1. This forum section is a read-only archive which contains old newsgroup posts. If you wish to post a query, please do so in one of our main forum sections (here). This way you will get a faster, better response from the members on Motherboard Point.

BFG Tech GeForce 7950 GTOC - PC Shutting Off - Temperature Problem?

Discussion in 'Nvidia' started by Damaeus, Nov 11, 2012.

  1. Damaeus

    Paul Guest

    It could be a portion of the VCore regulator is hiding under there.

    And using the heatpipe, the fans are also cooling the chipset (indirectly).

    Paul, Nov 15, 2012
    1. Advertisements

  2. Damaeus

    Damaeus Guest

    I had been reading online that ThermTrip doesn't work through the chipset,
    but only the CPU. Yet the CPU is just not getting all that hot compared
    to the temperature it must reach to start causing problems. I wonder if
    some motherboards actually do switch off if the chipset gets too hot. It
    does seem like it's important to keep that thing cool with a heatsink like

    The fans would be the least expensive thing to replace. I've read that
    they can be replaced, but it looks like I'm going to have to actually
    remove the motherboard from the case to do it.

    I know I should replace that fan since it's out, but do you think that bad
    fan could be causing the problem here?

    For now, I have lowered the temperature in this room until I'm shivering,
    hoping to keep the computer up at least for longer periods of time. I
    used it for several hours doing light-duty things, then my friend came on
    and kept sending me link after link to view on the web. After about an
    hour of that, I had a switch-off.

    All the regular case fans have been out in my case for a few years and for
    that time, I've just kept the side of the case off and I have a big fan
    next to it blowing on medium speed. That gives it far more airflow than
    they'd get from the case fans, anyway, and it's worked fine like this for
    all this time until the last week and a half. So I really don't think,
    myself, that this could be a problem. Even when the electricity has gone
    out and I run on the battery backup (the big fan is not connected to the
    battery), I've not have power-off issues.

    Brrrzzzzrrzrzrz... It's COLD in here! LOL It's in the 30s outside and I
    have my window unit running on 65 degrees and the central air on 70
    degrees. I didn't know it, but my roommate left the central heat on 74
    degrees earlier and it got a bit warmer than usual, then the PC did start
    switching itself off. I'm hoping this cooler air will help until I can
    get a fan. Meanwhile, I'll just try to avoid doing anything that might be
    too intense.

    Earlier it shut down after I'd been using Google Street View for a while,
    but generally, it remains on and running all night long with no programs
    running aside from system tray things like the virus scanner, firewall,
    monitor display management, Windows Update. I may turn off the indexing
    for this Windows Search 4.0 I just recently got when getting all those
    updates, and I'll re-enable it when I get the new fans, assuming I don't
    mess something up trying to put them on the motherboard. I'm not looking
    forward to this because I could cause more damage than I'm trying to fix.

    Damaeus, Nov 15, 2012
    1. Advertisements

  3. Damaeus

    Damaeus Guest

    Here's a scan of the specification chart that came with the power supply.
    I don't know why it says SP-850M. The "pretty painting" on the side of it
    says "Smart M850W" and before I installed it, I wrote down what was on the
    white label, which was "SP-850AH3CCB 850w". lol


    It looks like the 12v rail is rated for 70 amps.

    Damaeus, Nov 15, 2012
  4. Damaeus

    Paul Guest

    A motherboard designer is free to include thermal protection
    for anything they want. I report the "common" implementation,
    which is THERMTRIP, and it's virtually free in terms of
    hardware cost. Adding protection for other things, may involve
    adding some logic. It can even be implemented with transistors
    (Asus has done that when they needed to add logic functions).

    The disadvantage of adding too many protection features, is
    exactly what's happening to you. You can make a PC "too twitchy"
    by putting thermal monitoring on everything. For uncommon implementations,
    it helps if a LED is present, so you know what's going on.

    Again, using an Asus idea from a while back, they put in AGP
    slot protection, to protect 1.5V computers from 3.3V
    AGP video cards. They included a red LED next to the slot.
    If your computer would not start, you'd take the side off and
    notice the red LED was illuminated. That's an intelligent way
    to design an uncommon feature, so that users know what happened.
    Just silently turning off the computer, isn't nearly as


    Fans come in standard sizes, at least the square ones in brushless
    DC form. I have a local electronics store, where I could expect
    to find 40mm fans. You can see here, a couple thicknesses for
    40mm fans. A thicker fan moves more CFM (cubic feet per minute).
    The 10mm thick ones, are the kind you find in disk drive enclosures.
    Thicker ones, are in things like your motherboard, or in rackmount
    server computers. The small fans spin so fast, they tend to whine
    and be annoying, which is why computer cases tend to use larger ones.


