Crash, no reboot, have to clear CMOS

  • Thread starter Thread starter Fitz
  • Start date Start date
F

Fitz

AMD 64 3200+ on Gigabyte K8NS (f18 BIOS)
2 X 512MB Mushkin Blk Lvl 1 PC3500 (2-3-3-5)
2 X 36 GB Raptors (RAID 0)
1 Maxtor 80 GB IDE Drive
BFG 6800 GT OC
M-Audio Revolution 7.1 soundcard
Plextor 716A DVD/RW
Antec 430W PS
Floppy/card reader

Crashes, randomly during 3D games(monitor goes white and black
patchwork, occasionally sound loopback)
Pressing reset gets the monitor back, but computer usually won't
complete boot process- stops
at posting memory or prior to detecting drives. Have to clear CMOS and
re-enter BIOS preferences.

Passes Memtest86 (v3.2). Have updated DirectX. Tried reducing hardware
acceleration in Sound. Disabled sound.
Relaxed memory timings. Disconnected DVI input and used analog. Disabled
Fastwrites. Power supply checks good,
w/i 5%. Drivers all up to date.

I'm now beyond my level of knowledge and am reduced to guessing. Unless
someone has an idea, I'm betting on the
motherboard. If I replace it with a different brand, but same chipset
and raid controller, will I have to reinstall Windows
or should I be able to create a RAID array and use the installed OS?

Thanx,
Fitz
 
I am going to guess power supply and I'll tell you why.
I had the same exact thing happen, changed out a ps with a noisy fan
to a brand new 600 watt supply. Playing quake 4 it would crash and the only
thing I changed was the ps. Put the old one back and it ran fine. The
voltages check out too.
The 600 watt will run the system but not during that game/ go figure.
 
Fitz said:
AMD 64 3200+ on Gigabyte K8NS (f18 BIOS)
2 X 512MB Mushkin Blk Lvl 1 PC3500 (2-3-3-5)
2 X 36 GB Raptors (RAID 0)
1 Maxtor 80 GB IDE Drive
BFG 6800 GT OC
M-Audio Revolution 7.1 soundcard
Plextor 716A DVD/RW
Antec 430W PS
Floppy/card reader

Crashes, randomly during 3D games(monitor goes white and black patchwork,
occasionally sound loopback)
Pressing reset gets the monitor back, but computer usually won't complete
boot process- stops
at posting memory or prior to detecting drives. Have to clear CMOS and
re-enter BIOS preferences.

Passes Memtest86 (v3.2). Have updated DirectX. Tried reducing hardware
acceleration in Sound. Disabled sound.
Relaxed memory timings. Disconnected DVI input and used analog. Disabled
Fastwrites. Power supply checks good,
w/i 5%. Drivers all up to date.
Click to expand...

I know many people will disagree with me, but I would never check a power
supply with a multimeter. All that can tell you is if the voltages are
stable enough to look good on a multimeter. In other words, the multimeter
voltage readings can tell you if the power supply MIGHT be good, or is
definitely bad. But then, if the voltages are so far off that the
multimeter confirms the power supply is bad, you shouldn't have needed the
multimeter to KNOW that the power supply was bad. For example, if your +12V
reading is lower than +10V, the symptoms are going to be so obvious that you
will know you need a new power supply. Same with the other voltages. If
they are too far off, you are going to just KNOW that the power supply is
bad, even though you don't know specifically what the voltages are. So at
best, the multimeter isn't helpful (can't confirm the power supply is good),
or is redundant (you should have known the thing was bad without checking
voltages). So why bother with the multimeter?

Based on what you wrote, you have a bad power supply or a bad motherboard.
Considering that power supplies fail with alarming frequency, and that you
have not confirmed that your power supply is good (only that it isn't
totally dead), you should start there. A power supply can be defective, or
too weak to run your system, and still look GREAT on a multimeter. I'd be
willing to bet your power supply is bad. -Dave
 
BigJim said:
I am going to guess power supply and I'll tell you why.
I had the same exact thing happen, changed out a ps with a noisy fan
to a brand new 600 watt supply. Playing quake 4 it would crash and the
only
thing I changed was the ps. Put the old one back and it ran fine. The
voltages check out too.
The 600 watt will run the system but not during that game/ go figure.
Click to expand...

