Problem building new PC

  • Thread starter Thread starter pgnl
  • Start date Start date
P

pgnl

I have just bought a 'pre-tested' motherboard/CPU/Fan/Memory assembly from a
reputable supplier and am having problems. (MSI KT4V-L, AMD XP+2400, 256Mb
PC2700 memory).

I have installed everything in the case (the hard drive arrived dead, I
checked the motherboard with two other drives and they were OK, the jumpers
were set to master on the failed drive, so that's gone back to the supplier
but that's another story). I have attempted to install XP pro using one of
the working drives.

A few minutes after formatting the drive I received a blue screen
'error0x00000050 page_fault_in_non-paged-area', eventually I managed to get
past this by switching off one of the bios settings. Eventually after
rebooting a number of times I managed to get XP installed, but periodically
I get error messages requesting I send a report to MS, even worse - from
time to time the PC just reboots without prompting. Also the machine seems
slower than my existing P3 866....I don't think it's a temperature issue,
inside the box is 28 degrees C and the processor is 40 degrees C according
to the Bios.

Is this most likely to be bad memory? What do you think guys?

Thanks for your help,

Patrick
UK
 
I have just bought a 'pre-tested' motherboard/CPU/Fan/Memory assembly from a
reputable supplier and am having problems. (MSI KT4V-L, AMD XP+2400, 256Mb
PC2700 memory).

I have installed everything in the case (the hard drive arrived dead, I
checked the motherboard with two other drives and they were OK, the jumpers
were set to master on the failed drive, so that's gone back to the supplier
but that's another story). I have attempted to install XP pro using one of
the working drives.

A few minutes after formatting the drive I received a blue screen
'error0x00000050 page_fault_in_non-paged-area', eventually I managed to get
past this by switching off one of the bios settings. Eventually after
rebooting a number of times I managed to get XP installed, but periodically
I get error messages requesting I send a report to MS, even worse - from
time to time the PC just reboots without prompting. Also the machine seems
slower than my existing P3 866....I don't think it's a temperature issue,
inside the box is 28 degrees C and the processor is 40 degrees C according
to the Bios.

Is this most likely to be bad memory? What do you think guys?

Thanks for your help,

Patrick
UK
Click to expand...

Usually badly seated RAM/Bad contacts on the RAM/Incompatible or
Faulty RAM :(



--
Free Windows/PC help,
http://www.geocities.com/sheppola/trouble.html
It's a G not a J in gmx for email
Free songs download,
http://www.soundclick.com/bands/8/nomessiahsmusic.htm
 
In simple terms, the computer tried to access memory that
software was not suppose to access. It could be a software
programming error (unlikely) or hardware problem. Any number
of reasons could account for that error message. First
establish what is and is not working - which is why better
computer systems also provide a complete set of diagnostics
for free.

BTW, supplier only verified system with his unique
combination of attachments. Example, memory could provide
data anywhere from 20 to 35 nanoseconds later. But board only
works right with your particular CPU when data is provided 25
nanoseconds later - which is what most memory does. But this
one memory - well within specs - provides data anywhere from
25 to 30 nsec. Defect was in motherboard - not memory. But
motherboard tested just fine in shop.

Now you are having intermittent failure - but no part tests
defective. A hypothetical example of what W E Deming cited as
reasons for 'no quality'. Using test to confirm board was
good - a big time 'no no' if one wants quality.

In the meantime, the first item that can cause numerous
strange problems is power supply. No matter where it was
tested previously or what it did previously, power supply
integrity is still unknown until verified by measurement with
the essential 3.5 digit multimeter while driving THAT system.
Just like a house that is sagging; first check its foundation
In this case, first check power supply. Chart for limits is
also found at:
http://www.hardwaresite.net/faqpowersupply.html

Another has suggested memory. Yes, another of many
possibilities. They only checked that board worked in one
small example of possible combinations. And so we must check
memory using that manufacturer's diagnostics - or if he is a
cheap bastard, then use memory diagnostics available on web
such as MemTst86 or Docmem. Memory must be tested hot -
uncomfortable to touch which is normal and pleasant
temperature to memory (if in doubt, then get that memory's
data sheet). IOW memory must be heated with a hairdryer on
high or run in a 100+ degree F room while all diagnostics are
run.

Why heat? Heat is rarely reasons for computer failure.
Heat is more often how to aggravate symptoms of defective
hardware. Too many then try to fix the symptoms (more fans)
rather than fix a technical problem they don't understand.
Clone computer industry thrives on this technological
ignorance. Same ignorance that too often does not start with
diagnostics AND understand how to make diagnostics effective.

And so it goes with every possible reason for failure. Once
power supply and memory is verified, then maybe move on to
video subsystem. Of course the board was configured and
verified using the same CPU?

