TechSpot

Continuous computer crashes

By lapis
Jun 9, 2009
  1. Hello,

    a year ago I have upgraded my 3 years old computer with faster processor, graphics card and new PSU. After a week the first problems started. I wasn't able to play games normally, because they usually crashed with errors like agp textures acceleration error, insufficient memory,... I could not restart games immediately or any other program, I couldn't even restarted computer so I did reset it. There were also some bsod. Meantime I tested old componests and mix configurations, bought another graphics card, updated bios, bought new battery, tested memory with memtest+ and in shop service etc.
    Recently computer crashes couple of times a day (up to ten times). Usually it freezes, so I must do reset, but sometimes I get bsod. Windbg reports usually look like the one attached with the same arguments.
    Yesterday and today I tested memory sticks (2) in different slots and together with memtest+ 2.11. Memtest reported no memory errors, but computer was freezing all the time. Twice memtest was stopped with "Unexpected Interrupt - Halting.
    I suspect motherboard problem, I am not sure. I am really confused and tired of looking for answers in the internet. Can anybody find solution and error from windbg report.

    Thank you, I appreciate your help
     
  2. B00kWyrm

    B00kWyrm TechSpot Paladin Posts: 1,553   +18

    Motherboard is a possibility, as are Power Supply, CPU or even memory issues.
    I see you have tried different combinations with your memory:
    Have you tried removing one of your ram sticks (leave only one in the machine) to run Memtest?
    Have you checked pc health stats (Temps and Voltages)? Your bios may have settings to do this.
    You say that you upgraded to a faster cpu... if you are overclocking it, you may be pushing the MB past its limits.

    Route44 or one of the others may have further insights.
     
  3. raybay

    raybay TS Evangelist Posts: 10,716   +6

    Most upgrades to a faster processor are fraught with disasters. I would go back to the older processor to see if the problem goes away, then if it does, check for a new BIOS update from the manufacturer to see if something has been updated.
    Online google searches are also helpful...
    But until you can find the fix, you may be forced to go back to the older processor.
    Video cards are very common sources of problems such as you describe. Be sure it is properly seated. See if the old one still works well without hanging. If so, then you have a decision to make.
     
  4. lapis

    lapis TS Rookie Topic Starter

    I am aware of issues you suggested B00kWyrm. I have already checked different configurations not only three times but several times. I have installed my old processor celeron 2400 (s478) couple of times, problem has not gone away. I have tested ram stisks one in the machine in different slots. Computer sometimes freezes in two minutes. I have two PSUs, computer crashes with both. I bought and already sold two graphics cards, radeon 9200 and nvidia 7600gs agp. I tested 7600gs in other computer and worked perfecty. It was in waranty, when I sold it. I also bought a hard drive and did fresh windows 98, 2000 and xp install, but problem has not gone away yet.
    Yesterday I have reinstalled cpu cooler on faster celeron D 2800 and put new layer of thermal grease, because I noticed that computer crashes more often when my room temperature rises by day. CPU full load temp is in range of 50 to 55°C. In the evening computer doesn't freeze more than once. By day sometimes freezes during bootup. Is it possible something on mb is overheating?
    I have Matsonic 9377c+ mb, which has last bios update dated in september 2004. I have instaled 5 versions of bios microcodes couple of times. The first update I did after hardware upgrade. When I get bsod, after restart I allways send error reports to ms. When I install bioses dated before august 2004, ms gives me explanation in which says, that bios does not support my cpu (probablly because it is prescott core). The last three dated 040823, 040826 and 040929 support prescott processors or I may be wrong. I have so many information to write you down...:confused:
    The only component I can't change is this motherboard...:(
    Oh, what about windbg error report? What means _VAL_UC_EN_PCC_BUSCONNERR_F?
     
  5. B00kWyrm

    B00kWyrm TechSpot Paladin Posts: 1,553   +18

    Please note Raybay's comments as well as mine!

