PC won't power on

Discussion in 'Computer Support' started by alice, Mar 13, 2007.

  1. alice

    alice Guest

    I'm confused then, about how to make these measurements with the 20-
    something plug plugged into the MB. Where does the probe go?
    alice, Mar 27, 2007
    1. Advertisements

  2. alice

    w_tom Guest

    Follow each wire into holes atop that white nylon connector. Look
    inside. See that metal contact? Push the probe into that hole to
    measure voltage on that wire.

    Again - and this is most important. Never break a connection or
    replace anything. The only thing you need remove is chassis cover.
    Everything else remains connected. Even better is to not even remove
    dust. Change nothing. Simply take measurements according to a
    procedure in "When your computer dies without warning....." starting
    6 Feb 2007 in the newsgroup alt.windows-xp at:
    And then report those numbers here.

    Meter set to 20 Volts DC. Black probe either makes contact with
    black wires on any power supply connector, or attached to metal
    chassis. Red probe touches each color wire as described in that
    procedure. Purple wire measured with computer off and AC power cord
    attached to wall receptacle. Follow the procedure. Voltages also
    taken as front panel power switch is pressed. Nothing is disconnected
    since disconnecting only adds more problematic variables to a solution
    - and wastes time.

    When power switch is pressed, then green wire voltage should drop
    from well above 2 volts to near zero volts (ie 0.1 or 0.2 volts).
    w_tom, Mar 27, 2007
    1. Advertisements

  3. alice

    alice Guest

    I see. Thanks for sticking with me.
    Here are the numbers...of course they all drop to 0 eventually as the
    power quite completely.
    Purple 5.02
    Green 5.01 / .11
    Gray 5.02
    Orange 3.34
    Red 5.4
    Yellow 11.xx (it jumps around alot)
    alice, Mar 27, 2007
  4. alice

    w_tom Guest

    One voltage is too high. Not yet high enough to damage motherboard
    and peripherals. But approaching levels that would cause a power
    supply to shut down - to only restart by powering cycling.

    Voltage on red wire is 5.4 - max must be 5.25. If yellow is also
    dropping sometimes below 11.7, then this implies the CPU power supply
    might have trouble providing stable power to the CPU.

    Question is why 12 volts is so unstable - and therefore cause the
    excessively high 5 volts. Well, this problem started with actions
    back at the CPU heatsink. Adjacent to that heatsink is a CPU power
    supply powered by that 12 volts. Did something happen to that CPU
    adjacent power supply to cause bad 12 volts? Or has main power supply
    simply failed - maybe because it was always marginal (defective but
    not yet causing failure)?

    A visual inspection can now concentrate on the adjacent power supply
    area of power supply chips and capacitors (cylinders standing
    upright). The tops of those capacitors (cylinders) must be perfectly
    flat - aluminum does not bulge above capacitor's plastic covering.
    Nothing should be bent over - no incidental metal contact. And
    nothing should be so hot (when power is on) as to cause pain. Nothing
    in that area will harm human life; touching is safe.

    Knowing where to inspect with scrutiny, and if having seen no
    obvious defects, well, we move on to see what will make that 12
    volts stable and 5 volts normal. We have only two unknowns left.
    Either a 12 volt load on motherboard or the 12 volt supply.

    This test is done only if you are comfortable with removing the CPU
    and heatsink assembly as one. Generally, CPU socket has something (ie
    a lever) that clamps to hundreds of CPU pins. Once that lever is
    'released', then CPU almost falls out. Carefully note the CPU pin
    orientation because CPU must be put back in the exact same position.
    This removal is done with AC power cord disconnected (again so that
    the purple wire 5 volts does not exist).

    With CPU removed, then restore power cord, and power on system (that
    has no CPU). What happens to yellow and red wire voltages? If those
    voltages remain unstable, then something on motherboard (ie CPU
    adjacent power supply) has failed. Of course what those numbers are
    and do will be more informative.

    If not comfortable with removing CPU, then learn what happens with
    voltages from that other power supply. If that other supply also has
    unstable 12 volts (yellow wire) and higher 5 volts (red wire), then
    again, problem is probably in that CPU adjacent power supply.