    The fans on video cards are more irregular, and you're less likely
    to find an exact replacement when you need one. Replacing the
    whole cooler assembly is the answer there.

    Measure your fan, before you go shopping for a new one. The thickness
    is a variable, whereas the outer (square) dimension tends to be
    one of many standards. You also want to determine whether screw holes
    are needed or not, for mounting. And whether the fan is to have a two
    pin or three pin connector on the end.

    The fans these guys sell, just have loose wires on the end. The info
    here, says these fans use "locked rotor" for the third signal. If
    using this in a computer, you use two of three wires (red and black for
    +12V and GND), and tape up and don't use the yellow wire. The third
    wire can be "2PP" or two pulse per rotation, an RPM signal which is
    useful if the motherboard header has a pin for it. The "locked
    rotor" option is preferred on instrumentation, as a fan fail indicator,
    but a motherboard usually isn't prepared to deal with that. If you have
    a three pin fan header for that 40mm fan, you might at least connect
    up +12V and GND. And that's assuming they used a 12V fan. If they were
    mean, they could have used a 5V fan. Just hope enough details are
    printed on the fan hub, to guide you to a successful replacement.


    OK, this looks more like a computer-related one. This would be 2pp RPM
    on the yellow wire. It's only 5.5CFM. It has the standard computer
    connector on the end, so you don't need to add your own connector
    like I do with the Circuittest ones.


    This one is 15.8CFM and 9500 RPM. A screamer. You use fans like
    this in rackmount servers, where users don't sit next to the
    computer(s) for very long.


    So there is some variation in cooling capability. And you have
    to observe the existing fan, to know whether it's a wimpy (5.5)
    or a screamer (16) CFM fan. The screamer has an audio rating of
    41.9 dBA, and anything over 30 dBA you can hear. And because
    of the piercing "tone" those small fans make, even 30dBA is
    quite evident.

    Paul, Nov 15, 2012
  5. Damaeus

    Buffalo Guest

    Have you tried directing a strong air flow over that failed chipset heatsink
    with the failed fan.?
    Hopefully you get the new fans installed and things work out. Some of those
    chipset heatsinks (if that's what it is, are glued on with a material that
    doesn't always conduct heat too well) need to be pried off , cleaned up and
    reglued on with good stuff like Artic Silver stuff. There are methods to do
    that pretty safely on the Internet.
    Best of luck.
    Buffalo, Nov 17, 2012
  6. Damaeus

    Damaeus Guest

    For a long time, I've had the side off the case with a fan blowing inside
    onto the hardware. At first I just used a small desktop fan with no
    problem, but that fan died so I put a bigger room-cooling fan next to it.
    It moves more air and I had no problems with that, either. I still think
    it's this fan next to the chipset heatsink.

    I put my window A/C unit on 65 degrees and set the central air on 66
    degrees. I was able to keep the PC up and running for over 24 hours.
    Daytime set in, and you know it's harder to keep the house cool in the
    daytime. My PC shut down sometime late Sunday morning when I ran off and
    left a browser game running.

    Yes, I definitely think this is a heat issue.

    I'm wondering about what speed I should use on the big fan blowing into
    the case. I'm running on medium right now, while I had been running it on
    low until running into this. I wonder if I should run the fan on high
    until I get the little fan replaced. One of the fans sounds like it has
    its bearings going out. The fan doesn't make noise all the time, but I
    guess when the fan speed is increased or decreased, that's when I hear it,
    so it comes and goes. I wondered about any kind of static electricity
    from the big fan outside the case. The fan is as big as the case. I have
    it blowing in at an angle, but that probably blows against the airflow of
    the actual chipset fan that's working -- not that it matters too much. The
    top fan only blows air through the top couple of fins on the chipset's

    Damaeus, Nov 18, 2012
  7. Damaeus

    Paul Guest

    Have you ordered your replacement fans yet ?

