Self-built computer -Black screen
Home | Reviews and Features | Special Reports | Forums |

Page 1 of 4 123 ... LastLast
Results 1 to 15 of 46

Thread: Self-built computer -Black screen

  1. #1
    Join Date
    Jul 2007
    Location
    NYC
    Posts
    55

    Self-built computer -Black screen

    Built this from a bare-bone kit:

    Mobo XFX MG-63MI-7109
    Intel Pentium D 830 (Smithfield), Dual-core, 3.0GHz / 2MB Cache Socket 775, 800 FSB Processor with Fan
    NVIDIA nForce 630i/7100 Audio, Video, both onboard
    MicroATX case
    450-Watt Power Supply -Generic (Oops!)
    Crucial 2048MB PC5400 DDR2 667MHz Memory

    AMI BIOS
    Manual for mobo, but not BIOS settings

    New Seagate SATA 80 GB HD, my own CD-RW
    Win XP Pro SP2

    My usage: no gaming, no video editing, nothing stressful on this system.

    ---------------------------------------------------------------------
    Be gentle. I had old slow computers and longstanding financial problem, and this was the best I could afford. This was the first one I've done without help, but I'm experienced in parts replacement, system repair.

    After about 2 months, upon bootup, I would get the first black screen enumerating the drives and memory, then the screen went black with a flashing white dot in the upper left corner. Could not reboot normally.

    The BIOS produced these beeps: 1 short, 2 short, 1 short 2 short. Not definitive on the AMI beep list, but 3 short = "Base 64K memory failure A memory failure has occurred in the first 64K of RAM. The RAM IC is probably bad " http://www.computerhope.com/beep.htm#04 WHAT"S RAM IC?

    Searching the net produced no other definitive results. XFX took back the mobo, tested it, said it was defective, said probably the bad PSU had shot it. The techie said he's seen it many times.

    Used an old faithful PSU ( PSU #2) with mobo #2, but got the same results. XFX said mobo was NOT damaged, but replaced it anyway to give me a fresh start. Mobo #3 was thoroughly tested before shipping.

    Here's where I am now: Mobo #3, I replaced PSU#2 with another old faithful, (PSU #3) 430W, fully functional. Some bootups produce the black screen, some don't. PSU #2
    and #3 work fine in my old comp. . I no longer get the 3 beeps.

    I have no technical testing equipment, but I have :Everest Home Edition.

    Temp readings from BIOS after running a while:

    CPU temp-- 152-158
    SYS TEMP-- 98-102
    CPU fan -- about 2616, fluctuates slightly
    SYS FAN -- 0 RPM (not monitored?)

    -----------------------------------------------------------------------------
    I could return the whole thing to TigerDirect (another Oops! And yes, I know they're awful regarding returns.) but I've been without it for 2 months, and I've needed a fast comp for work for several years. If it's something minor, although they're responsible, I 'd rather just fix it myself, considering the initial coast was only $200 with the new HD. Major--it goes back to them. And better if I can pinpoint the problem before I call them.

    *vB tags corrected. - Moderator

  2. #2
    Join Date
    May 2006
    Location
    Arkham Asylum, Cell 13
    Posts
    3,373
    Are you testing your setup OUTSIDE the case on a nonconductive surface with just minimal hardware? Did you check for extra standoffs under the motherboard?

  3. #3
    Join Date
    Jul 2007
    Location
    NYC
    Posts
    55
    There are no extra standoffs. I've assembled the thing 3 times and know every inch.. As far as I see, no circuits are touching the screws. Wouldn't a short circuit have fried things by now? I'm functioning perfectly well after bootup.

    And no, I wasn't aware of being able to test a system outside. Can I set it up on the nonconductive bag high above the standoffs? Otherwise how would I hit the Power button if the setup is outside the case?

    I was just inspecting the IRQ's, and #22 is shared by "NVidia Gforce etc" and "Standard Enhanced PCI to USB Host Controller". Is that OK? Would I see a yellow exclamation point somewhere if I had a conflict?
    Last edited by sneakers; May 6th, 2009 at 05:24 PM.

