PC shuts down by itself following power surge incident

Discussion in 'Computer Support' started by Jimmy Dean, Mar 24, 2005.

  1. Jimmy Dean

    Jimmy Dean Guest

    There was a storm during which the PC remained connected to mains
    power and phone line.

    The internal modem was damaged and had to be replaced.

    Since then the PC shuts down by itself every now and then.

    My question: Will replacing the powersupply fix this problem, or is it
    more likely to be something else e.g. damaged mainboard

    tia
    jd
     
    Jimmy Dean, Mar 24, 2005
    #1
    1. Advertisements

  2. Jimmy Dean

    dadiOH Guest


    Not likely. Sounds like the poor thing is just frightened to
    death...once bitten...


    --
    dadiOH
    ____________________________

    dadiOH's dandies v3.06...
    ....a help file of info about MP3s, recording from
    LP/cassette and tips & tricks on this and that.
    Get it at http://mysite.verizon.net/xico
     
    dadiOH, Mar 24, 2005
    #2
    1. Advertisements

  3. Jimmy Dean

    pengulin Guest

    Seems the more likely problem is the motherboard, given the path. In through
    phone line, takes out modem, motherboard is next in line. However power
    supply is less expensive, obviously, it might be prudent to replace the
    power supply first, anyway, just to see.

    Is there any specific pattern to the shutdown. Does it completely shut down,
    or does it reboot. It's also possible that the problem could be coincidental
    to the lightning strike. I know this sounds unlikely given the
    circumstances, but a modem will typically take the hit for the rest of the
    system before anything else gets fried (if the source is the phone line).
    The circuits in a way act like a fuse. They are designed to blow quickly in
    the event of a lightning strike.

    One thing that frequently causes system reboots is a processor that
    overheats. This is somewhat common if the processor's heatsink is clogged
    with lint (packed tightly between the heatsink and fan). It might be worth
    opening your system case and removing the fan from the heatsink to check. It
    can be blown out with a can of compressed air.
     
    pengulin, Mar 25, 2005
    #3
  4. Jimmy Dean

    w_tom Guest

    First off, others are making assumptions about electricity
    that are even contrary to second grade science class. A
    destructive transient passes through everything in a circuit.
    Only then does one thing in that circuit fail. The complete
    circuit requirement is 2nd grade science. No complete
    circuit? Then no surge to damage electronics.

    So first we define what that circuit may have been. Another
    again speculates the transient entered on phone line, damaged
    modem, then stops. No possible. No circuit defined. AND
    your phone lines would be surge protected, for free, by the
    telco. But the utility wires most often struck and connecting
    directly into a computer are AC electric. Incoming from wires
    down the street, into household wires and into computer via
    green safety ground wire. Through motherboard, into modem,
    through DAA section of that modem to phone line. Down phone
    line to earth via the telco provided 'whole house' protector.
    Now we have a complete circuit from cloud to earth. This is
    the most common type of modem damage. The most commonly
    damaged part is a PNP transistor for modem's off hook relay -
    a little hint how the modem can be repaired.

    Notice the transient also confronted everything else in
    computer. For example, memory has an incoming path but no
    outgoing path. If true, then memory is not on the list of
    potentially harmed components. Again, first there must be a
    complete circuit.

    Second, important data must be obtained. Immediately, what
    do Device Manager and the Event (system) logs report (assuming
    an NT type OS). IOW don't try to repair everything. Obtain
    this important information. Also a responsible computer
    manufacturer will provide comprehensive diagnostics just for
    this reason. If your computer manufacturer was not so
    responsible, then download those diagnostics from component
    manufacturers or from third party sources. Diagnostics
    confirm hardware integrity without the complications of
    Windows.

    In the meantime, is the power supply compromised? In two
    minutes is a definitive yes or no answer using the multimeter
    and procedures defined here (but reading will take many times
    longer): "Computer doesnt start at all" in
    alt.comp.hardware on 10 Jan 2004 at
    http://tinyurl.com/2t69q and
    "I think my power supply is dead" in alt.comp.hardware on 5
    Feb 2004 at
    http://www.tinyurl.com/2musa

    Article with pictures demonstrates the tools:
    http://techrepublic.com.com/5102-10586-5566528.html

    Provided in those posts are a voltage chart. Measured
    voltages must reside in the upper 3/4 range of those limits.
    In your case, the important numbers need only be taken when
    power is on and again when programs are accessing most all
    peripheral simultaneously - max load on the power supply. If
    those numbers are good, then move on to other suspects such as
    those found by the comprehensive hardware diagnostics.

