Video Goes Bonkers

[tjustin]

Yesterday my computer worked fine. Today I turn it on, and the monitor
is filled with colored lines with semi-readable words embedded in the
colored chaos I am seeing. I should note this occurred when I first
turned on the computer and before booting. Even when I boot to Dos, I
cant see the screen since its all covered with this mass of colors and
lines. No one else has used the computer since I last used it.

The computer is a 700mhz running Windows 98.

I first tightened the video connector, then tried another monitor.
Nothing changed. Both monitors get the same mess on the screen. I
booted to a floppy, again no change. This mobo has a built in video.
I pluged in an old pci video card and now I am getting a normal
picture again.

Did my onboard video fry (I sure cant see how since the computer was
shut off. Or could this be some CMOS setting issue, which would mean
my battery is weak.

Anyone?

TJ

 

[Paul]

Yesterday my computer worked fine. Today I turn it on, and the monitor
is filled with colored lines with semi-readable words embedded in the
colored chaos I am seeing. I should note this occurred when I first
turned on the computer and before booting. Even when I boot to Dos, I
cant see the screen since its all covered with this mass of colors and
lines. No one else has used the computer since I last used it.

The computer is a 700mhz running Windows 98.

I first tightened the video connector, then tried another monitor.
Nothing changed. Both monitors get the same mess on the screen. I
booted to a floppy, again no change. This mobo has a built in video.
I pluged in an old pci video card and now I am getting a normal
picture again.

Did my onboard video fry (I sure cant see how since the computer was
shut off. Or could this be some CMOS setting issue, which would mean
my battery is weak.

Anyone?

TJ

It could be bad memory, in the section of memory used by the built-in
graphics. With the PCI video card in place, you could try memtest86+
from memtest.org . (The theory being, that the internal graphics would
be disabled, and no longer reserving a chunk of system RAM.)

For the CMOS battery, you can touch a voltmeter to the top of the
battery and take a reading. The battery is virtually unloaded by
the motherboard (only 2 microamps or so of current flow), so a voltmeter
reading is valid in this situation. It should read 3V. If below 2.4V,
it could fail any day. (The CMOS data may have checksum protection,
which means the BIOS may be able to detect trouble, and use defaults
instead.)

One way something can fail, from one day to the next, is a cracked
solder joint. Ball grid array chips, have stresses on the solder
balls on the bottom of the package. As the packaging materials heat
up and cool down, the stresses work on the balls. I've seen designs,
where the corner balls would fail, and each footprint design on
the bottom of the chip, will put the peak stress in a different place.
(Some chips leave blank spaces, which provides some stress relief.)
So it could be, the last time the PC cooled down, a pad related to
the graphics section, broke loose.

As long as the PCI video card is working, I'd be pretty happy

Paul

 

[tjustin]

It could be bad memory, in the section of memory used by the built-in
graphics. With the PCI video card in place, you could try memtest86+
from memtest.org . (The theory being, that the internal graphics would
be disabled, and no longer reserving a chunk of system RAM.)

For the CMOS battery, you can touch a voltmeter to the top of the
battery and take a reading. The battery is virtually unloaded by
the motherboard (only 2 microamps or so of current flow), so a voltmeter
reading is valid in this situation. It should read 3V. If below 2.4V,
it could fail any day. (The CMOS data may have checksum protection,
which means the BIOS may be able to detect trouble, and use defaults
instead.)

One way something can fail, from one day to the next, is a cracked
solder joint. Ball grid array chips, have stresses on the solder
balls on the bottom of the package. As the packaging materials heat
up and cool down, the stresses work on the balls. I've seen designs,
where the corner balls would fail, and each footprint design on
the bottom of the chip, will put the peak stress in a different place.
(Some chips leave blank spaces, which provides some stress relief.)
So it could be, the last time the PC cooled down, a pad related to
the graphics section, broke loose.