That's exactly why I tell people not to check power supply voltages with a
multimeter. At best, it can't confirm that a power supply is good. It can
only tell that the power supply isn't totally dead. -Dave
 
Fitz said:
AMD 64 3200+ on Gigabyte K8NS (f18 BIOS)
2 X 512MB Mushkin Blk Lvl 1 PC3500 (2-3-3-5)
2 X 36 GB Raptors (RAID 0)
1 Maxtor 80 GB IDE Drive
BFG 6800 GT OC
M-Audio Revolution 7.1 soundcard
Plextor 716A DVD/RW
Antec 430W PS
Floppy/card reader

Crashes, randomly during 3D games(monitor goes white and black
patchwork, occasionally sound loopback)
Pressing reset gets the monitor back, but computer usually won't
complete boot process- stops
at posting memory or prior to detecting drives. Have to clear CMOS and
re-enter BIOS preferences.
Click to expand...
why do you have to clear the MOS? checksum error?
Passes Memtest86 (v3.2). Have updated DirectX. Tried reducing hardware
acceleration in Sound. Disabled sound.
Relaxed memory timings. Disconnected DVI input and used analog. Disabled
Fastwrites. Power supply checks good,
w/i 5%. Drivers all up to date.
Click to expand...

power supply would be my guess also...thats where I would start.
its easy for me because I always have a PSU to check with.
 
why do you have to clear the MOS? checksum error?<

No error msg. Stops boot process and locks up. No beeps, no msg, and no go.
Clear CMOS and it boots fine. Incidentally, it managed to lock up during
a reboot
without having froze up during game play first. I had restarted after
turning off
some processes (another attempt to solve the problem) and it locked up
during
boot sequence. The power supply wasn't tested with a multimeter- just
checked
with an Antec PS tester (has a 20 pin connector and is supposed to put a
load
on the PS when checking). Rails were monitored using Everest. If the
software monitor
is anywhere near accurate, then all 3 rails check normal - with the
computer running
and under load (though certainly not being stressed).

Thanx,
Fitz
 
Mike T. said:
That's exactly why I tell people not to check power supply voltages
with a multimeter. At best, it can't confirm that a power supply is
good. It can only tell that the power supply isn't totally dead.
Click to expand...

Wrong, it can also tell you when the rail is out of spec.

Sure, its not much use with a higher than acceptible noise level.
 
I know many people will disagree with me, but I would never check a power supply with a
multimeter.
Click to expand...

More fool you.
All that can tell you is if the voltages are stable enough to look good on a multimeter.
Click to expand...

Wrong, it also tells you if the rails are within spec and with a decent
min max multimeter also tells you if the rail ever gets out of spec too.
In other words, the multimeter voltage readings can tell you if the power supply MIGHT be good, or
is definitely bad.
Click to expand...

Which is what you normally want to know with the exception of noise levels.
But then, if the voltages are so far off that the multimeter confirms the power supply is bad, you
shouldn't have needed the multimeter to KNOW that the power supply was bad.
Click to expand...

Oh bullshit. Most obviously when the rail occasionally
goes out of spec due to an intermittent or because
the power supply cant keep up with the varying load.
For example, if your +12V reading is lower than +10V, the symptoms are going to be so obvious that
you will know you need a new power supply.
Click to expand...

Wrong, particularly when that is regulated down to the
cpu voltage on the motherboard and isnt on the hard
drives, particularly if that sag isnt there all the time.
Same with the other voltages. If they are too far off, you are going to just KNOW that the power
supply is bad, even though you don't know specifically what the voltages are.
Click to expand...

Wrong again. And what you are trying to do is decide if the symptoms
you see are due to a bad power supply or a bad motherboard. If the
rails are within spec with a decent max min power supply, over the
time you see the symptom, and the noise levels look fine, you can
be reasonably confident that its worth trying another motherboard.
So at best, the multimeter isn't helpful (can't confirm the power supply is good), or is redundant
(you should have known the thing was bad without checking voltages). So why bother with the
multimeter?
Click to expand...

Because you may not have a spare power supply to try.
Based on what you wrote, you have a bad power supply or a bad motherboard. Considering that power
supplies fail with alarming frequency, and that you have not confirmed that your power supply is
good (only that it isn't totally dead), you should start there. A power supply can be defective,
or too weak to run your system, and still look GREAT on a multimeter.
Click to expand...

Not with a decent max min multimeter.
I'd be willing to bet your power supply is bad.
Click to expand...

Yeah, but you get so many things completely wrong that what you
will bet on is completely irrelevant. Just another mug punter basically.
 
If you do not replace the motherboard with the EXACT same model, you will
need to reformat and reinstall.
 
DaveW said:
If you do not replace the motherboard with the EXACT same model, you will
need to reformat and reinstall.
Click to expand...
Thanks- that's what I was afraid of. This is going downhill fast.

Fitz
 
DaveW said:
If you do not replace the motherboard with the EXACT same model, you will
need to reformat and reinstall.
Click to expand...


Not true...
as long as the new mobo is at least similar...a repair install should do the
trick...
 
Fitz said:
No error msg. Stops boot process and locks up. No beeps, no msg, and no
go.
Clear CMOS and it boots fine.
Click to expand...

It is possible that the CMOS clearing is a red herring. In order to really
clear the CMOS, you'd have to unplug the power supply. If a power supply
shuts itself down due to a momentary low voltage condition, unplugging it
could clear that condition and allow the power supply to be ready to try to
work again. -Dave
 
Not true...
as long as the new mobo is at least similar...a repair install should do
the
trick...
Click to expand...


I'm not at all convinced this guy's motherboard is toast. Everything I've
read seems to point to a bad power supply. And considering that power
supplies fail with alarming frequency, it is not logical at this point to be
leaning toward replacing the mainboard. -Dave
 
I'm not at all convinced this guy's motherboard is toast. Everything I've
read seems to point to a bad power supply. And considering that power
supplies fail with alarming frequency, it is not logical at this point to be
leaning toward replacing the mainboard. -Dave
Click to expand...

I'm not convinced the board is toast either. But aside from the already
described
problem, I also find it interesting that the board tries to run my
memory at 333MHz.
It Mushkin Lvl 1 PC3500, which in the past it has detected and run at
400MHz when
left in "Auto". Now I have to manually set it to 200MHz.

Couple other reasons I'm replacing the board first....it's cheaper than
a good quality
power supply (754 pin technology- not give away, but damn near). I'm not
real pleased
with the board I've got in there, and ordered an Abit NF8-V2. I live in
Alaska, so my
shopping choices for PS's are CompUSA ($) or Best Buy (only carry one
brand, some no-name
product). I could order one, but shipping to AK isn't cheap (FedEx 2nd
Day is cheapest
option)- $15 on shipping for the board from Newegg- PS would be more.

So, against the recommendation of most who responded, I'm going board
first, then PS.
If I'm wrong, I'm out $60. If I go PS first, I'm down twice that if it
turns out to be the
board. Besides, if I end up getting a PS, I'll be getting one suitable
for the next (re)build,
which would mean SLI capable, 24 pin, etc.

Fitz
 
After little thought and some poor planning, I ordered a new motherboard
to replace
the existing one and a new power supply (Corsair 620HX). In the event I
don't need the
PS (if the motherboard fixes the problem), then I have a new PS that
supports PCI-E
SLI for the next build. It'll look nice sitting on a shelf waiting for
the next set of parts.

Fitz
 
Back
Top