  4. #4
    Join Date
    May 2006
    Location
    Arkham Asylum, Cell 13
    Posts
    3,373
    I've seen weird things happen with extra standoffs. It's a pretty common mistake.

    Otherwise how would I hit the Power button if the setup is outside the case?
    you manually short the pins for the power switch with a flat head screwdriver or something similar.

    Your CPU temp is 158F? That's like 70C, which is really hot. Are you sure the heatsink is mounted properly?

    Did you try running with just 1 stick of ram? Did you run Memtest86+ on the ram?

  5. #5
    Join Date
    Jul 2007
    Location
    NYC
    Posts
    55
    It runs at about 158 degrees after 15 minutes and stays there. I mounted the heatsink properly, outside the case, because of the problematic plastic pins. But would a hot CPU cause problematic cold bootups? Wouldn't it be more likely to shut the comp down while running? It's starting top sound like RAM.

    I didn't run Memtest or try 1 stick of RAM yet. I thought the first 64K of RAM was on the mobo. I'll get back to you on those.

  6. #6
    Join Date
    May 2006
    Location
    Arkham Asylum, Cell 13
    Posts
    3,373
    The temps might not cause your cold boot problems, but they need to be addressed.

    I would run a RAM diag just to make sure your hardware is good. Don't assume anything.

  7. #7
    Join Date
    Jul 2007
    Location
    NYC
    Posts
    55
    OK, will run tonight. Does it give me a report on the screen?

    About the temp, I might have made an error. I used the same white grease from a local repair place that I had used on my old 750 GHZ computer. Do I need a better grease with the hotter CPU? Also, weirdly, the temp goes from 152 to 158 while I'm staring at the BIOS and Windows ISN'T running.The SYSTEM TEMP does the same, from 98 to 102. . Is this normal?

  8. #8
    Join Date
    May 2006
    Location
    Arkham Asylum, Cell 13
    Posts
    3,373
    The thermal paste should not make THAT much difference. A couple degrees maybe, but not 10+. 158F at idle is really high, even for a Pentium D. 102F is acceptable for the chipset. Did you choose Fahrenheit or was that default? Most of the time it's in Celsius.

    Does it give me a report on the screen?
    Yes, Memtest86+ will have red errors on the bottom of the screen. You should get ZERO errors if the ram is good. If you get any errors, make sure to test each stick of ram individually. I'm still concerned about your CPU temps while running Memtest.

    UBCD has Memtest86+ and a bunch of other tools. Link is in my sig.

  9. #9
    Join Date
    Jun 2000
    Location
    Vancouver, B.C., Canada
    Posts
    8,069
    you manually short the pins for the power switch with a flat head screwdriver or something similar.
    You can also use a touch with a spare jumper. Less chance of shorting against other pins.

    You should follow up on Midknyte's suggestions on checking out the cooling and testing ram. After that if you're still unable to solve the problem and willing to try yourself, I'd be suspicious of the Pentium D. I doubt their longevity. That board can take Core2Duo cpus and the E1200 is an overlooked model. Unloved by being branded a Celeron, it is not crippled in any way, it just has smaller cache. And from all the tests I've seen, the cache only makes a significant difference in gaming. I'm not sure if your board has decent overclocking options, but the 1.6G is only a starting point. You should be able to easily do 2.4G on nearly default voltage, something to consider if you can find it cheaply.

  10. #10
    Join Date
    Jun 2000
    Location
    Vancouver, B.C., Canada
    Posts
    8,069
    After catching up on the recent posts:

    The type of thermal grease won't make a large difference, but using too much might. The very best instructions for applying it are on Arctic Silver's website, and they are now different for different types of cpus. Considering those temps, might be worth checking.

  11. #11
    Join Date
    May 2006
    Location
    Arkham Asylum, Cell 13
    Posts
    3,373
    A Pentium D at 70+C idle will definitely have a shorter lifespan than a Core2.

    http://www.arcticsilver.com/instructions.htm

    Check out the one for AS5.