As long as the PCI video card is working, I'd be pretty happy

Paul

Thanks for the info Paul. I'll do some checking as you said and post
the results. Although I have another slower computer, I am glad this
spare video card works for now. The built in did seem to have faster
graphics though, which is strange because as far as I know, builtin
video is usually crappy, or so I have been told. Of course PCI is not
the fastest either.

You did sort of lose me on the "Ball grid array chips". Why are they
called that? I never seen chips with balls (not that kind <lol>).
I always thought all chips just had pins that were either soldered or
plugged in.

I do have on RAM socket that lost the clip on the end. Actually not
lost, just wont stay on the slot. The Ram is pushed tightly in the
socket, just that clip wont stay attached.

A new mobo may be in the works soon.

 

[Paul]

You did sort of lose me on the "Ball grid array chips". Why are they
called that? I never seen chips with balls (not that kind <lol>).
I always thought all chips just had pins that were either soldered or
plugged in.

The contacts on the bottom of the chip, are only ball shaped
until soldering is complete. The solder melts, and joins the
chip to the PCB.

http://www.siliconfareast.com/pbga-xs.jpg

http://en.wikipedia.org/wiki/Ball_grid_array

You can see the grid on the bottom of this chip here is not
complete. The strategic absence of balls in certain
areas, helps control the stress. The second (PDF)
document, shows more pictures of the patterns
used for BGA packages.

http://www.pcb123.com/images/promotion/bga2.gif
http://www.intel.com/design/packtech/ch_14.pdf

BGAs are part of the SMT (surface mount) technology
world. Surface mount allows components to be soldered
to the surface of the PCB.

Paul

 

[tjustin]

It could be bad memory, in the section of memory used by the built-in
graphics. With the PCI video card in place, you could try memtest86+
from memtest.org . (The theory being, that the internal graphics would
be disabled, and no longer reserving a chunk of system RAM.)

For the CMOS battery, you can touch a voltmeter to the top of the
battery and take a reading. The battery is virtually unloaded by
the motherboard (only 2 microamps or so of current flow), so a voltmeter
reading is valid in this situation. It should read 3V. If below 2.4V,
it could fail any day. (The CMOS data may have checksum protection,
which means the BIOS may be able to detect trouble, and use defaults
instead.)

One way something can fail, from one day to the next, is a cracked
solder joint. Ball grid array chips, have stresses on the solder
balls on the bottom of the package. As the packaging materials heat
up and cool down, the stresses work on the balls. I've seen designs,
where the corner balls would fail, and each footprint design on
the bottom of the chip, will put the peak stress in a different place.
(Some chips leave blank spaces, which provides some stress relief.)
So it could be, the last time the PC cooled down, a pad related to
the graphics section, broke loose.

As long as the PCI video card is working, I'd be pretty happy

Paul

As a followup on this, the computer seemed to fix itself ????????
I did push the cable that goes from the motherboard to the mon. plug
on the case. Thats about all I did. On the other hand, I had a glass
of water that dumped on the floor next to my desk. It was a few feet
from the computer (case open). I now wonder if some water got in the
case?

This does open up another question. When I stuck that PCI video card
in the computer, the onboard video did not work at all until I removed
that PCI card. Is there any way to use TWO video cards at once (and
two monitors)? I never tried this before....

 

[Paul]

As a followup on this, the computer seemed to fix itself ????????
I did push the cable that goes from the motherboard to the mon. plug
on the case. Thats about all I did. On the other hand, I had a glass
of water that dumped on the floor next to my desk. It was a few feet
from the computer (case open). I now wonder if some water got in the
case?

This does open up another question. When I stuck that PCI video card
in the computer, the onboard video did not work at all until I removed
that PCI card. Is there any way to use TWO video cards at once (and
two monitors)? I never tried this before....

That is a function of the BIOS. Some BIOS like to disable the onboard,
when a card is detected. And there are cases, where it does work, and
you can run both at once.

If you want to see how other people set up multiple monitors, take a
look through this collection...

http://www.realtimesoft.com/multimon/gallery.asp

Paul