    From either test, if the CPU adjacent supply is suspect and if
    nothing is bent over 'touching', then motherboard failure is suspect.
    We have a good idea where the failure is AND we cannot fix it - only
    replace it. CPU is most likely just fine.

    Hopefully, that unstable 12 volts is due to the load - not due to a
    marginal and getting worse power supply 12 volts. This problem
    started when you were working around that CPU area - removed the
    heatsink. That is the region that consumes so much 12 volts (yellow
    wire currents).

    It would be nice to know that power supply voltages were good when
    the system was first purchased - that voltages were in spec when
    system was working and accessing all peripherals. Then we knew with
    certainty that the power supply did not have a marginal 12 volt
    output. Unfortunately most computer techs have little appreciation
    for the meter. However we would know far more either with CPU
    removed or with voltage numbers from a second power supply. Removing
    the CPU is a better test because a removed CPU turns off the adjacent
    power supply causes near zero power draw from power supply.

    Looking forward to learning what you disicover. Hopefully it is
    only a bent over component adjacent to the power supply - a least
    expensive solution.
    w_tom, Mar 28, 2007
  5. alice

    Leythos Guest

    Checking the CAPs and other items on the board is one of the first things
    that should have been done, repeat, one of the first things. You always do
    a visual inspection first, it's part of doing a proper diagnosis.

    Second, once you have the visual done and there are no signs, you could
    have saved a LOT of time and wasted days by just purchasing a cheap PSU
    that meets the spec's for the computer needs. More times than not, 90%,
    the PSU will fix the problem and you only need enough time to purchase
    one, not a multimeter that you can't return and that you can't read
    without understanding it first and without learning about PSU
    testing/voltage levels.

    So, as standard, it's been proven to be cheaper and quicker in most cases,
    about 90%, to just buy a cheap PSU that you can even return if wrong, and
    disconnect the drives, then see if the board/post works....

    All of this could have been resolved in a few hours time.
    Leythos, Mar 28, 2007
  6. alice

    w_tom Guest

    Had Leythos bothered to read or learn before posting, then he knew
    shotgunning was tried - and failed. Shotugnning provided no useful
    information and took longer. The OP then did what is advocated even
    in CSI - follow the evidence. As a result, actual problem is known.
    In but two minutes, area of failure has been isolated to but a few
    items. Shotgunning takes many times longer AND discovered nothing.

    But then shotgunning is advocated by those who don't first learn
    technology - did not even first read previous posts. Leythos
    advocates shotgunning because he did not bother to learn how
    shotgunning already failed to find or fix anything.
    w_tom, Mar 28, 2007
  7. alice

    Leythos Guest

    I followed the thread, noticed that shotgunning was not tried, noticed
    that a proper method was not followed, noticed that you keep taking people
    down a path that takes days instead of hours.

    There is nothing you can do with a multimeter and testing voltages that
    can't be done without it in shorter time, with less training, with less

    Why don't you list, for everyone's benefit, the steps one should take, in
    proper order, when trying to diagnose a computer that won't properly power
    on - you could then cut/past it for them so that everyone could benefit
    from clear directions. (I don't actually expect you to list the steps,
    you've never tried and don't seem to be capable of it).

    Want to know what PCBUTTS1 is really about?
    *** WARNING - this links contains foul/pornographic content of an
    abusive nature created by PCBUTTS1 and still hosted on his public
    website ***
    Leythos, Mar 28, 2007
  8. alice

    w_tom Guest

    Had Leythos read before posting, then he would have read this
    Or Leythos could have read this attempted shotgunning:
    After so much labor, then what was learned? Almost nothing. Why
    does Leythos recommend shotgunning? Well, this same person with so
    little knowledge of electricity even claimed fuses protect appliances
    from damage. Then the newsgroup taught Leythos that fuses blow AFTER
    the damage - to protect humans from danger such as fire. One who did
    not even know that would then shotgun: symptoms of insufficient basic

    Leythos can learn by reading. Instead, Leythos did not even read
    this discussion - and magically knew how to fix things. Leythos
    recommends shotgunning even after shotgunning failed. He did not even
    read previous posts.
    w_tom, Mar 29, 2007
  9. alice

    Leythos Guest

    I almost never read an entire thread that you are part of because you take
    the long way around everything and waste soooo much time.

    How come you failed to answer the steps needed for proper diagnostic work?

    How come you never take the quick path?

    How come you've spend days trying to fix something that could have been
    fixed in an hour?

    Want to know what PCBUTTS1 is really about?
    *** WARNING - this links contains foul/pornographic content of an
    abusive nature created by PCBUTTS1 and still hosted on his public
    website ***
    Leythos, Mar 29, 2007
  10. alice

    w_tom Guest

    You never read that shotgunning tried and failed because you are too
    busy posting rather than learning facts - ie how electricity works.

    Apparently Leythos has been caught lynig so often about 'magic
    fuses' and 'invisible surge protectors' that he sits on the internet
    waiting to defend himself - rather than learn basic technology. Each
    post gets a immediate response from him - 20 minutes and faster. He
    must sit on the computer all day long learning from internet myths -
    rather than learn how electricity works from text books. No wonder he
    cannot walk down to Home Depot to see prices. No wonder he has no
    design experience. No wonder his entire knowledge base is from
    internet myths. It explains why he shotguns rather than do as
    advocated on CSI - 'follow the evidence'.
    w_tom, Mar 29, 2007
  11. alice

    Leythos Guest

    You know, you really have no clue, and your constant frabication of BS is
    almost as amazing as the best of trolls I've seen.

    I've said that if you spend days learning how to use a meter, what
    voltages should be present on the PC in the different states, and 99% of
    the time a simple shot-gun method can resolve it a LOT quicker than having
    to find a store to purchase a multimeter, learning how to read it,
    learning what the signal levels should be, etc...

    In 99% of the cases, if you shotgun a PSU it fixes the problem, in the
    rest, if you disconnect the devices you've narrowed it down to a device or
    the motherboard in about 10 minutes - and if it's the motherboard/CPU your
    multimeter can't tell you which one it is.

    So, you saying that I'm wrong with the above?

    Want to know what PCBUTTS1 is really about?
    *** WARNING - this links contains foul/pornographic content of an
    abusive nature created by PCBUTTS1 and still hosted on his public
    website ***
    Leythos, Mar 29, 2007
  12. alice

    lobo Guest

    <BIG SNIP>
    Way too much back and forth going on here about this subject. Bottom
    line.....did she ever actually find out what the problem was?
    lobo, Mar 29, 2007
  13. alice

    alice Guest

    OK, with the CPU removed, the yellow reading is 11.91 and stable, the
    red 5.14 and stable (and the power never turns off).
    There is one component/cylinder right next to the CPU that -might- not
    be perfectly flat on top, but it's hard to tell.
    alice, Apr 3, 2007
  14. alice

    alice Guest

    This time, when I put the heatsink back on, I added some paste. And
    guess what? It now turns on and stays on.
    Now if I can get it to not overheat and shutdown, which was the
    original problem.
    alice, Apr 4, 2007
  15. alice

    Leythos Guest

    And to think, sorry for being sarcastic, the multimeter didn't help one
    bit and could never have indicated a improperly installed heat-sink, but,
    allowing the PC to cool, then checking the BIOS, would have given an
    indication of a temperature problem.

    To properly install HSP (heat sink paste) you need a small razor knife,
    you clean both surfaces with something like alcohol and a very clean
    cloth, something that doesn't leave anything behind. Next, you place a
    very small amount of HSP on the CPU and using a razor knife you spread it
    around - don't do it on the heat sink itself - the HS surface is often
    larger than the mating surface of the CPU.

    Now, with a single motion, install the HS on the CPU, don't push it
    around, don't grind it, don't move it more than needed. If you've done it
    right you should not seen an excess paste - the past is designed to ONLY
    FILL THE SURFACE IMPERFECATIONS - to much past acts as a insulator and
    will still let the CPU overheat.


    Check the above link, pick the CPU that matches your description, check to
    make sure that you installed the HSP properly - it really doesn't matter
    what the CPU looks like (intel or amd) the method for proper HSP is the

    Want to know what PCBUTTS1 is really about?
    *** WARNING - this links contains foul/pornographic content of an
    abusive nature created by PCBUTTS1 and still hosted on his public
    website ***
    Leythos, Apr 4, 2007
  16. alice

    w_tom Guest

    Your numbers completely eliminated the power supply system as
    defective. And those numbers took you to where the problem was.
    Therefore you also did not waste money on a new motherboard, new CPU,
    and who knows what else.

    I don't remember if you posted it as an AMD or Intel CPU - which was
    relevant information. Based upon results, it probably was an AMD.
    However reinstalling of everything may have corrected a mechanical

    BTW, - caution: thermal paste must never ooze out the side if not
    mentioned elsewhere. Too much thermal paste, indicated by oozing, can
    cause future computer problems.

    Is that capacitor starting to bulge? Well, remember what it looks
    like and view it again in a few months. If the bulging has started,
    then increased bulging will be obvious later and probably also in
    other adjacent capacitors. How to identify another problem before it
    creates failures.

    Only if necessary, those capacitors can be replaced for about $1 per
    capacitor by someone with a good soldering iron; if bulging does
    increase and if you don't want to scrap the entire motherboard.

    Good to hear problem solved without spending money on unnecessary
    parts. You now know more about fixing computers than many 'experts'
    in some computer stores - how to 'follow the evidence'. Good luck.
    w_tom, Apr 4, 2007
  17. alice

    alice Guest

    Well, I'm still not gettting very far. I'm able to get to the BIOS
    setup and see that it is not overheating. But when I try to boot up on
    the HD, I just get the windows 'loading your settings' screen, and it
    hangs there forever. The keyboard light is on, but the mouse light is
    not, which I think it usually is a that point. I am able to boot from
    a floppy, but I'm not sure how to get the HD to fully boot at this
    point. Before all this happened, the HD worked just fine.
    alice, Apr 6, 2007
  18. alice

    w_tom Guest

    Some basic hardware is now working. Does software on motherboard
    complete and Windows start to load? Well repeatedly pressing F8 (or
    maybe F5) key will provide a list of options. One should permit each
    file that is loaded to be reported as it loads. Simply look at the
    last file reported - the one that does not complete - to see where
    Windows is failing to load. This should point to that problem

    That assumes BIOS does start a Windows boot loading.

    Another useful fact may be from hardware diagnostic from that disk
    drive manufacturer- loaded for free from their web site. These
    diagnostics sometimes may be cryptic. But the diagnostic will report
    what the CPU sees on disk drive. Are cables connected - and maybe 50
    other possible problems? Do any of these problems exist? That disk
    drive hardware diagnostics would confirm disk drive (and interface)
    integrity so that we can move on to other suspects.

    Additional information are BIOS settings for that disk drive as
    stored in the CMOS. Unfortunately we don't know what those setting
    once were. Hopefully the BIOS is set to automatically select the
    correct disk drive configuration settings such as numbers of
    cylinders, sectors, heads, and the selection for how all those are
    read. At minimum, record and report what those CMOS settings are for
    the hard drive.

    And of course, does the system boot from some other peripheral - CD-
    Rom. Does Windows boot from the CD? If so, repair tools exist on the
    CD. But we don't yet dare start fixing Windows until we know hardware
    is good AND CMOS settings make sense. Otherwise irreparable damage
    may result.

    If computer boots from floppy - what exactly boots? Can you load
    DOS? If so, then a command line program can read disk drive without
    damage. Enter-
    dir c:*.*
    to see files on hard drive. If so, we have another test to confirm
    disk drive hardware is working. Enter:
    dir c:\*.* /s or
    dir c:\*.* /s /p
    to see if all of drive can be read. This same command can be entered
    from Command Prompt if system can boot Windows from CD. Again, see
    what does and does not work before trying to fix anything.

    Numerous ways to 'follow the evidence' - to see why computer is not
    booting before replacing or changing anything.
    w_tom, Apr 6, 2007
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.