Sudden crash and restart

[Nicholas]

This problem happened a few times and is now a regular occurrence. I
will be working on the machine and it will suddenly die (like a power
failure) and then immediately restart. It runs through POST etc., the
crashes again, repeats, then eventually hangs after memory check. If I
switch off the machine. It will then start up, do a disk check and load
windows. I work for a while and then the happen again. Is this is some
sort of Hardware problem? Could something be conflicting with windows
and doing this?

Pollywell computer
Leadtek K7NCR18DPRO2 MoBo
Athalon 2500
512 ram
gforce FX5200

 

[Bob Day]

This problem happened a few times and is now a regular occurrence. I
will be working on the machine and it will suddenly die (like a power
failure) and then immediately restart.

Check out your hardware:

1. Make sure your CPU and case temperatures are OK, and
that the CPU fan, case fans, and any other fans are
working.

2. Monitor your power supply voltages for a while -- check for
fluctuations and voltages out of spec.

3. Shut down your computer and turn off power *at the
surge protector* (which cuts off standby power while
leaving your computer grounded) or unplug it. Take
the standard protections against static electricity
(wear a wrist strap or keep one hand grounded at all
times).

4. Open up your computer and dust it out. *Do not use a
vacuum cleaner* - it might cause static discharge.
Get some dust remover spray (RadioShack has it) and
a dust mask, and take your computer outside and spray
out the dust. Keep the end of the spray straw at least
six inches from any components and keep it moving so as
to avoid excessive thermal shock to the components.

5. Reseat all the modules, including the memory modules,
and cables inside your computer.

6. Turn on your computer, and if it boots, run a few cycles
of "memtest86" (download from http://www.memtest86.com)
to thoroughly check out your memory. You should get zero
errors. If it doesn't boot, shut down and remove your
CMOS battery and check its voltage (or just replace it).
In any case, Clear CMOS (see your mainboard manual for
how to do it) and set up your BIOS again.

--Bob Day
http://bobday.vze.com

 

[Nicholas]

Bob,

thanks for the reply. I'll try your advise. Any suggestions on how to
monitor the case temp and power supply? Is there some software that logs
the info?

 

[kony]

Bob,

thanks for the reply. I'll try your advise. Any suggestions on how to
monitor the case temp and power supply? Is there some software that logs
the info?

Most motherboards do not take the temp of the case ambient air,
rather the temp of an integrated sensor chip that creates it's
own heat. To get a reliable (or sometimes even remotely
accurate) reading you'd need make or buy a temp sensor that
either has it's own separate display (standalone) or that
connects to a system bus for reading by software.

Likewise you'd need a separate sensor to take PSU readings unless
you have a PSU that has it's own temp output, and a motherboard
with a spare pin header for this input. Most PSU and many
motherboards do not have these features.

 

[Trent©]

This problem happened a few times and is now a regular occurrence. I
will be working on the machine and it will suddenly die (like a power
failure) and then immediately restart. It runs through POST etc., the
crashes again, repeats, then eventually hangs after memory check. If I
switch off the machine. It will then start up, do a disk check and load
windows. I work for a while and then the happen again. Is this is some
sort of Hardware problem? Could something be conflicting with windows
and doing this?

Pollywell computer
Leadtek K7NCR18DPRO2 MoBo
Athalon 2500
512 ram
gforce FX5200

What operating system?


Have a nice week...

Trent©

Follow Joan Rivers' example --- get pre-embalmed!

 

[Nicholas]

Win XP home.

I have downloaded the memtest86 as recommended (have not run it yet.)
and a temp monitoring program (hmonitor). I installed hmonitor and the
program lists my mainboard at about 33 c (yellow) on start up and after
some use it goes up to 50 c (red) None of the other temps seem to rise
to the red level. It 50 c very hot for a mobo? When it reaches that
temp I shut down the machine. Using this method I have not had the
crash. This weekend I will do all the tests that Bob recommended.

thanks again to all--i'm learning.

 

[kony]

Win XP home.