    Either locate the defective suspect OR confirm each
    component as good. I suspect as you do these tests and review
    historical data, other pertinent trends will become apparent.
    Just don't try to fix things until the problem is either hard
    and reproducible or you have clear and identifying symptoms.
    Further, information from these procedures now makes it
    possible for other more knowledgeable posters to provide
    useful support.

    You asked about the power supply. Post the numbers. Get a
    definitive answer. Currently with information provided, no
    one can do anything but wildly speculate.
     
    w_tom, Mar 25, 2005
    #4
  5. Jimmy Dean

    pengulin Guest

    Oops. I (we) stand corrected. Here I thought I had gotten away with it, all
    those years ago. Who'd of ever thought all those little naps I snuck in
    second grade would have caught up to me and bit me in the ass like this. As
    to the speculative nature of Usenet... Well, you get what you pay for.
     
    pengulin, Mar 25, 2005
    #5
  6. Jimmy Dean

    pengulin Guest

    First off, others are making assumptions about electricity
    OK, I plead guilty. I slept through second grade science, and since have
    acquired minimal formal training in electronic theory. .
    That would be me!
    This must be one of the new star services to which I have yet to subscribe.
    But then I guess it's free, so I probably wouldn't see a line item for it
    on my phone bill.
    Ok, now here's where you lost me. Oh, it certainly sounds plausible. It
    definitely seems like you've got some electrical engineering in your
    background, something that would really put my feeble education to shame. If
    I understand you correctly, and again that could be a stretch for my limited
    cognitive capacity, AC lines are the most commonly hit by lightning, and the
    only possible circuit path to ground starts at the AC line, goes through the
    power supply, motherboard, modem, and the telephone line acts to ground.

    I've dealt with a number of customers with fried modems. (telephone tech
    support). Most of the problems have had to deal with the modem's DAA. Now
    the DAA is either solid state (on most new modems) or electro mechanical
    relay (on older modems) with solid state ring detector. Nearly all of the
    modems that were fried had the solid state DAA. I understand the mechanical
    relay DAA's were hard to fry because there was really no circuit when the
    modem is off hook. There would have to be sufficient charge to jump the gap.
    I'm thinking this is what you meant when you said there was no circuit
    defined. Given that most of the customers I dealt with had newer modems
    (most internal modems newer than 1997 use solid state DAA-especially those
    crappy winmodems) that explains why I personally started to see a lot more
    fried modems as the years progressed.

    Now my own experience tells me, contrary to your theory, that the telephone
    line, not AC lines, are the source from which most lightning strikes will
    impact a computer's hardware. Data I've seen seems to back this up. In fact,
    AC lines may be responsible for as little as 30 percent of all damage
    reported as caused by lightning. However, In my own experience I remember
    customers frequently advising me specifically, in the line of questioning,
    that they had indeed unplugged their power before storms. Now I recognize
    some may have been lying because they knew a lightning hit would void their
    warranty. However, few if any of them denied that the phone line was plugged
    in during the storm in question, and that lends toward credibility. Even if
    only a fraction of the people who told me they disconnected the power during
    a storm were in fact telling the truth, your theory (that it's not possible
    for the telephone line to be the entry point of the hit) just doesn't stand
    up.
    That is correct. We tend to operate on the best guess principle because this
    is usenet, people don't tend to provide all the details, and that's all we
    can do. Get it? Kind of like meatball techery.
     
    pengulin, Mar 25, 2005
    #6
  7. Jimmy Dean

    w_tom Guest

    Too much basic technical teaching is required to reply
    properly to pengulin's post. This is not a discussion about
    teaching basic electrical concept. Why the modem is damaged
    is secondary to the OP.

    Just one reason why a modem's DAA section is so easily
    damaged. Pengulin feels only solid state modems suffer such
    damage since electromechanical modems have isolation. He says:
    First get and learn from datasheets for those DIP relays
    found in modem DAA section. A coil that drives the wiper
    (switch) has an electrical connection to that wiper.
    Connection is listed in that data sheet as a breakdown
    voltage. Once that voltage is exceeded, then solid state
    electronics gets connected to phone line via that relay. The
    electromechanical relays were quite easy to damage as I have
    demonstrated by repairing so many lightning damaged modems -
    including this one (repairing them became a source of free
    modems that don't fail).