  12. #12
    Join Date
    Jul 2007
    Location
    NYC
    Posts
    55
    [QUOTE=Midknyte;1038729]The thermal paste should not make THAT much difference. A couple degrees maybe, but not 10+. 158F at idle is really high, even for a Pentium D. 102F is acceptable for the chipset. Did you choose Fahrenheit or was that default? Most of the time it's in Celsius.
    I think I chose Fahrenheit because I'm familiar with it. Or maybe it was the default. Would have to convert Celsius to F anyway.

    Yes, Memtest86+ will have red errors on the bottom of the screen. You should get ZERO errors if the ram is good. If you get any errors, make sure to test each stick of ram individually. I'm still concerned about your CPU temps while running Memtest.

    I tried swapping the memory card in the slots, and tried each alone. Got about an equal number of bad boots . Then I ran Memtest and another one, msinst, and there are no errors. So I called a techie at XFX and asked his opinion. He said the memory test sometimes don't work, and that if the CPU weren't working, I wouldn't get any post.

  13. #13
    Join Date
    May 2006
    Location
    Arkham Asylum, Cell 13
    Posts
    3,373
    Something got messed up with your quote code. You need an open and close tag.

    I knew you were on F since 150C would be 1.5x boiling point. Most of the World uses C.

    Memtest86+ will usually get errors within the first few passes (8 tests per pass), but some people run it overnight to be sure. How many passes did you run? Please specify version of Memtest that you used also. Memtest86 and Memtest86+ are different apps.

    You won't get POST with a dead CPU, but bad CPUs can still post. I've seen some with bad cache that POST.

  14. #14
    Join Date
    Jul 2007
    Location
    NYC
    Posts
    55
    Quote Originally Posted by DanceMan View Post
    You can also use a touch with a spare jumper. Less chance of shorting against other pins.

    You should follow up on Midknyte's suggestions on checking out the cooling and testing ram. After that if you're still unable to solve the problem and willing to try yourself, I'd be suspicious of the Pentium D. I doubt their longevity. That board can take Core2Duo cpus and the E1200 is an overlooked model. Unloved by being branded a Celeron, it is not crippled in any way, it just has smaller cache. And from all the tests I've seen, the cache only makes a significant difference in gaming. I'm not sure if your board has decent overclocking options, but the 1.6G is only a starting point. You should be able to easily do 2.4G on nearly default voltage, something to consider if you can find it cheaply.
    I don't game or overclock, and the 3GHZ CPU is more than enough for me. I bought the kit because it was reasonable. When you say you don't trust the CPU, do you mean after 2 months it could have died? I built this in January. The techie at XFX says it would't have posted at all if the CPU were damaged.

    I'm looking into the cooling problem next. Do you think the fan is running too slowly? The one in this old 750 MHZ runs at over 4000 RPM.

  15. #15
    Join Date
    Jul 2007
    Location
    NYC
    Posts
    55

    Question

    Quote Originally Posted by Midknyte View Post
    Something got messed up with your quote code. You need an open and close tag.

    I knew you were on F since 150C would be 1.5x boiling point. Most of the World uses C.

    Memtest86+ will usually get errors within the first few passes (8 tests per pass), but some people run it overnight to be sure. How many passes did you run? Please specify version of Memtest that you used also. Memtest86 and Memtest86+ are different apps.

    You won't get POST with a dead CPU, but bad CPUs can still post. I've seen some with bad cache that POST.
    ----------------------------------------------------------------------
    Don't know anything about making tags.

    Most of the world uses C, but we use F, and that's what I know. :P

    I ran 30 passes over a period of 13 hours. Don't know the version, but I had made a note on the floppy: 1/2008. The other program was mtinst.exe, don't remember the exact name, from Microsoft, 2007. Are they too old?

    I'm on my old reliable clunker, one of two I built a number of years ago. Slow but reliable , and at least I know all the parts like the back of my hand. Can't get any boot from the new one. I do need to pinpoint this, because the CPU has a 3-year warranty.How about a CPU testing program, but one a non-professional can understand?But it has o run in DOS. Or if need be, I'll submit the report here.
    Last edited by sneakers; May 8th, 2009 at 01:10 AM. Reason: addition

Thread Information

Users Browsing this Thread

There are currently 1 users browsing this thread. (0 members and 1 guests)

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •