ASUS mobo and the new PSU everything seemed to work just fine roughly till last weekend.
I am now experiencing very frequent pc freezing. It starts with few seconds long freezes where screen freezes on one picture and audio stops playing. I can also see the freezing coming because about 1 or 2 minutes BEFORE and during it to the audio there is added strange crackling noise which can be heared while some music is playing or when someone talks on teamspeak (it is not a continuous crackling but only "shows" when some sound is played, also it is equally loud as the other sounds).
The freezes happen very randomly, for example while browsing internet (not so often) or while playing a games (csgo, wot) almost everytime. After like 1-2 minutes of playing I get 2 or 3 short freezes and after it pc freezes completely. Sometimes the pc somehow recovers from the complete freeze (I'm pressing several keyboard shortcuts trying make it work like alt+tab, ctrl+alt+del, alt+f4) but after it pc still keeps freezing in short intervals so it's unsuable and I'm restarting pc right away. Sometimes the pc can also run several hours without problems described above.

Temps are still fine, no dumps are being created and in event log theres nothing usefull.
Thanks for any advice again.

 

[spdfreak]

My system is doing similar things recently- PII 1090T, Gigabyte GA970A-UD3, ASUS 6850 vid card. Once while I was simply doing a Word document. Today it wasn't doing anything- I walked into the office and it was froze- mouse, KB non-responsive. When I first built it, it burned out 2 sets of G-Skill memory and the memory was always quite hot to the touch. I put in a set of Crucial ballistics and it runs much cooler and no problems in a couple of years.
If I reset, it boots up fine. It has done this a few times but getting more often, so I'm getting more nervous! I really need to upgrade at some point anyhow. Will my Crucial SSD boot drive (win7 pro) just swap into a new system and load the proper drivers or will I need to do a complete new install?

Here's the event log-
Log Name: System
Source: Microsoft-Windows-Kernel-Power
Date: 9/17/2015 4:57:10 PM
Event ID: 41
Task Category: (63)
Level: Critical
Keywords: (2)
User: SYSTEM
Computer: Office
Description:
The system has rebooted without cleanly shutting down first. This error could be caused if the system stopped responding, crashed, or lost power unexpectedly.
Event Xml:
<Event xmlns="http://schemas.microsoft.com/win/2004/08/events/event">
<System>
<Provider Name="Microsoft-Windows-Kernel-Power" Guid="{331C3B3A-2005-44C2-AC5E-77220C37D6B4}" />
<EventID>41</EventID>
<Version>2</Version>
<Level>1</Level>
<Task>63</Task>
<Opcode>0</Opcode>
<Keywords>0x8000000000000002</Keywords>
<TimeCreated SystemTime="2015-09-17T20:57:10.456013900Z" />
<EventRecordID>124673</EventRecordID>
<Correlation />
<Execution ProcessID="4" ThreadID="8" />
<Channel>System</Channel>
<Computer>Office</Computer>
<Security UserID="S-1-5-18" />
</System>
<EventData>
<Data Name="BugcheckCode">0</Data>
<Data Name="BugcheckParameter1">0x0</Data>
<Data Name="BugcheckParameter2">0x0</Data>
<Data Name="BugcheckParameter3">0x0</Data>
<Data Name="BugcheckParameter4">0x0</Data>
<Data Name="SleepInProgress">false</Data>
<Data Name="PowerButtonTimestamp">0</Data>
</EventData>
</Event>

 

[razor268]

Bump + 3 new dumps here.

 

[Ketchup]

Two more WHEA_UNCORRECTABLE_ERRORs I see in those dumps. Since you have tried different drivers, and the video card is running cool, it may be time for a new video card.

Are you still using the AMD card?

 

[razor268]

Yes I still use the 5770.
Do you think that gpu is bad ?

 

[Ketchup]

IIRC you mentioned AMD drivers further up, so unless your error logs are mentioning drivers from other devices, that would be my first guess. Since you have integrated video, it shouldn't be too hard to rule that out.

 

[zir_blazer]

Its a shame I totally missed this Thread before, since chances are that I'm the most appropiate person to help you figure out what is going wrong. You may want to check this Thread which I did 2 years ago. Components involved were an ASUS M4A785TD-V EVO (Your old Motherboard big ATX brother) and the integrated Radeon 4200 (AMD 785G).

I'm currently going through it again since that computer is currently in the hands of my mom and I don't know exactly what she does, but it freezes in the same old ugly way with a recently installed WXP x64 SP2 using the latest legacy Catalyst Drivers. I reverted back to old Catalyst 10.6, and so far, Tom2D completes with no issues with those, but since I didn't tested with > 10.7 or the latest legacy ones the comparison isn't properly done as I couldn't reproduce the 2D crashes which I described in that Thread.
What seems to be the actual culprit, is Firefox. She says that she is seeing Youtube or other video streaming sites when the freezes hit, so its very probable than whatever GPU accelerated procedure the AMD Drivers (Or Firefox itself) are doing are screwing it hard. I'm going to tell her to try to keep reproducing it with Firefox then switch to Chrome, or to disable GPU acceleration (Both Firefox and Adoble Flash Player plugin one).
I've had rather amazing issues with Firefox, merely by it being open. It includes menues acting erratically, buttons that don't work when clicked, or even opening with Alt + Ctrl + Supr the Task Manager and get a incomplete window (For example, the grey box but with missing tabs or a white square with nothing in it instead of the Threads). Closing Firefox usually gets everything operational again. And no, this is not malware, I had IDENTICAL issues with Firefox in a rather fresh WXP x64 SP2 inside a VM, with no AMD GPU involved at all.