    A most common path for surges is AC mains to phone line and
    earth ground. AC mains have no protection and connect a
    destructive surge directly into computer electronics. Damage
    that is even made easier by using an adjacent plug-in
    protector.

    How would pengulin know where that transient came from when
    he does not even have basic electronic knowledge? He just
    knows? Or he assumes the damage came from phone line only
    because the modem's DAA section is damaged. This 'damage from
    phone line' is a myth promoted by so many 'meatball' computer
    experts. If the DAA section were damaged, then it must have
    been from phone line? Using 2nd grade science, I demonstrated
    why that is only wild speculation and typically does not
    happen. They assume surges come like waves in the ocean.
    That even violates how 2nd grade science describes
    electricity. And they don't even know about the phone line
    surge protector.

    Pengulin somehow has confused a most common source of modem
    destructive transients to state:
    No. There are other possible paths for damage. The
    sentence was quite clear; "the most common". But when one
    assumes (without first learning the circuits) that modem
    damage comes from the phone line AND does not even know of
    phone line protectors (installed even in the 1950s), well, it
    is best called junk science reasoning. Surge protectors have
    been installed on phone lines for generations - longer than
    most lurkers have existed. So how many 'meatball' computer
    experts even knew about these protectors when they
    speculated? Mostly none. Pengulin also did not know about
    protectors that have existed long before PCs. Therefore how
    can he really know what damaged modems. He also speculates.
    It is damaged. It was connected to phone line. Therefore a
    surge must have existed AND it must have come from the phone
    line.

    Again, that incoming and outgoing path must first be defined
    before anyone can know what caused the damage. That usually
    means analysis at the IC level. Worse still, an adjacent
    power strip protector can even contribute to damage of a
    powered off computer and modem (internal or external).

    Meanwhile, the OP has procedures to locate his intermittent,
    or to provide those with electrical knowledge the facts
    (numbers) necessary to provide assistance. This sidebar is
    about the first of many points made to solve his intermittent
    failure problem. The OP is also encourages to learn about why
    and to correct the building failure that permitted modem and
    computer damage.
     
    w_tom, Mar 25, 2005
    #7
  8. Jimmy Dean

    pengulin Guest

    I'm not even pretending to possess a fraction of the knowledge you have on
    electrical circuits and electrical theory. Nor am I disputing that telco
    lines have built in surge protection. But there is no way you'll convince me
    that said protection is enough to block a direct strike. Nor will you
    convince me that telephone lines do not act as conduit for lightning
    strikes. One other question? Why would anyone in this day and age take the
    time to repair a modem when you can buy one for under $20. Additionally, I
    did not mean to imply that a hit through the telco line would not cause any
    other damage than to the modem. I can see where some people might have
    gotten that impression from my first post. I simply meant to suggest that in
    some cases the modem would take the brunt of the hit possibly shielding the
    other components from damage. I agree that testing all the components for
    damage would be prudent.
     
    pengulin, Mar 25, 2005
    #8
  9. Jimmy Dean

    Jimmy Dean Guest

    Thanks for everyone's comments.

    In practical terms, it seems better to simply replace the power supply
    and see what happens. Checking CPU cooling is another simple measure
    also worh doing.

    The aforesaid PCI modem was definitely monged. I have seen many like
    it, damaged after storms. Fitting a new modem ($15) fixes the problem
    instantly.

    Is it possible something in Windows has been changed, to cause the
    undesired shutdowns?

    jd
     
    Jimmy Dean, Mar 25, 2005
    #9
  10. Jimmy Dean

    w_tom Guest

    This post answers a peripheral question as to why the modem
    was damaged. The other post discusses the OP's original
    problem.

    Does the town protect from a flood by damming the river? Of
    course not. And yet that is exactly what myth promoted
    protectors do. I never said that a transient is blocked,
    stopped, or absorbed. Myths purveyors make that claim. That
    would be like damming the river. Effective town protection is
    a dike and a deeper riverbed - to divert the flood
    downstream. Effective protection does as Ben Franklin
    demonstrated in 1752. Divert to earth ground. Destructive
    surges found a path to earth ground via church steeples (or in
    your case, electronics). To protect, he placed an 'earthed'
    lightning rod (surge protector) so that the surge would be
    earthed by a non-destructive path. All properly installed
    facilitates suffer direct lightning strikes with no damage.
    They don't block surges. The properly installed protector
    shunts (divert, redirect). Only those ineffective and grossly
    overpriced plug-in protectors hope you believe the myth: a
    protectors will block surges. It is another myth along with
    the one that defines surges as ocean waves.

