Many many errors...

In recent months, my system has become very unstable. at 1st, never a
problem, rock solid at any time. then maybe 2x a month I would get an
occasional crash that a simple reboot would fix. then it came to be once or
2x a week (at this point I still had not changed any hardware, software or
bios settings) and now crash-o-matic every stinking minute of every day!
This has happened over a course of 4-6 months. Now I have changed out ALL
parts EXCEPT the motherboard and cpu which these are the only parts I do not
have to spare or test with. Seems like I mostly get memory errors such as
MEMORY_MANAGEMENT BSOD's, such & such data could not be written/read at such
& such memory address (address is different HEX every time), the referenced
data pointed to an invalid memory location, other problems consist of
corrupt files, registry corruption, IRQL BSOD's and other BSOD's that don't
have a name or code. Even errors such as Internet explorer has caused a
problem and needs to close, for no apparent reason. Its never ending now it
seems. I have been told by several tech-savvy people that its the CPU as the
memory controller is built onto the chips now adays however I'm not so sure.
one strange thing is that if I remove 2 of my 4 stix of ram the system is
completely stable. once I give it 4 stix of any ram the errors happen again.
all I have are 512mb stix of ram and 1g is not enough to get me thru the day
altho the past week I've had no choice but to run 1gb of ram. all memory
used in this system tested with memtest86 and all passed with flying colors,
even when tested on this system. something else that is odd to me, I will go
through hours/days of continuous errors as mentioned above then all of the
sudden the system will be fine and dandy for a few hours then out of the
blue it starts again. this thing is driving me insane! Additionally the
crashes happen from it being idle at the login screen to playing not such a
high demanding game like WoW or even simpler games like solitaire. anyway,
thanks for your time and thanks a ton for any info provided.

Gigabyte GA-K8NS Ultra-939 AGP 8X nForce 3 based chipset (almost 3 yrs old)
AMD 4800+ X2 dual core CPU, never overclocked and runs about 40C under heavy
load. (1 yr old)
Currently an ATi X800XT AIW AGP (tested 2 others, one nvidia and another
ATi)
2.0 GB RAM OCZ plat. rev 2. (Also tested Rev 1 and some corsair value
select)
2x western dig raptors 36gbx2 (also tested a single western dig 120gig and a
Seagate 80.
tried 2 different PSU, a 550W and currently a 600W
have tried win xp pro 32 bit, win xp pro x64 edition, windows vista prem.
x64 to no avail.
currently running win xp pro 32 bit with all current drivers, still non stop
crashes
system resides in a mesh filtered case with plenty of fans and airflow and
runs continuously ice cold.

I tried to be as specific as possible but if I left anything out please feel
free to ask.

"AnotherAnonymous" wrote in message
t...
In recent months, my system has become very unstable. at 1st, never a
problem, rock solid at any time. then maybe 2x a month I would get an
occasional crash that a simple reboot would fix. then it came to be once
or
2x a week (at this point I still had not changed any hardware, software or
bios settings) and now crash-o-matic every stinking minute of every day!
This has happened over a course of 4-6 months. Now I have changed out ALL
parts EXCEPT the motherboard and cpu which these are the only parts I do
not
have to spare or test with. Seems like I mostly get memory errors such as
MEMORY_MANAGEMENT BSOD's, such & such data could not be written/read at
such
& such memory address (address is different HEX every time), the
referenced
data pointed to an invalid memory location, other problems consist of
corrupt files, registry corruption, IRQL BSOD's and other BSOD's that
don't
have a name or code. Even errors such as Internet explorer has caused a
problem and needs to close, for no apparent reason. Its never ending now
it
seems. I have been told by several tech-savvy people that its the CPU as
the
memory controller is built onto the chips now adays however I'm not so
sure.
one strange thing is that if I remove 2 of my 4 stix of ram the system is
completely stable. once I give it 4 stix of any ram the errors happen
again.
all I have are 512mb stix of ram and 1g is not enough to get me thru the
day
altho the past week I've had no choice but to run 1gb of ram. all memory
used in this system tested with memtest86 and all passed with flying
colors,
even when tested on this system. something else that is odd to me, I will
go
through hours/days of continuous errors as mentioned above then all of the
sudden the system will be fine and dandy for a few hours then out of the
blue it starts again. this thing is driving me insane! Additionally the
crashes happen from it being idle at the login screen to playing not such
a
high demanding game like WoW or even simpler games like solitaire. anyway,
thanks for your time and thanks a ton for any info provided.

Gigabyte GA-K8NS Ultra-939 AGP 8X nForce 3 based chipset (almost 3 yrs
old)
AMD 4800+ X2 dual core CPU, never overclocked and runs about 40C under
heavy
load. (1 yr old)
Currently an ATi X800XT AIW AGP (tested 2 others, one nvidia and another
ATi)
2.0 GB RAM OCZ plat. rev 2. (Also tested Rev 1 and some corsair value
select)
2x western dig raptors 36gbx2 (also tested a single western dig 120gig and
a
Seagate 80.
tried 2 different PSU, a 550W and currently a 600W
have tried win xp pro 32 bit, win xp pro x64 edition, windows vista prem.
x64 to no avail.
currently running win xp pro 32 bit with all current drivers, still non
stop
crashes
system resides in a mesh filtered case with plenty of fans and airflow and
runs continuously ice cold.