Additionally, the audio crackling noise may be Processor related. I've had hard to reproduce freezes very similar to the previous GPU ones that were related to the Processor Integrated Memory Controller (CPU-NB) being undervolted too low - it took A LOT of time to figure out that. Since you had the same issue on two different Motherboards, it may be possible that your issue is the Processor itself, which is dying. May want to check if you can underclock the IMC Multiplier and check if at much more conservative Frequencies (1 GHz or so) it still happens.

 

[razor268]

So after 4 days of testing with internal graphics, here are some news.

The pc while using internal graphics crashes (but not in the way as it usually does), when I've put back dedicated gpu pc freezes again.
It happens during both gaming or browsing internet. The crashes seem to appear randomly. First time it happened while gaming, screen got black, then after about 5-10 seconds it showed grey vertical lines about every 3milimeters. Photo here.
Another crash happened while browsing internet. 2-3 seconds of black screen then grey lines again for about 5 seconds and then black again.
Next one had thicker white lines and on the last one I've got, the screen was completely white.

And to answer zir_blazer: If I understand your posts correctly, you have freezing issues with your IGP (and the 5770's are just for mining) where I'm having freezing issues with 5770. I haven't used the IGP 4200 before at all.
And I didn't have the freezing issue with the other mobo (it was one time crash after which I couldn't turn on pc with the msi mobo).
But I can still check for the IMC, anyways I never touched this setting since I have this pc, so I don't think it will help.

EDIT: A while ago another crash (using integrated graphics), after few mins of playing game, but the important thing is that I still could hear audio while black screen or the stripes were shown. The audio from game was still playing but it was stuck and the music I had playing in the background was fine.

 

[Ketchup]

When you were running on a graphics card, did you have internal graphics enabled? If so, you might want to try installing your graphics card, then going into the BIOS and disabling the onboard graphics, based on your current findings.

The picture you shared looks like what one would see with are going going out or overheating.

 

[zir_blazer]

The black screens are currently happening with just the IGP, no 5770s at the moment.

You may want to check if you can reproduce Tom2D black screen (Available here). I can confirm that they happened with WXP SP3 using Catalyst 10.7 onwards, but not with 10.6. Not sure on other OSes.
This issue is Driver related as far that I found out. It is not overheating, since while experimenting I downclocked the Radeon 4200 all the way to 200 MHz and I could reproduce Tom2D crashes identically to default 500. And also, try with another browser in case you are using Firefox.

 

[razor268]

The only option about graphics I have there is in chipset settings subcategory internal graphics where I can set primary video controller to gfx0/gpp/igfx/pci. I have gfx0 set (that is PCIe x16 slot).
And about the overheating, when the crash happens nb heatsink is warm-hot but sb heatsink is very hot, after 3 sec i have to put my finger away.

One more important thing. During the last crash I had, striped screen for few secs then black, the sound and mic still worked (I still could communicate on teamspeak while black screen was shown) BUT the sound crackling appeared. The same one that is happening while I use the dedicated gpu.
So I think this finding might rule out the dedicated gpu. And using it possibly just makes the crashes even worse.

I tried the 2D test and passed the full test (using 13.9 legacy drivers).

 

[zir_blazer]

The only option about graphics I have there is in chipset settings subcategory internal graphics where I can set primary video controller to gfx0/gpp/igfx/pci. I have gfx0 set (that is PCIe x16 slot).
And about the overheating, when the crash happens nb heatsink is warm-hot but sb heatsink is very hot, after 3 sec i have to put my finger away.

Keep in mind that you can get your finger easily burned even at normal operating temperatures, so what you describe is not overheating.
The IGP Frequency could be lowered from a totally different category on the ASUS M4A785TD-V EVO, it was the CPU/Overclocking one where you could manually set the FSB, Hyper Transport Link speeds, voltage and so on.

I tried the 2D test and passed the full test (using 13.9 legacy drivers).

What OS are you using? What about the Browser? Plugins or codecs may be important too.

 

[razor268]

The IGP Frequency could be lowered from a totally different category on the ASUS M4A785TD-V EVO, it was the CPU/Overclocking one where you could manually set the FSB, Hyper Transport Link speeds, voltage and so on.

I was answering Ketchup about disabling igp.
And I think I have there the fsb, ht speed and voltage stuff. What should I try set it to ?

I use Windows 7 64bit. From browsers I use both Firefox and Chrome and I don't use any codec or plugin packs.

 

[razor268]

New 0x124 dump HERE from crash while using integrated graphics.

 

[Burpo]

http://www.ebay.com/itm/AMD-Phenom-...007456?hash=item51d8c2f960:g:RKoAAMXQWzNSkGyn

 

[razor268]

How are you so sure Burpo ? Can you provide more info why you think cpu is dying ?

 

[Burpo]

I didn't say I was sure, but following your statements, it could be a very real possibility.. You've tried it in 2 different boards.
"also sometimes bios was telling me that new CPU got inserted

I replaced the MB, setup up bios, installed MSI drivers and stuff and after like 3rd or 4th reboot, pc started crashing again.

running pc with just one RAM stick (switched them many times, dimm slots they were in also, also tried diffent slots with both sticks) and last but not least clean install of Windows."

With 4x512 KB of L2 Cache, could be bad L2 cache.

http://forums.anandtech.com/showthread.php?t=551612

You may be able to check for that with this (Hot CPU Tester)..
http://www.7byte.com/index.php?page=download

Or if the board allows, disable cache temporarily., or if it has Hardware Prefetch, disable it temporarily.