I have downloaded the memtest86 as recommended (have not run it yet.)
and a temp monitoring program (hmonitor). I installed hmonitor and the
program lists my mainboard at about 33 c (yellow) on start up and after
some use it goes up to 50 c (red) None of the other temps seem to rise
to the red level. It 50 c very hot for a mobo? When it reaches that
temp I shut down the machine. Using this method I have not had the
crash. This weekend I will do all the tests that Bob recommended.

thanks again to all--i'm learning.

When using a hardware monitor program not specifically for your
motherboard (or even for it, if manufacturer isn't debugging
enough) it may easily report the wrong sensor for a component.
IN other words, it's most likely that the temp you saw was for
CPU, not motherboard. A CPU (at stock speed) should not crash at
50C though, but if it rises above 60C there is much greater odds
of crash, would be an indicator that some additional case
cooling, better heatsink, or removal/regrease/dusting of
heatsink-fan is due.

 

[Trent©]

Win XP home.

Did you check your Event Viewer? With XP, it could be crashing
because of a software problem.

Turn the reboot on error option off...then see how it acts.

But it sounds like a CPU heat problem...especially if the computer
runs fine when you first start using it.

BTW...its ALWAYS a bad idea to take out components and reseat
them...unless you have a specific defective component in mind.

Good luck...let us know.


Have a nice week...

Trent©

Follow Joan Rivers' example --- get pre-embalmed!

 

[Bob Day]

Did you check your Event Viewer? With XP, it could be crashing
because of a software problem.

Turn the reboot on error option off...then see how it acts.

But it sounds like a CPU heat problem...especially if the computer
runs fine when you first start using it.

BTW...its ALWAYS a bad idea to take out components and reseat
them...unless you have a specific defective component in mind.

Nonsense. I wouldn't do it for no reason, but reseating all
the modules and cables often fixes mysterious problems.

-- Bob Day

 

[Trent©]

Nonsense. I wouldn't do it for no reason, but reseating all
the modules and cables often fixes mysterious problems.

For ONE particular problem...of which you know not...you'll take out
ALL the modules and cables?

Never a good idea...and usually a last resort of amateurs.


Have a nice week...

Trent©

Follow Joan Rivers' example --- get pre-embalmed!

 

[kony]

For ONE particular problem...of which you know not...you'll take out
ALL the modules and cables?

Never a good idea...and usually a last resort of amateurs.

It can depend on the situation. If you built it yourself and
took your time, it isn't too old, then that might seem like a
waste, but there are some situations where systems were moved,
cheap cases flex, drunk lunatics were dancing around inside the
case or who-knows-what else has happened.

Certainly the best attempt at resolution is to directly attack
the problem, but in some cases it could waste hours to days of
troubleshooting if at least the parts weren't checked for being
well-seated in the slots, aligned with rear of board, cables
oriented correctly, etc, etc. Sometimes it's just easier to
write "reseat" than itemize every connection in the system that
might've been bad, and also we have to consider that "hopefully"
before someone has posted here they've at least opened the case,
which is good, but also yet another opportunity to disturb
something.

 

[Bob Day]

It can depend on the situation. If you built it yourself and
took your time, it isn't too old, then that might seem like a
waste, but there are some situations where systems were moved,
cheap cases flex, drunk lunatics were dancing around inside the
case or who-knows-what else has happened.

Also, over time, the contacts on modules and cables can tarnish,
degrading the quality of the electrical contact. Reseating components
can clean the contacts a little bit, and improve the electrical quality
of the contact.

-- Bob Day

 

[Trent©]

Also, over time, the contacts on modules and cables can tarnish,
degrading the quality of the electrical contact. Reseating components
can clean the contacts a little bit, and improve the electrical quality
of the contact.

-- Bob Day

That is NOT true. Whoever told you that? Have you tested it
yourself?


Have a nice week...

Trent©

Follow Joan Rivers' example --- get pre-embalmed!

 

[Trent©]

It can depend on the situation. If you built it yourself and
took your time, it isn't too old, then that might seem like a
waste, but there are some situations where systems were moved,
cheap cases flex, drunk lunatics were dancing around inside the
case or who-knows-what else has happened.

Certainly the best attempt at resolution is to directly attack
the problem, but in some cases it could waste hours to days of
troubleshooting if at least the parts weren't checked for being
well-seated in the slots, aligned with rear of board, cables
oriented correctly, etc, etc. Sometimes it's just easier to
write "reseat" than itemize every connection in the system that
might've been bad, and also we have to consider that "hopefully"
before someone has posted here they've at least opened the case,
which is good, but also yet another opportunity to disturb
something.

Well, call me old-fashioned, I guess.

But if someone brings in a machine with a sound problem, I never pull
the video card...and the CPU...and the modem, etc.

Maybe its just me. But I try to evaluate the problem...then try to
solve it.

YMMV.


Have a nice week...

Trent©

Follow Joan Rivers' example --- get pre-embalmed!

 

[bearman]

That is NOT true. Whoever told you that? Have you tested it
yourself?


Have a nice week...

Trent©

Follow Joan Rivers' example --- get pre-embalmed!

Gee, Trent. I guess using a pencil eraser on contacts is out too, huh? But
it works for me.

Given the probable thinness of the "gold" plating on the contacts, I would
expect them to wear through to the base metal when first installed and be
subject to some kind of corrosion.

Same principle applies to cable connectors.

Bearman

 

[Trent©]

Gee, Trent. I guess using a pencil eraser on contacts is out too, huh? But
it works for me.

Works for me, too...I do it often. But I never take a card OUT to
clean it up. More damage is caused by amateurs...but taking out a
card and letting dust fall into the slots.

Given the probable thinness of the "gold" plating on the contacts, I would
expect them to wear through to the base metal when first installed and be
subject to some kind of corrosion.
Nonsense.

Same principle applies to cable connectors.

What principle?


Hope ya'll have a nice 4th of July weekend...

Trent©


Proud member of the Roy Rogers fan club!

 

[kony]

Works for me, too...I do it often. But I never take a card OUT to
clean it up. More damage is caused by amateurs...but taking out a
card and letting dust fall into the slots.

Dust is not so much of a problem as a film buildup, like tobacco
tar, especially if mixed with the dust.

Nonsense.

It could depend on the manufacturer, how thick they make it, but
typically that plating is good for a few hundred insertion
cycles, IIRC. In cases where factories are often swapping test
cards they do follow best procedures by using an intermediary
contact/joiner thingy (exact name escapes me at the moment) so
that it's contacts are worn instead of device contacts wearing...
but of course this isn't going to happen outside of a mass
testing scenario.

What principle?

Many aren't corrosion resistant. Most have no, or insufficient,
corrosion inhibitor. Over time they will oxidize and/or build up
dirt. The amount of time could vary based on environment and
metal, type of contact. It is the same in a PC as in any other
equipment, that the mechanical contacts (and mechanically moving
parts) are earlier failure points. At least a PC is usually in a
relatively mild environment so the effect takes longer.

 

[bearman]

Works for me, too...I do it often. But I never take a card OUT to
clean it up. More damage is caused by amateurs...but taking out a
card and letting dust fall into the slots.


What principle?

Removing them and reinstalling them.

 

[Trent©]

Removing them and reinstalling them.

Do you think its because yer CLEANING something?


Hope ya'll have a nice 4th of July weekend...

Trent©


Proud member of the Roy Rogers fan club!

 

[Nicholas]

Thanks again to all who responded. Just an update. I bumped up the
air-condition in my house 3 degrees live in Florida and the temps are in
the 90's everyday) and opened the case on the machine. It has not
crashed since. The software monitor has maintained a 41-42 degree temp
down from the over 50 previously. Although I am not certain, the
Hmonitor program seems to work well with my mobo. It seems to list
correctly all the parts including the removable hard drives.

I am convinced that it is the heat in this room that caused the problem.
This room stays much hotter than the rest of the house. Any
recommendations on a powerful/quiet case fan that I can replace mine with.
This is the case I have
http://www.polywell.com/us/desktop/images/case/BlueEdge-Silver-Tower.jpg

Any other case cooling recommendations that do not include "move the
computer to another room?"

Thanks again