I tried to be as specific as possible but if I left anything out please
feel
free to ask.



Although it's always risky to answer a post such as this, what you describe
is so nearly identical (board and all) to mine and how it behaved that I am
pretty confident I am not sending you a bum steer. Mine did all this and
when I finally established that, not only would it only run with one stick,
but that it would only run with it in the second slot, I was confident
enough I was dealing with bad electrolytic capacitors to ask for an RMA from
GigaByte. My board was still under the three-year limit, though yours
probably no longer is.

They did repair it. In fact, when the board was returned to me, every
single electrolytic cap on it had been replaced. (I can't imagine how much
that cost them, but the board was out of production, and the CPU a problem
in transferring to any new one). You might *try* getting them to accept
responsibility for it, but unless you were able to show just how common that
model board failure might be, I doubt you'd get very far.

Next best is to hope that it's an *isolated* failure, best of all possible
situations: an isolated leaking or bulging electrolytic. If you don't know
what you're looking for, I'd suggest you Google for 'bad cap' which would be
likely to bring some examples of the so common failure of motherboards and
PSUs due to inadequate or poor-quality electrolytic capacitors. I don't
have any good pictures right now, only one of a cap I removed from this
machine I'm using right now, and it isn't clear enough in what you'd be
looking for. If you don't recognize electrolytics, they look like tubes,
standing upright with closed ends. The up-end is generally a flat aluminum
disk, with, say, a cross pattern embossed in it, or, worse news, a
three-spoke pattern, but other patterns as well. The cylindrical side is
usually encased in a light plastic film sleeve, and the capacitor is mounted
on the motherboard with two leads from the (should be flat)-bottom soldered
into eyelets on the board, so that it just looks like an upright small can.
Size varies widely, from half the height of an 'AAA' alkaline cell to the
size of a 'D' cell, thinking in flashlight terms. Signs of trouble a
bulging top or bottom (it should be flat); leakage (evident as an
oily-looking liquid, or perhaps a dried-up bubbled residue). Sometimes they
are split open. The one on this machine was split open, but the machine
generally worked acceptably, so where the cap is means a lot as to how
evident will be your trouble. Look around your memory slots to see if
there's anything obvious. Or around your chipset.

Even so, unless you are extremely handy with soldering and desoldering,
you'll probably find that the cost of repair will exceed your valuation of
the board and CPU, given the age of the pair and the performance of current
equipment. Don't go to a big box store, many of which seem to be folding,
anyhow. Choose some shop that makes its living repairing commercial
electronic equipment and they'll give you the straight up and down on it.

Sorry to not bring more positive news :-|

Joe

"jt3" wrote in message
...

Although it's always risky to answer a post such as this, what you
describe
is so nearly identical (board and all) to mine and how it behaved that I
am
pretty confident I am not sending you a bum steer. Mine did all this and
when I finally established that, not only would it only run with one
stick,
but that it would only run with it in the second slot, I was confident
enough I was dealing with bad electrolytic capacitors to ask for an RMA
from
GigaByte. My board was still under the three-year limit, though yours
probably no longer is...

Thanks for your reply. I am familiar with leaky caps and such (unlucky me,
from past experiences). I have since removed the board and very carefully
visually inspected it. Visibly, I can't see any signs of leaky caps, or
obvious problems on the board although I do not have any testing equipment
to test the components for any hidden issues. I did however call AMD today
to start an RMA process on my chip since its probably the only piece of
hardware that is still under warranty. I explained the situation to them and
thoroughly described my setup to the technician. He then told me that due to
a limitation with their chips, when running 4x stix of ram, their processors
(or atleast the socket 939 versions) will automatically clock the memory
down to 166/333MHz. I do recall noticing this happening in the bios when I
installed my 4x stix of ram when I 1st built my pc nearly 3 yrs ago so I
manually set the memory to 200/400MHz then and it has been that way since
day one and running without a hitch. What I (nor he, the tech) don't
understand is why now, all of the sudden this would happen. I told him I had
hard-set the mem to 400 and he said to set it to 333 or auto. Well I have
done this (as much as I hated to) and so far it seems completely stable
again. its been running prime95 now for about 2 hrs with no errors while
prime95 caused BSOD's the second it was started prior to this. He still
accepted my RMA and offered to replace the cpu. I may give it some time, or
go ahead and replace the chip while I still can before I start throwing $
around. Thanks alot for your reply, I had not thought about checking for
caps prior. hopefully theres no hidden probs with them.

Here's what bad 'caps' look like-:
http://home.earthlink.net/~doniteli/index27.htm here's a couple of
links that 'may prove interesting....

apparently some bad caps was due to industrial espionage....!
http://www.niccomp.com/taiwanlowesr.htm and
http://www.pcmag.com/article2/0,4149,939887,00.asp The ones affected
are certain ones from Taiwan only.... the Japanese ones are ok.


I had a couple bulge in my GA7VA rev 1 mobo which I changed, in my
case it did not affect it's operation, the rev 2 mobo's used superior
ones. I can't say for your's as is a different mobo from mine.

I never like the look of bulging caps, they can leak gunge which can
be corrosive and eat the copper tracks away.... mine didn't leak,
might have done if I had left them in though.


Some Gigabyte mobos I understand could be unstable when running near
maximum with certain brands of RAM, wonder if thats worth checking..?

Davy