    When you remove the fans, also visually inspect that the heatsink
    those fans blow on, is making good contact with the source of
    the heat. They could use a sil-pad between the heatsink and components,
    plus a couple plastic push-pins to hold down the heatsink. You want
    to verify the heatsink is still secure, and making good thermal
    contact. I would only really be concerned, if it fell off
    in your hand.

    At least on some motherboards, users discovered the joint between
    a chipset-style heatsink and the thing underneath, was "dry". And
    there was no thermal compound or sil-pad used. In which case,
    the user can make an "improvement" to it, as long as no active
    electrical components will end up with thermal compound on them.
    So if correcting a manufacturing mistake like that, make sure
    you won't accidentally be affecting something, that shouldn't
    have paste on it.

    (Example of a paste that can fill a small gap. If a large
    gap is evident, this wouldn't do the job. It could ooze out.
    Large gaps require other kinds of solutions, such as a
    silicon rubber that cures in place.)


    It's just as likely the heatsink isn't fitting properly, as it
    is that the loss of one fan is causing it to overheat.

    Paul, Nov 18, 2012
  8. Damaeus

    Damaeus Guest

    No, I'm going to go buy them at a local computer store if they have them,
    after I procure funding. I'm doing some work for someone and I should be
    getting something to work with soon.
    There are four pushpins+IBQ-two on the top, and two on the bottom on each side
    of the "heat pipe".

    (270 KB)

    I have never removed that heatsink, nor have I even touched it, and I hope
    I don't have to take it off to get the fan off. The head of the screw,
    unfortunately, is between the heatsink and the fan housing. But there's a
    snap-on panel on the back+IBQ-the one sent by Abit to match the holes to all
    the plugs and whatnot on the motherboard so they can be used. I'm hoping
    that pulling that panel off will reveal an easy way to change the fans
    without having to actually take the motherboard out of the case. If I
    have to turn the screws so close the heatsink, I can't see how I can get
    around taking that heatsink off. And then, if something is between the
    chip and the heatsink, there's the problem of cleaning the heatsink, and
    trying to clean the chip underneath without messing it up.

    (201 KB)

    As it is now, the heatsink should be secure.

    By the way, those pictures amaze me because I snapped them while the PC
    was running. There is no blur on the CPU fan, but it was spinning when I
    took the picture.
    Well, the one fan that's out is the fan that blows through the biggest
    part of the heatsink. If it was the top fan, I'd be more concerned. Plus,
    this computer has never behaved like this before. It's always been really

    Here's what the fan and PC look like as a pair. lol

    (96 KB)

    That setup with the big fan running on high and a temperature in the room
    of about 67 degrees, I've seen lower temperature readings across the board
    in HWiNFO32, and it has not shut down on me with these lower temperatures.

    It took some time, but I'm glad to finally be getting to the root of it.

    I actually had HWiNFO32 logging the readings when my PC did shut down
    yesterday. The last log in the entry was the last thing my computer
    remembered. It's a CSV file in a ZIP. The ZIP is 22k, if you're
    interested in seeing. I think that was when I ran off and left a browser
    game running, specifically Total Domination: Nuclear Strategy. There was
    a lot of CPU activity going on when it powered down.


    Oh oh oh! I just noticed something in that log. The last entry for the
    core clock shows a sudden jump from a normal 2210.2 to 2349.3! What on
    Earth would cause my core clock to suddenly jump? That's when it stopped
    logging because the PC shut down!

    Damaeus, Nov 18, 2012
  9. Damaeus

    Damaeus Guest

    In Damaeus
    I must have gotten excited for nothing. HWiNFO32 shows that my core clock
    has been as high as 3142.1 MHz on Core #0, and 3548.5 MHz on Core #1.


    PS - It's so cold in here I can hardly type, but the PC loves it.
    Damaeus, Nov 18, 2012
  10. Damaeus

    Paul Guest

    In the Newegg photo, I can't see any screws in those fans.
    I do see a white material underneath the fans, like they
    were held in place with double-sided tape.


    ( from http://www.newegg.com/Product/Product.aspx?Item=N82E16813127237 )

    You're going to be pulling that board, to do the repair.

    If that is double-sided tape they've used, you'll need a
    bit of that as well. So the repair will actually be an
    ugly job, not a nice neat one.