Other tests include running the system with a minimal
configuration - keyboard, mouse, video card, memory, CPU, and
minimum number of mass storage devices (CD-Rom, disk drive).

Was board mounted to chassis so that it only makes a single
conductive connection between chassis and motherboard? Just
an example of other more trivial reasons for failure.
 
pgnl said:
I have just bought a 'pre-tested' motherboard/CPU/Fan/Memory assembly from a
reputable supplier and am having problems. (MSI KT4V-L, AMD XP+2400, 256Mb
PC2700 memory).

I have installed everything in the case (the hard drive arrived dead, I
checked the motherboard with two other drives and they were OK, the jumpers
were set to master on the failed drive, so that's gone back to the supplier
but that's another story). I have attempted to install XP pro using one of
the working drives.

A few minutes after formatting the drive I received a blue screen
'error0x00000050 page_fault_in_non-paged-area', eventually I managed to get
past this by switching off one of the bios settings. Eventually after
rebooting a number of times I managed to get XP installed, but periodically
I get error messages requesting I send a report to MS, even worse - from
time to time the PC just reboots without prompting. Also the machine seems
slower than my existing P3 866....I don't think it's a temperature issue,
inside the box is 28 degrees C and the processor is 40 degrees C according
to the Bios.

Is this most likely to be bad memory? What do you think guys?

Thanks for your help,

Patrick
UK
Click to expand...


PSU may not be strong enough. You need at least a 350 or 400 watt.

bluestringer
 
Well that's reassuring to hear it from a third party.

So maybe it's a good idea to try a different slot for the ram then? (it was
supposedly tested and found to be working ;(.......)
Click to expand...

Another thing you can try, (not as a solution, just to check on ram)
is to reduce the aggressivity for ram. Your bios should have some
alternative.
Change auto to manual and you can also lower the memorybus
(-memorybus! not FSB-clock or FSB) frequency slightly.

(I've had this problem with an 'almost' working system where the ram
reported OK to all ramtests. Still, dropping frequency and CL cured
the problem, as did the replacement ram-stick.)


ancra
 
Ancra said:
Another thing you can try, (not as a solution, just to check on ram)
is to reduce the aggressivity for ram. Your bios should have some
alternative.
Change auto to manual and you can also lower the memorybus
(-memorybus! not FSB-clock or FSB) frequency slightly.

(I've had this problem with an 'almost' working system where the ram
reported OK to all ramtests. Still, dropping frequency and CL cured
the problem, as did the replacement ram-stick.)


ancra
Click to expand...

Thanks for all your suggestions esp the long reply by w_tom. Unfortunately,
the kit came with a single stick of memory but I intended to upgrade anyway
so have ordered some morewhich will arrive shortly, then I will certainly
try out your advice Shep.
Bluestringer - the power supply is 350W so should be OK, hard to determine
the quality of it though.
Ancra-not sure I can change the memory bus on my motherboard you can
certainly change the CAS Latency though...
Thanks for your help once again guys.

Patrick
UK
 
JAD said:
do you have memory hole enabled? in bios?
Click to expand...
Thanks for the response JAD but I am not familiar with the term 'Memory
Hole' - there is no reference to it in the motherboard manual, if you mean
enabling of that particular memory slot on the motherboard, it doesn't seem
to allow for enabling of a particular slot...Perhaps you could elaborate for
me?

Patrick
UK
 
Thanks for the response JAD but I am not familiar with the term 'Memory
Hole' - there is no reference to it in the motherboard manual, if you mean
enabling of that particular memory slot on the motherboard, it doesn't seem
to allow for enabling of a particular slot...Perhaps you could elaborate for
me?

Patrick
UK
Click to expand...

It's a very old BIOS option and should no be enabled on modern systems
even if it exists.



--
Free Windows/PC help,
http://www.geocities.com/sheppola/trouble.html
It's a G not a J in gmx for email
Free songs download,
http://www.soundclick.com/bands/8/nomessiahsmusic.htm
 
P4B266 -533 both still have it...and the damn thing defaults, after clearing, to 'enable' on both. That can give all kinds of page /
memory faults... I had an ISA modem that required that setting to be enabled(jumpers were no help on the card) turns out it would
only work in warp...2 I think it was. anyway...it was a reach, but since the problem still persists ....................
 
A plastic pencil eraser is good for cleaning the contacts on the bottom of
the ram sticks. Also, check the RAM settings in BIOS, eg your RAM may be
333 but the BIOS is set for 222 (CAS Lat, RAS to CAS, RAS precharge). Set
the RAM timings slower (bigger numbers) for a start and then get something
like SiSoft Sandra to read the RAM modules to tell you their speed -
assuming the lower RAM settings work long enough to boot up. Otherwise, try
the RAM in another machine and see how that works, trying the RAM one stick
at a time, you may still have to drop the RAM speed on that machine as well.

