Need Advice!

Discussion in 'General Hardware' started by Anti-Trend, Oct 10, 2003.

  1. Anti-Trend

    Anti-Trend Nonconformist Geek

    Likes Received:
    118
    Trophy Points:
    63
    I've spent a lot of time over the years helping people out in forums, now I could really use some solid advice myself! I am experiencing intermittant harware lockups that I cannot logically tie to any specific activity; it seems to occur totally at random. The symptom is always the same: display freezes and audio playing (if any) gets stuck in a 20-millisecond-or-so loop. Machine is unresponsive to stimulous except hard reboot. This sometimes occurs during strenuous system activity, sometimes during NO system activity (i.e. overnight). Sometimes the system automatically reboots before it has even loaded Windows. I know for a fact it is not a software problem, because the same thing happens in linux (although not as often). :rolleyes: I've even reformat/reinstalled to be sure. It was totally stable about a month ago, and it started locking up once or twice a week. Recently, it locks up 2-4 times daily, at random intervals.

    Here's my system specs, maybe you can make some sense of this:

    OS1: Windows 2000 Pro SP4
    OS2: Redhat Linux 9.0
    PSU: 500-watt (550max) UL-rated
    MB: ABIT KG7 RAID
    CPU: AMD Athlon XP 2000+ (Non-Overclocked)
    RAM: 512mb PC2100 Kingston ValueRAM (Default Clocks)
    Video: ATI Radeon 128mb 8500LE
    HDD: Western Digital Caviar 80gb w/8mb cache
    Sound: SB Live!
    NIC: D-Link DFE 530TX+

    I thought at first it might be a bad video card, as lockups almost always occur after a while of 3D Gaming. However, I recently ran ATI's tech demos overnight with no lockups. Just to be certain though, I tried a different video card.... still no love. :( Now I'm leaning towards a possible failing power supply. If anyone has a clue, please share!
     
  2. harrack52

    harrack52 Supreme Geek

    Likes Received:
    5
    Trophy Points:
    0
    did you test the RAM ?

    Faulty ram can cause weird problems. Download memtest86 (www.memtest86.com) and see if you get any errors. If you do, it's time to change the ram.
     
  3. Big B

    Big B HWF Godfather

    Likes Received:
    145
    Trophy Points:
    63
    That, and also check your event logs to see if anything crops up.
     
  4. Anti-Trend

    Anti-Trend Nonconformist Geek

    Likes Received:
    118
    Trophy Points:
    63
    Not a (Purely) RAM Issue

    Wow, great forum. :D Anyway, I've already run 2 different RAM tests (results: OK), but on harrack52's recommendation, I tried out that Memtest 'nux kernel... It freezes up at the exact same point every time I run it, leading me to believe that the RAM may well be bad after all. To verify this theory, I tried running the system with a different, fully-functioning stick, and I still get the same lockups in W2k. Doh! On a hunch, I tested out my PSU with a multimeter, and lo-and-behold, the 12v+ is about .78v higher than it should be. My current hypothesis is that the erratic voltage toasted my RAM, and I need to replace the power supply (and ship the memory back to Kingston). Any further input?

    P.S. - the event logs were the very first thing I checked, no help there, but thanks anyway Big B :(
     
  5. harrack52

    harrack52 Supreme Geek

    Likes Received:
    5
    Trophy Points:
    0
    Maybe try running Prime95, and if it hangs at the exact same spot every single time you run it, then it's another vote for the ram.

    One thing you can do to test the psu is stripping down the system to its bare minimum (but with one working stick of ram, since we don't know yet). If you have no problems, then you know it's the psu.

    EDIT: Oh and one thing I always do when I suspect something is I always load the default bios options.
     
  6. Anti-Trend

    Anti-Trend Nonconformist Geek

    Likes Received:
    118
    Trophy Points:
    63
    Mixed Problems?

    I ran a good long torture test with Prime95, and gave it all the RAM I could... No lockups. I decided to test to make certain that the PSU wasn't shorting the RAM any voltage, so I unpluged all devices except 1 CD drive and ran Memtest a few more times. This didn't help at all, as it still locks up on certain tests, namely any "cached" memory tests. I think the PSU is giving too much power, not too little. So far my results are jiving with that theory, and I think the RAM may be a casualty...
    As far as the BIOS is concerned, early on in my troubleshooting I tried resetting it to factory stock, and various combinations of different settings. Unfortunately it hasn't made any difference in stability.
     
  7. harrack52

    harrack52 Supreme Geek

    Likes Received:
    5
    Trophy Points:
    0
    Then I think what'S left to be done is testing the system with a known good working psu.

    I do not think it's the ram.

    Prime95 will usually fail within the first hour if it's gonna' fail.
     
  8. Anti-Trend

    Anti-Trend Nonconformist Geek

    Likes Received:
    118
    Trophy Points:
    63
    I've got a working, high-grade PSU en route, so I'll know if that does the trick soon. In the mean time, I'll try testing the RAM module in question in a stable PC. I'm curious to see if it still locks up at the same point in Memtest...
     
  9. Anti-Trend

    Anti-Trend Nonconformist Geek

    Likes Received:
    118
    Trophy Points:
    63
    At least the RAM's not bad...

    I just finished testing the questionable RAM in a stable system, and it tested out just fine -- even the tests that consistently failed before. Since the system always crashes during memory tests that use certain cache methods, I think the power supply issue may have caused the demise of the north-bridge (aging AMD 761). Would you agree? It would also explain why hardware overlay typically causes system hangups. I'll know for certain whether the MoBo is bad when I get my hands on that other PSU.

    Thanks for all the help, guys :)
     
  10. Anti-Trend

    Anti-Trend Nonconformist Geek

    Likes Received:
    118
    Trophy Points:
    63
    Problem = Motherboard

    I can now state with certainty that the problem was indeed the motherboard. The northbridge had fried, which explains the bad memory tests and hardware lockups during video. I replaced the MoBo with an ABIT KV7-SATA. Along with the restored system stability, I'm enjoying the faster video, RAM, and HDD throughput this board offers over my old KG7-RAID. Thanks for the advice...
     
  11. harrack52

    harrack52 Supreme Geek

    Likes Received:
    5
    Trophy Points:
    0
    No problem :) Glad you found the problem and fixed it.
     
  12. Mike Rogers

    Mike Rogers Geek Trainee

    Likes Received:
    0
    Trophy Points:
    0
    Mainboard Diagnostics

    I know that you figured the problem but I dunno if this might be of help to others with similar problems.
    I already posted a thread on this
    http://www.hardwareforums.com/showthread.php?s=&threadid=1286
    but; i believe anti-trend could have used a mobo diagnostic application like Fit-It Utilities4 or SiSoft Sandra 2004 or other such diagnostic applications.
    They MAY have picked up the fried chipset problem???
    Kinda worked for me...
    :D
     
  13. Anti-Trend

    Anti-Trend Nonconformist Geek

    Likes Received:
    118
    Trophy Points:
    63
    Thanks for the info, but I had tried that, and the lockups made it impossible to diagnose that way. In the end, I figured it out through a process of elimination. Thanks though.
     
  14. ProcalX

    ProcalX all grown up

    Likes Received:
    26
    Trophy Points:
    48
    stop posting stuff like this! its making me anxious as to go ahead with my system build in the next few weeks :(

    i just know something is going to fail or be incompatible :|
     
  15. Anti-Trend

    Anti-Trend Nonconformist Geek

    Likes Received:
    118
    Trophy Points:
    63
    I wouldn't be worried, if I were you. That machine (it's my wife's BTW) ran 100% stable for a long time. As you can tell from the hardware config, that machine's a little older. It turned out to be the old AMD 761 Northbridge that fried, and AMD doesn't even make chipsets anymore (except a few for 64-bit development). I replaced the mobo, and all the other components survived. So, nothing for you to be worried about, eh? :p
     
  16. harrack52

    harrack52 Supreme Geek

    Likes Received:
    5
    Trophy Points:
    0
    Well, if you wanna' be on the safe side, you could always buy a calculator :p
     
  17. ProcalX

    ProcalX all grown up

    Likes Received:
    26
    Trophy Points:
    48
    ? what ? :confused:
     
  18. Big B

    Big B HWF Godfather

    Likes Received:
    145
    Trophy Points:
    63
    All the power of an Apple I, in your $5 calculator.:p
     
  19. harrack52

    harrack52 Supreme Geek

    Likes Received:
    5
    Trophy Points:
    0
    I'm just messing with you.

    Youy said you didn't wanna' have problems. So I suggested a calculator because I'm sure you won't have any problems if you buy one.

    There's no secret, you WILL have some problems eventually. But you can make clever buying decisions buy choosing the right hardware, which will bring down the risk a lot, and right now you have good hardware, so don't worry.
     

Share This Page