    You don't want to use any material you can't get off again,
    such as two component epoxy. That would be a mistake (unless
    your plan is to replace the motherboard/CPU/RAM when the next
    failure occurs).

    And no matter how you do it, the clearances there are rather tight.
    You have to make sure the fans don't jam on the backplate, or
    on the heatsink on the other side of the fan.

    You would have thought they'd have made the aluminum heatsink,
    as a retainer for the fans. That would have been a logical
    way to do it. And making heatsinks is easy when
    they're extruded. And the companies that make heatsinks, they
    can make any shape you want in cross-section. But things
    you'd want to avoid, would be post-machining, such as tapping
    the holes and putting nice threads in them. That's why, when
    fans are fitted in other places, they use self-tapping
    screws. But screws on an assembly line are a no-no,
    as most of the other steps in manufacturing are automated.
    Only the two minute final test, is manual. And that's
    a pretty expensive operation, if you make four million
    motherboards a month. Using tape, as it appears they've
    done, avoids metal filings from self-tapping screws, and
    it's a bit cleaner. It is still time consuming, and I don't
    think any automatic pick-n-place machine can "do tape".

    Um, Good Luck,

    Paul, Nov 18, 2012
  11. Damaeus

    Damaeus Guest

    Maybe I'll just take the motherboard out and take it to the computer store
    to let them install new fans. But at the computer store I use, I once
    asked the guy if he'd put the processor on the motherboard with some
    arctic silver for me. He pulled out a screwdriver to scrape the thermal
    pad off the heatsink and that's when I told him I'd just do it myself. I
    didn't see how he could use a screwdriver on a heatsink without scarring
    the surface. That was the Abit KT7A-RAID I built back around 2000 or so.

    Haha... Thanks. It'd be a shame to say this motherboard is "dead" just
    because of a little $5.00 fan.

    Damaeus, Nov 19, 2012
  12. Damaeus

    Damaeus Guest

    In Damaeus
    Well, I took out the motherboard and did manage to remove the two
    heatsinks and the "heat bar" to get to the fan, which I did remove. But
    before I even got that far, I noticed one capacitor by the PCIe slot that
    was bulging a little and leaking some black stuff. Just to make sure, I
    took the board to a computer repair shop and he verified that the
    capacitor was bad. I asked if he could solder on a new one; he said it
    could be done, but there would be no guarantee that the result would work.
    So I re-installed an older Abit NF7 motherboard that I was using before.

    I'm now starting a new thread with a different kind of problem: The
    GeForce FX 5600 XT graphics card I'm using shows on the web that it will
    pump out a 1920 x 1080 display, but the options panel doesn't have a
    setting higher than 1600 x 1024.

    Damaeus, Nov 23, 2012
  13. Damaeus

    Paul Guest

    On digital outputs such as DVI or HDMI, some early cards
    didn't fully meet spec on clock speed. The clock can go
    up to 165MHz on a compliant first-generation port. Some
    of the "defective" designs, only meet eye opening at

    In this picture, the DVI is only compliant at 141MHz.
    And by the way, 141MHz is just the clock signal - there
    are ten data bits serially, per clock bit, so the data
    stream is 1410 Mbit/sec in that case. It's actually
    pretty fast. (It's non-compliant,when the "colored" part
    touches the deep blue "fixed template", on a normalized


    In order to hide their shame, the driver writers put in
    resolution restrictions, so you won't see the side-effects
    of non-compliant DVI or HDMI ports. In one case, the driver
    writer even did the math wrong, and blocked something like
    1440x900 as well. By preventing users from selecting high
    resolutions on digital output ports, it prevents the
    users from seeing "colored snow" mixed with their image info.

    Now, a VGA port, the driver writer shouldn't have coded
    a restriction in there. And 1920 or higher could well be
    possible in that case. Of course, not all monitors have VGA
    as an option, so there's no workaround if you don't have
    VGA. You could perhaps, buy a VGA to DVI converter, but for
    the price, you could also buy a new video card. Preferably,
    a card at least one generation later than the "defective" ones.

    Paul, Nov 23, 2012
  14. Damaeus

    Damaeus Guest

    This monitor has an input that LOOKS like a VGA input, but it's labeled
    "D-SUB". And really, man, this PC is S-L-O-W. It amazes me now that I
    could ever have enjoyed using it before. The graphics card only has 128
    megabytes of video memory. I can't wait to get some new hardware, but
    having this is better than having nothing at all...sort of. It won't even
    play Farmville 2 worth a diddily damn.