    Do you really think a one inch protector was going to stop
    what miles of sky could not? Effective protectors don't
    block. It is a difficult concept to change one the propaganda
    has you wishing protectors 'block'. Effective protectors
    'shunt' - and to earth.

    What is the most critical component of a protection
    'system'? Single point earth ground. Protectors are simply
    another type of wire that 'shunts'. Therefore 911 Emergency
    and telephone operators don't remove headsets during storms.
    Commercial radio and cell phone towers suffer direct strikes
    and remain operational. The concept is that well proven for
    generations and that reliable. A solution not provided by
    plug-in protectors. No earth ground means no effective
    protection. No blocking. It is called a 'shunt' mode device.

    Why do I repair anything? Almost no electronics is worth
    repairing to save money. But I do it to learn. I can report
    beyond a doubt that most modem damage comes from AC electric.
    Repair to trace the path of destructive surges.

    But again, you are making erroneous and wild assumptions
    that transients enter as ocean waves. Everything in the
    circuit from cloud to ground suffers the same current flow
    simultaneously. There is no 'shielding'. The modem failed
    first. Somehow the 2nd grade science eludes you. Electricity
    is not a wave. Current occurs virtually simultaneously in
    every part of that circuit. Electricity must have a circuit.
    Only after everything is experiencing the current, then does
    something maybe fail. The modem only shields components when
    a human denies basic science.

    I have been through this before. Myths are so entrenched
    that many still have difficulty. They still associate ocean
    wave with electrical transient. They feel protectors somehow
    absorb or block surges. Myths are more widespread than basic
    science.

    Another concept. At what point does voltage from that
    transient first appear? Near end (where the transient enters
    the circuit) or at the far end? Those using the 'ocean wave'
    assumptions would say voltage is highest first where the
    transient enters. Again wrong. The voltage first appears at
    the farthest end; farthest from where a transient first
    enters. Another example of why electrical transients are not
    like ocean waves. Why knowledge of electromagnetic wave
    theory is important.

    All electronics contain internal protection. Protection
    that assumes the building will earth a destructive transient:
    a 'whole house' protector is installed. If not - as is too
    typical for AC electric - then appliance internal protection
    can be overwhelmed. Typically why modems are so often
    damaged. And again, the protector being only as effective as
    its earth ground - and irrelevant to the OP's current problem.
    This well proven science is valuable to all who lurk. Such
    protection sold even in Home Depot and Lowes - but not in most
    other stores that only sell ineffective and overpriced
    products - such as Sears.
     
    w_tom, Mar 26, 2005
    #10
  11. Jimmy Dean

    w_tom Guest

    Power supply is a least likely suspect. Furthermore,
    swapping the power supply may still not fix a power supply
    'system' failure. Power supply system is more than just a
    power supply. Again, two minutes with the meter is far more
    conclusive, done so much faster, teaches how the system really
    works, AND empowers others who currently cannot tell you
    anything useful. Why? You have not yet provided a useful
    fact or number so that people with knowledge can assist.
    Which again is why two minutes with the meter is ... well the
    Japanese talk about doing thing smarter - not harder. You are
    trying to do things harder- a wild assumption without any
    reason to make that assumption. And then swapping parts which
    is so painfully more difficult.

    BTW, many power supplies are missing essential functions.
    Therefore you could swap the power supply, and still have
    strange problems. However that is another discussion for
    another post. The solution with more words is actually the
    easiest and fastest solution. And again, power supply is a
    less likely suspect. Yes, you can keep swapping parts until
    something works. Much more expensive as well as more
    difficult.

    Start with the multimeter, two minutes, and that above
    procedure. End up being smarter and fixing the system faster.
     
    w_tom, Mar 26, 2005
    #11
  12. Jimmy Dean

    Guest

    Jimmy Dean wrote on 25-03-05 15:36 :
    [...]
    yes, if you have spare cash and a desperate need to spend it somewhere,
    and if you don't care whether that will fix things. or you could try
    something simpler:

    do you have a lot of other components in this box? if so, remove
    hardware that isn't needed for normal boot-up -- the dead modem, for
    one; sound cards, yadda yadda... leave only enough to boot up and run.
    if you still get a hard reset, then my guess is something got fried on
    the mainboard. if 'no hard reset', replace components one at a time
    until it happens again.

    [...]
    you're asking whether an ESD event could somehow modify, say, a
    registry key, or change a *.dll or re-program an executable in a way
    that could lead to a hard reset, while not affecting other behavior?

    i would put that at about the same probability as, through brownian
    motion, all the molecules in your monitor move in the same direction at
    the same time, and your monitor jump off the table, unassisted.
     
    , Mar 26, 2005
    #12
  13. Jimmy Dean

    Guest

    w_tom wrote on 25-03-05 18:05 :
    [...]
    wow... whutta badly verbose, misinformed, tedious load of meaningless,
    irrelevant, pseudo-technical crap. there may be one or two useful items
    in this, but i don't know how anyone will find them. and that makes it
    kind of impressive, in its own way.
     
    , Mar 26, 2005
    #13
  14. Jimmy Dean

    Dave Lear Guest

    Brownian Motion is...

    The random movement of microscopic particles suspended
    in a liquid or gas, caused by collisions with molecules of the
    surrounding medium.

    As you appear to be such an expert on statistical probability, what do you
    think the chances are that the original poster has a monitor that is both
    microscopic and suspended in a fluid?
     
    Dave Lear, Mar 26, 2005
    #14
  15. Jimmy Dean

    w_tom Guest

    Lurkers take note useful facts posted by naysayers. None.
    In the meantime, the verbose is also what professionals have
    been doing for generations to not suffer damage from direct
    lightning strikes. Whether it be IEEE papers, those legendary
    application notes from Polyphaser, or even discussion about
    effective protection from the National Institute of Standards
    and Technology: effective protection is always about earthing
    before a transient can get to appliances. Effective
    protection is sold by in better stores such as Home Depot
    (Intermatic products) and Lowes (Cutler Hammer and GE
    products). The naive will recommend spending tens of times
    more money on hyped protectors sold in K-mart, Staples, the
    grocery store, Circuit City, Walmart, Office Max, etc. And
    the naive don't post numbers. In this latest example, a
    naysayer posts insults; not even saying which part he does or
    does not agree with.

    Provided are sources of credibility; and more important,
    the numbers. Naysayers instead post attacks without technical
    fact - nor numbers. A naysayers posts, what is by definition,
    junk science reasoning. In this case, cited as proof is the
    word "wow". In the meantime, another industry professional
    demonstrates in "The Need for Coordinated Protection" how
    destructive transient enter even via underground wires AND how
    damage is eliminated - as was done even 50+ years ago:

    http://www.erico.com/public/library/fep/technotes/tncr002.pdf
     
    w_tom, Mar 26, 2005
    #15
  16. Jimmy Dean

    Top Guest

    Question for you somewhat related.

    How would you protect from the EMP after a nuclear blast? Do
    you think it can be done?

    Top
     
    Top, Mar 27, 2005
    #16
  17. Jimmy Dean

    Jimmy Dean Guest

    I am not talking about a direct lightning strike - I do not believe
    anything really protects from that - rather erratic behaviour in the
    general power supply which seems to be associated with storm activity.
    The latter may cause a power surge which will damage unprotected
    equipment. On the other hand, equipment which is connected to mains by
    a quality surge protector (there are many cheapies not worth having)
    will protect against such.

    jd
     
    Jimmy Dean, Mar 27, 2005
    #17
  18. Jimmy Dean

    pengulin Guest

    Guilty as charged. I didn't even realize I used the word "block" until I
    read back through my comment. But this just brings another question to mind.
    Are there no shunts in place to similarly "divert" a strike away from AC
    lines?
     
    pengulin, Mar 27, 2005
    #18
  19. Jimmy Dean

    Guest

    Dave Lear wrote on 26-03-05 10:07 :
    [...]
    i'd guess it's on-par with your ability to read and interpret, ya daft
    bint. why do you ask?
     
    , Mar 27, 2005
    #19
  20. Jimmy Dean

    w_tom Guest

    If a power supply contains minimally essential functions
    that were required even 30 years ago, then where is the entry
    point for surge damage? Over-voltage protection is required
    on power supplies (but is so often missing in clone systems
    built only on price).

    Again, you can swap power supplies and still not know if the
    problem is solve. But the procedure that firmly says 'yes or
    no' has been provided above. You should be asking for the
    best prices on multimeter - and not speculating suspects. Two
    minutes to determine what is good or bad. That is less time
    than it takes you to repost wild speculation that the power
    supply is defective. You don't know until you first take two
    minutes to get numnbers.
     
    w_tom, Mar 27, 2005
    #20
    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.