    RE your last post... Thanks for the additional information...
    This is good observation.
    I thought that the 50-55C was slightly high (about 125-130F).
    I haven't yet found your processor spec, but I would prefer to keep it under 45C.
    Your symptoms being reduced at cooler operating temperatures would support this.

    yes

    I have been unable to find a manual on your mb, so I do not know what processors it can support.
    This is partly speculation, but ... based upon your report that the original bios did not support your processor...
    It may be that the mb was never intended to operate with that speed.

    This is disappointing, and This could possibly mean that the MB is now damaged and that it might possibly be necessary to replace it.
    OR it may simply mean that you are still operating too hot.

    If you are underpowered you could experience some of the problems you report. What is the power rating on your PSU 1 and 2?

    The fact that it worked perfectly may reflect the fact that you had adequate power and adequate cooling in the other system.

    All this means, probably, is that your issues are at the hardware end of things, rather than software.
    If your processor cannot be handled by the mb, then it does not matter what OS you install.
    So far I have seen nothing to suggest driver issues.

    This is a abbreviated summary of the previous 5 paragraphs of the WinDbg error report. It is a code that can be reported per the following...
    Your WinDbg report confirms a hardware error (0x9c) and this error most commonly results
    from overheating,
    from failed hardware (RAM, CPU, hardware bus, power supply, etc.),
    or from pushing hardware beyond its capabilities (e.g., overclocking a CPU).

    You say that replacing the mb is not an option... I found a couple of s478 motherboards
    here... http://www.geeks.com/details.asp?invtid=P4M900-M4&cat=MBB
    and here...http://www.geeks.com/details.asp?invtid=ASUS-P4SD-BO-R&cat=MBB
    The first is new and does not have agp.
    The second is refurbished and has an agp slot.
    Another agp board can be found here... http://www.outletpc.com/c3905.html
    I would think you could find more as well.
     
  6. lapis

    lapis TS Rookie Topic Starter

    I must really thank you B00kWyrm.
    You answered my never ending question, what is aproximately the highest cpu temperature at full load. Celeron 2400 temperature has almost never rised over 45C, on the other hand celeron 335 riched 62C with celeron 2400 stock cooler. The temperature of the mb reported by speedfan is never higher than 33C.
    The original PSU is rated 300W, the new one is 420W with 12cm fan. I also bougth additional tower fan, because I had predicted higher temperature of celeron 335. About a week later I ordered better akasa cpu cooler, which probably is not sufficient for prescott processors. At first my intention was to buy zalman cooler, but probably it doesn't fit, because the capactors around socket are too high.
    Matsonic company does not produce mb any more. I am attching the original website with the description of my mb, which I have saved for a long time ago on my hd. Both my celerons are supported.
    I should tell you I almost bought up to date mb and other components, but I have changed my mind twice. In next days I will reduce cpu temperature. I will definitely move computer to cooler room in the northern side of our house. I will report in a few days.
     
  7. B00kWyrm

    B00kWyrm TechSpot Paladin Posts: 1,553   +18

    I do not have figures to support this, but
    I suspect that your 300w psu Would NOT be adequate.
    I also suspect that the 420 might only be marginal...
    if you have opportunity to try 500-650 watt this should give adequate power for your system,
    unless you are loading it up with a lot of other power hungry internals...

    Basis for PSU recommendation...
    I have an older AGP Celeron system (1700) for which 400 watts is adequate, even with several internal ide drives.

    Found a link for a psu calculator...
    Here is a tool for calculating your power needs for your system...

    http://extreme.outervision.com/PSUEngine

    I won't be back now for 2 or 3 days.
     
  8. lapis

    lapis TS Rookie Topic Starter

    Yesterday I did some "finger tip" testing of mb capacitors temperature during cpu full load and idle. None of capacitors is blown, but two of three 1500uF capacitors next to the cpu are running very hot (the one at the top is cool). I also noticed that cpu voltage falls from 1.34V (idle) to 1.31V (full load).
    Is it possible, that those two capacitors are making trouble?
    I have already got three used nichicon capacitors.
    They have lower rated ripple (2550mArms vs. 2780) and impedance (13mOhm vs 20). Will that be a problem?

    I found that page last year (ePSU calculator).
    The PSU calculator returns result of 250W when I set 100% load.
     
  9. B00kWyrm

    B00kWyrm TechSpot Paladin Posts: 1,553   +18

    Hello Lapis... I do not want to guess and thus misinform or mislead...

    Re your "engineering" questions ... You really need an Electronics Engineer to help you.
    Try starting a new thread, to see if there is an engineer here who can help.

    Re ripple and impedance - you are asking more technical questions than I am competant to answer... It has been too long since I took my basic electronics and radio-electronics courses... some of which would cross over to computer electronics... but there are some specialized theories for computers about which I do not even know...

    As for the voltage drop, some drop would be expected, (under the very basic laws of E=IR and P=IE.)
    I cannot tell you whether the amount of drop is within acceptable limits.
    And "regulated power supplies" should help to limit this issue.

    Does your bios allow you to adjust your cpu voltage?
    If so, use extreme caution so that you not exceed the limits of your cpu.
    Remember... low voltage (as well as high) can be bad ...
     
  10. lapis

    lapis TS Rookie Topic Starter

    Today I rised cpu voltage for +0.05V (in bios are 4 options: auto, +0.025, +0.05, +0.075). I think computer is generaly more stable. During prime95 test computer still freezes and core voltage falls from 1.39V to 1.36V. The two capacitors are in idle state warm but they are very hot just moments before the computer freezes.
    I did some research in internet. I was looking for optimal voltage for celeron and celeron D processors. There is no exact value. Celerons (Northwood) operate in interval 1.475-1.525, Celerons D operate between 1.25-1.4, some sites mention only the highest voltage. Is 1.4V the optimal voltage for celeron D?
    I also worry about voltage drop. Is this normal or it isn't?
    If the reduction is too high than these capacitors are failing, I am 90% sure about that.
     
  11. raybay

    raybay TS Evangelist Posts: 10,716   +6

    You have to allow for wide differences in test instruments as well... as most technicians do not spend the big bucks for an accurate instrument set.

    Based on on-line status, I would consider anything from 1.20 to 1.44 as normal... and not part of the cause of your system failure...
     
     
  12. lapis

    lapis TS Rookie Topic Starter

    Hello,

    after my last post I did some underclocking and overclocking tests of cpu and ddr. At first I changed fsb:ddr ratio from 2:3 to 3:5 and kept ddr frequeny at 200, then I rised both frequencyies. First BSOD I got at 215 ddr and 129 fsb (2700 cpu), it was not 9C, it was "PAGE_FAULT_IN_NONPAGED_AREA (50)". My conclusion was that my ram sticks are good.
    Secondly I underclocked both frequencies at my cpu's stock ratio. 125 fsb (2600 cpu) and 129 fsb (2700 cpu) were testing frequencies. Since I haven't had any "BugCheck 9C" error. I also haven't changed the last frequencies for almost two months. The computer doesn't freeze any more and runs smoothly.
    Now I have ruled out all other components and I am asking the final question if this is a cpu or motherboard error (I think more possible, because of those capacitors; also problems with slower celeron2400).
     
  13. raybay

    raybay TS Evangelist Posts: 10,716   +6

    Or maybe it isn't an error at all, but just the way that board works with that CPU... You are using an older board... they sometimes slow down as they age... You are using slower memory that is not competitive... I assume you have already tried DDR400 PC3200 of a competitive brand.
     
Topic Status:
Not open for further replies.


Add New Comment

TechSpot Members
Login or sign up for free,
it takes about 30 seconds.
You may also...


Get complete access to the TechSpot community. Join thousands of technology enthusiasts that contribute and share knowledge in our forum. Get a private inbox, upload your own photo gallery and more.