    The fan on the graphics card made a horrible, wavering whining noise. I
    checked and it was spinning, but not very fast. I could actually make out
    the fan blades spinning, but then it speeded up and isn't making noise
    anymore. If I try to run at 1920 x 1080 through the VGA port, I might
    just finish this card off. I only have a couple of other video cards. One
    is a GeForce FX 5200, and the other is a GeForce 2 Ultra. Unfortunately I
    gave my old GeForce 6800GT to a friend, thinking I'd never need it again.
    How wrong I was! I'd LOVE to have that card back right now because it
    would run fine in this system. It was what I was using before I built the
    one that died. :(

    Web research shows this monitor does have a VGA input, so I guess that's
    what the "D-SUB" actually is. I don't know why it's labeled that way. I'm
    tired enough of this non-preferred resolution and weird appearance of
    fonts that I'm willing to try VGA to see if I can get there without frying
    the video card. It's supposed to do up to 2550 x something at 100 Hz,
    which is pretty amazing, but I only need 60 Hz for this monitor. Maybe
    that'll help keep the burners down.

    Damaeus, Nov 23, 2012
  15. Damaeus

    Damaeus Guest

    Okay! I'm now running with the VGA cable and 1920 x 1080! It looks a lot
    better like this than it did running subresolution with the DVI cable. I'm
    happy with this until my video card fries or I get a new rig.

    Thanks for all the help. You've been very patient. :)

    I doubt this thing will run Final Fantasy XI at 1920 x 1080 at a decent
    frame rate, but the game itself is limited to 30FPS, so maybe it'll be
    okay. In areas of high density traffic from other players, even with my
    GeForce 7950 GTOC, the frame rate would sometimes drop to 5FPS, but it's
    still fun.

    Damaeus, Nov 23, 2012
  16. Damaeus

    Paul Guest

    There are some suggestions here. I take it one of the suggestions, moves
    some registry stuff out of the way.


    But what I'd be searching for is "Hardware Acceleration" settings.

    Adobe Flash has a hardware acceleration setting. There is a dialog
    that opens in the middle of a Flash movie window, that includes a
    tick box for hardware acceleration.

    Browsers have also taken to hardware acceleration.

    But with your hardware (and mine), our cards don't have a lot to
    offer. Our programmable shaders standard is probably too old for
    that sort of thing. Video cards have had 2D acceleration (bitblt)
    for a long time. I don't know if they bother with that or not,
    as it's hardly worth it. Video cards have had IDCT for some
    forms of movie decoding, but that isn't a "general purpose" form
    of acceleration. I can't imagine what construct would be dying
    with the hardware acceleration.

    When the browser session starts, you want to tell it to *not* load
    the last windows. Where you were surfing, the source code
    tipped over your browser. So you'd want to start a fresh session,
    and see if the browser will stay up with just a blank page
    in view. If your home page is "Youtube", then that might immediately
    fire up Adobe Flash, and test Flash hardware acceleration.

    Maybe the browser has a "Safe Mode" or load without plugins ?
    At least for one session, before you turn it all back on again
    and re-test.

    Anyway, that's where I'd head. At least, until you indicate that
    something unrelated to video seems to be crashing. If Wordpad
    was crashing, you'd be looking at CPU or RAM perhaps, as much as
    anything else. But with browsers, there are a few forms of
    hardware acceleration. Either in plugins, or in the main
    body of the browser code. Older browser code, doesn't use hardware
    acceleration, and in a way, you can blame the work Adobe Flash did,
    for attracting others to take that route. The thing is, the code
    they write for that, doesn't seem to be quite the same quality
    as code for 3D games. It's almost like they don't have a big
    enough testing group, to make sure it all works.

    Paul, Nov 24, 2012
  17. Damaeus

    Damaeus Guest

    After I let Chrome rest for a while, I was magically able to relaunch it
    normally. So far it hasn't happened again, but since that was the first
    time I ever saw that message (and the first time I had ever used this
    monitor with a VGA cable), of course, all my flags flew out.

    Thanks again!

    Damaeus, Nov 24, 2012
    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.