Dave
 
Well I replaced the supplied memory with some that I bought from Crucial,
reset the BIOS to default (and reset the processor speed to 133). Now all
seems stable (it's been on for 4 hours) - I now have another problem though,
which I will post separately. Thanks for your advice guys.

Patrick
UK


JAD said:
P4B266 -533 both still have it...and the damn thing defaults, after
Click to expand...
clearing, to 'enable' on both. That can give all kinds of page /
memory faults... I had an ISA modem that required that setting to be
Click to expand...
enabled(jumpers were no help on the card) turns out it would
only work in warp...2 I think it was. anyway...it was a reach, but since
Click to expand...
the problem still persists ....................
 
Using eraser to clean contacts was a classic benchmark for
technicians who did not learn their trade. Another benchmark
were liberal use of freeze spray.

Connectors are self cleaning. And if any residual impurity
affects a connection, then the circuit has more problems than
dirty contacts. Digital noise requirements in any properly
designed electronics must be far more than what any 'dirty'
contact will affect. However some 'feel' that shiny means
better and advocate erasing contacts.

If cleaning the contacts avoids failure, then there are
serious timing problems or defective connectors on motherboard
or in memory. Fix the problem. Don't cure symptoms. If
fixing connector problems with an eraser, then also buy a
voodoo doll to avoid future problems.
 
However some 'feel' that shiny means
better and advocate erasing contacts.
Click to expand...

_________________________________________________________

Of course shiny is better.

Or do you recommend applying a nice even film of
Krud-Kote Kontact Kover? :-)
 
If you don't like it then don't do it. To me it sounded like he had second
hand parts, obtained cheaply. This was a cheap solution. I also have used
a paintbrush to clean out the ram sockets on old boards to get them to work.
Isopropyl and a cotton bud/q-tip/whatever works as well on the RAM sticks
too.

You have criticised but have offered no alternative and a load of
opinionated drivel.

Freeze spray is damn good stuff, as is a hair dryer and the back of a
screwdriver, each used where appropriate. What do you do, keep replacing
parts until it works?
 
Using eraser to clean contacts was a classic benchmark for
technicians who did not learn their trade. Another benchmark
were liberal use of freeze spray.
Click to expand...

What? Ever lived by the ocean? I've seen plenty of corrosion and an
eraser is a great way to clean them. I do agree that techs who
troubleshoot with freeze exclusively can chase their tail by causing
all kinds of timing issues with digital equipment, but it can come in
handy for that old stereo or TV and no schematic.
Connectors are self cleaning. And if any residual impurity
affects a connection, then the circuit has more problems than
dirty contacts. Digital noise requirements in any properly
designed electronics must be far more than what any 'dirty'
contact will affect. However some 'feel' that shiny means
better and advocate erasing contacts.
Click to expand...

Shiny is clean and clean allows proper contact between connections.
Most PC sockets and RAM modules are far from top quality and tend to
make poor contact on some connections. Are you saying a guy should
throw away salvagable equipment because it isn't top quality? That
doesn't make sense.
If cleaning the contacts avoids failure, then there are
serious timing problems or defective connectors on motherboard
or in memory. Fix the problem. Don't cure symptoms. If
fixing connector problems with an eraser, then also buy a
voodoo doll to avoid future problems.
Click to expand...

Cleaning the socket and RAM modules *is* fixing the problem. If
contacts aren't mating because of corrosion, cleaning is the fix. A
nice touch is to add a spot of Progold G5 on both connector and module
to help keep the oxidation from reoccuring. Remember a little goes a
long way. A can should be in every techs kit.
 
nice touch is to add a spot of Progold G5 on both connector and module
to help keep the oxidation from reoccuring. Remember a little goes a
long way. A can should be in every techs kit.
Click to expand...


DeoxIT works better for me.....
 
Do you read before posting? Posted on 19 Oct were solutions
- fixing a problem without wildly speculating.

But many can only fix things by what they see. Its shiny;
therefore it works better. Good electrical connectors are
self cleaning. Break and make the connection - at the
connection is sufficient clean. Digital designs (properly
done) include worst case parameters for any connector 'dirt'.
Therefore if cleaning the 'dirt' fixes a problem, then human
is only curing symptoms.

This post is about those who fix things without sufficient
background knowledge - who fix things using freeze spray,
erasers, and worry about dirty dust. Nonsense - or what we
used to call the technician who would not long have a job.
Solutions based upon science were provided previously in this
thread. Eraser is not a solution to digital electronic
problems. Eraser is just to make the human happy because it
is clean.
 
Back
Top