Hardware problem

[kony]

Which occurs right after the POST, well within the 4min.
mentioned.

There is no software factor including boot sector viri,
which impede entering the bios menus. The machine has
already demonstrated it is a hardware failure at that point.
While it is /possible/ there is also a software failure,
untill the initial hardware failure is resolved it would be
premature to assume it.

You are aware that there are programs that even rewrite
the BIOS from XP?

Yes, they have to have some intelligence of the specific
board to work. Show us even one virus that can write to
this board.

If interaction at the level of totally
rewriting the BIOS are possible, why do you find it so hard
to understand that some of the data that the BIOS must
use can also be changed, is it just because I'm the one
pointing it out?

What is theoretically possible, but never shown to exist
with capability to flash the OP's board, is a scenrio so
unlikely it would not be reasonable to consider unless there
was evidence suggesting it.

There are too many things to even enumerate that are
possible if we can't play odds at least a little bit. There
is no reason to think the bios has been written to, and
particularly not that it had been written to but remained
intact enough to POST the system only locking up soon
thereafter. The odds are far lower than 1 in a million.

But I would also direct you to any definition of the term
"Long Shot". This is based on an actual startup problem
I had with one system of mine, and it did turn out to be
caused by a keyboard driver problem. It would lock
things up so that it wouldn't even POST. Pressing the
Power Button or shorting the pins wouldn't start the sys,
even with the little green light on the MB glowing.

On a restart maybe, but this was a problem from a
powered-off, cold start. No driver will effect getting into
the bios menu from a cold start.

I have no way of knowing if this has anything to do
with the OP's problem, but it is possible. It is a long
shot worth the trouble of plugging in a different simple
keyboard. (One not designed to work with whatever
enhanced drivers there may be on the problem sys.)

If the OP feels the keyboard is faulty, certainly another
should be tried. If the keyboard otherwise works once the
system is running (which it seems to do once ran and
restarted), having the keyboard only exhibit this problem
upon a cold start is also exceedingly unlikely.

 

[Ken Maltby]

There is no software factor including boot sector viri,
which impede entering the bios menus. The machine has
already demonstrated it is a hardware failure at that point.
While it is /possible/ there is also a software failure,
untill the initial hardware failure is resolved it would be
premature to assume it.



Yes, they have to have some intelligence of the specific
board to work. Show us even one virus that can write to
this board.


What is theoretically possible, but never shown to exist
with capability to flash the OP's board, is a scenrio so
unlikely it would not be reasonable to consider unless there
was evidence suggesting it.

There are too many things to even enumerate that are
possible if we can't play odds at least a little bit. There
is no reason to think the bios has been written to, and
particularly not that it had been written to but remained
intact enough to POST the system only locking up soon
thereafter. The odds are far lower than 1 in a million.





On a restart maybe, but this was a problem from a
powered-off, cold start. No driver will effect getting into
the bios menu from a cold start.




If the OP feels the keyboard is faulty, certainly another
should be tried. If the keyboard otherwise works once the
system is running (which it seems to do once ran and
restarted), having the keyboard only exhibit this problem
upon a cold start is also exceedingly unlikely.

Well the keyboard that I replaced worked fine when the
system was running, and still runs fine in another system,
(without the enhanced drivers). It ran fine as a keyboard
except when it was killing the restart of the system. Even
when it was shutdown with the power key and the switch
on the PSU turned off for a while, cold enough for you?

I could agree that it is an odd, probably rare occurrence,
but it did happen to this one system of mine. It gets a
little tiring to listen to you saying that what I've seen with
my own eyes, is not possible. Let's see, Kony and his
theory or objective reality, I'll take the evidence of my own
eyes. It could be that it was a coincidence that when the
keyboard was changed and the drivers removed, the system
could be started reliably again. But I find that highly unlikely.

The OP has nothing to loose trying another keyboard, and
since it is a long shot, the odds are that it won't be the answer
to his problem. Why you are so dead set against even
considering the possibility, could be another question.

Luck;
Ken

 

[CBFalconer]

.... snip ...

There is no software factor including boot sector viri, which
impede entering the bios menus. The machine has already
demonstrated it is a hardware failure at that point. While it is
/possible/ there is also a software failure, until the initial
hardware failure is resolved it would be premature to assume it.

I have to disagree. To illustrate, consider a machine with a
pluggable ROM bios. Unnplug the bios. You will no longer be able
to boot, or to enter any bios menus, etc. Yet normal operation
would be unaffected if you could remove the ROM while running,
because normally the entire bios is over-ridden by opsys code.

If you just want to say 'unlikely', I will agree.

 

[Maurice Batey]

What does this have to do with what?

Merely to add that those tests were tried, rather than
overlooked.

 

[Maurice Batey]

nothing to loose trying another keyboard

Which I have made a note to do, when I can lay my hands on one.

 

[Maurice Batey]

Will resume battle on return on Sunday.

Many thanks for all your efforts to help; much appreciated...

Have nice weekends, folks!

 

[kony]

I have to disagree. To illustrate, consider a machine with a
pluggable ROM bios. Unnplug the bios. You will no longer be able
to boot, or to enter any bios menus, etc. Yet normal operation
would be unaffected if you could remove the ROM while running,
because normally the entire bios is over-ridden by opsys code.

If you just want to say 'unlikely', I will agree.

Remember the system when cold/off, does post and run up
until the point where it locks up. This means it does read
and run the bios from the EPROM, but is then instable.

Unplugging the EPROM would make this much impossible, so
based upon what the system could do, but then where it
stopped, we have eliminated some things.

 

[CBFalconer]

Merely to add that those tests were tried, rather than overlooked.

?? What tests? What problems? On what?

 

[CBFalconer]

Remember the system when cold/off, does post and run up
until the point where it locks up. This means it does read
and run the bios from the EPROM, but is then instable.

Unplugging the EPROM would make this much impossible, so
based upon what the system could do, but then where it
stopped, we have eliminated some things.

Ah, that's different than the scene I envisioned. Conceded.

 

[w_tom]

Will resume battle on return on Sunday.
Many thanks for all your efforts to help; much appreciated...
Have nice weekends, folks!

You have a long list of wild and wanton suggestions. A new keyboard
is not relevant to your problem. Either is some mysterious virus that
exists only in wild speculation.

Kony has provided a responsible list of suspect. This poster was
doing this stuff multiple generations ago. Kony's list is what you
should be concentrating on - and ignore so many other posts based only
in speculation - it could be this or could be that follows by "so
replace it".

The suspect list is where to begin looking for irrefutable facts -
where you 'follow the evidence'. Too many are saying "I suspect it
must be this, therefore it must be this; so replace this". That can
even make the problem exponentially more complex.

Notice after so much work, your accomplishment is zero. You still
don't even know what in the computer is good. Swapping also does not
verify things good.

Get a 3.5 digit multimeter. A tool so simple and so ubiquitous as
to be sold even to K-mart shoppers. IOW if you fear the meter, then
you have no business even trying to fix the computer. Get the meter.
Take some voltage measurements. Where to get those DC voltage numbers
from is listed in "When your computer dies without warning....."
starting 6 Feb 2007 in the newsgroup alt.windows-xp at:
http://tinyurl.com/yvf9vh

In your case, most important numbers are voltages on any one of
orange, red, purple, and yellow wire from power supply to motherboard
just after power switch is pressed. Post those numbers here to learn
even more useful information. Now we have definitive facts - no more
"it could be ..."

IOW first establish power supply 'system' integrity. Yes the power
supply is only one part of that 'system'. If you don't establish
'system' integrity, then 'try this and try that' will take weeks.
Above procedure means the few technically informed here can then post
something useful. Numbers obtained in two minutes.

Another suggested BIOS battery. Only ten second with the meter
provides that number. We remove nothing (because a better number
means the battery stays connected to motherboard). We then know if a
battery is defective OR we know that battery is good. IOW we
accomplish something. But that means numbers and a meter.

Why so few useful responses? A direct reflection of information
provided in your posts. Having provided so few useful facts, then one
is even claiming some nonsense about a virus. Ignore those posters.
There is nothing wrong with the Eprom, as Kony posted - obviously.
But then so many who *know* computers don't even understand what he
posted about BIOS operation. They are the sources of wild speculation
even about overheating.

Take only two minutes. Get those power supply voltage numbers. Post
them here. Obtain useful replies. Only then have you accomplished
something by establishing one subsystem as good. Notice what it takes
to 'follow the evidence' AND accomplish something.

Currently you have a near zero list of good subsystems - and too
many useless replies. You want replies that are useful? Start by
defining what is 'definitively good' and 'definitively bad'. That
means numbers. Provided is where to look for facts:

Cold power up problems tend to be either power supply, flaky
motherboard capacitors, a bad mechanical connection or a PCB
problem such as bad solder joints or board cracks, though
the former two are a lot more common than the latter two.

Since the power supply can make anything or everything look
defective, then we first establish its integrity - two minutes with a
multimeter and then post those numbers.

IOW we fix things only by 'following the evidence'. Which item on
that list is know good after all this time? None. Nothing was
accomplished. That virus? Show me how he concluded by following the
evidence. He did not. That post demonstrated how many *computer
experts* don't even know how a computer works - don't even know how
electricity works - why they are experts only because they speculate.

 

[Ken Maltby]

You have a long list of wild and wanton suggestions. A new keyboard
is not relevant to your problem. Either is some mysterious virus that
exists only in wild speculation.

Kony has provided a responsible list of suspect. This poster was
doing this stuff multiple generations ago. Kony's list is what you
should be concentrating on - and ignore so many other posts based only
in speculation - it could be this or could be that follows by "so
replace it".

The suspect list is where to begin looking for irrefutable facts -
where you 'follow the evidence'. Too many are saying "I suspect it
must be this, therefore it must be this; so replace this". That can
even make the problem exponentially more complex.

Notice after so much work, your accomplishment is zero. You still
don't even know what in the computer is good. Swapping also does not
verify things good.

Get a 3.5 digit multimeter. A tool so simple and so ubiquitous as
to be sold even to K-mart shoppers. IOW if you fear the meter, then
you have no business even trying to fix the computer. Get the meter.
Take some voltage measurements. Where to get those DC voltage numbers
from is listed in "When your computer dies without warning....."
starting 6 Feb 2007 in the newsgroup alt.windows-xp at:
http://tinyurl.com/yvf9vh

In your case, most important numbers are voltages on any one of
orange, red, purple, and yellow wire from power supply to motherboard
just after power switch is pressed. Post those numbers here to learn
even more useful information. Now we have definitive facts - no more
"it could be ..."

IOW first establish power supply 'system' integrity. Yes the power
supply is only one part of that 'system'. If you don't establish
'system' integrity, then 'try this and try that' will take weeks.
Above procedure means the few technically informed here can then post
something useful. Numbers obtained in two minutes.

Another suggested BIOS battery. Only ten second with the meter
provides that number. We remove nothing (because a better number
means the battery stays connected to motherboard). We then know if a
battery is defective OR we know that battery is good. IOW we
accomplish something. But that means numbers and a meter.

Why so few useful responses? A direct reflection of information
provided in your posts. Having provided so few useful facts, then one
is even claiming some nonsense about a virus. Ignore those posters.
There is nothing wrong with the Eprom, as Kony posted - obviously.
But then so many who *know* computers don't even understand what he
posted about BIOS operation. They are the sources of wild speculation
even about overheating.

Take only two minutes. Get those power supply voltage numbers. Post
them here. Obtain useful replies. Only then have you accomplished
something by establishing one subsystem as good. Notice what it takes
to 'follow the evidence' AND accomplish something.

Currently you have a near zero list of good subsystems - and too
many useless replies. You want replies that are useful? Start by
defining what is 'definitively good' and 'definitively bad'. That
means numbers. Provided is where to look for facts:

Since the power supply can make anything or everything look
defective, then we first establish its integrity - two minutes with a
multimeter and then post those numbers.

IOW we fix things only by 'following the evidence'. Which item on
that list is know good after all this time? None. Nothing was
accomplished. That virus? Show me how he concluded by following the
evidence. He did not. That post demonstrated how many *computer
experts* don't even know how a computer works - don't even know how
electricity works - why they are experts only because they speculate.

You again, why don't you an Kony get a room?

No one suggested that a virus was the cause of the OP's
problem. I was the only one to even mention a virus and
that was just as an example of how code/software can
effect the startup process. You talk of your adherence
to the scientific principle/requirement for numeric data, but
you deliberately misinterpreted statements or totally invent
them to try and make your point.

Should the OP follow standard troubleshooting practice;
of course he should. Should the power supply be checked
and established as good, yes that is a given. But a quick
review of our previous encounters suggests that, for you,
the problem is always the power supply or grounding. Just
like for Kony it's always a heat related problem or failing
capacitors.

My posting about a similar situation where the cure turned
out to be replacing the keyboard and its drivers, was offered
because it was one rare case where traditional troubleshooting
didn't provide the answer. It was a "long shot" that I wouldn't
expect to be considered, in the normal troubleshooting process.
(Which is the very reason I felt it needed bringing up.) Because
it fixed a very similar problem I thought it might benefit the OP.

Most of the ammunition for your arguments against the postings
here comes not from statements addressing the OP's problem, but
from my replies to Kony's knee jerk reaction to my posting. I
foolishly thought I could provide some information to show why
it is possible, countering Kony's claims that it wasn't. I should
have known that nothing I could say, would allow Kony (or such
as you) to see that anything but their suggestions could have merit.

I don't actually know if the possible explanations I provided were
the ones at work in the situation I was describing, I do know that
the actions I took and described, fixed the problem. It is certainly
possible that totally different factors were in play, that I did not
detect. I do know that what I did worked, when changing good
MBs, cases, processors, memory, and yes power supplies, did
not isolate or fix the problem.

I certainly hope that the readers don't buy into Kony's or your
position that only postings that meet your approval or that you
originate, are worth considering.

Luck;
Ken

 

[kony]

Just
like for Kony it's always a heat related problem or failing
capacitors.

I again invite you to look over my past posts and prove your
statement that I frequently claim "it's always a heat
related problem", with the exception of cases where the OP
had already indicated as much, or there was a direct link
like a failing fan.

If you haven't seen much gear with failing capacitors, I
suggest you haven't seen much gear. I can't will a cap to
pop, but once it does there is little speculation needed -
to merely look inside a computer is an obvious enough
suggestion to make when troubleshooting hardware, and while
one is inside, they might as well move an eyeball or two
towards the vicinity of the capacitors. It is quick, easy,
free, and requires no special ability on the part of the
person looking except to recognize that a cap should be flat
on top, stand up straight, and not have crusty residue
leaking out.

If I had ever seen or even heard of one single virus that
flashed any semi-modern motherboard bios and this resulted
in the board posting but then locking up, I might then
consider it worth mentioned from time to time. Instead I
only mention things that have any reasonable chance of
happening without some indication of an unusual
circumstance.

My posting about a similar situation where the cure turned
out to be replacing the keyboard and its drivers, was offered
because it was one rare case where traditional troubleshooting
didn't provide the answer. It was a "long shot" that I wouldn't
expect to be considered, in the normal troubleshooting process.
(Which is the very reason I felt it needed bringing up.) Because
it fixed a very similar problem I thought it might benefit the OP.

There's no need to be defensive, we're all just throwing out
ideas. Nobody is going to guess right immediately, 100% of
the time, but some things are far less likely and/or
contrary to the evidence presented. Group collaboration can
reduce the magnitude of things someone troubleshooting would
need try, and hopefully reduce the time and expense to get
the system working right again.

Most of the ammunition for your arguments against the postings
here comes not from statements addressing the OP's problem, but
from my replies to Kony's knee jerk reaction to my posting.

That's an interesting statement, considering I had posted
nothing to this thread yet when you wrote:

"Have you and "Grinder" been replaced by "kony" clones?"

While your suggestion about drivers and bios can help, maybe
even solve some seemingly related problems, in this case
there was a specific reason to believe it could be ruled
out. The only way to narrow the focus and have someone try
a few targeted things, is to also rule out the hundreds if
not thousands of things that don't need to be considered.

When a cold-off system locks up before one can even get into
the bios menu, which is necessarily before it has ever tried
to read anything from the hard drive, when the only thing
running as of yet is the decompressed bios, we know there is
no software issue because there is no software running yet,
nor any chance a volatile register held the wrong values -
because it was a clean boot from cold off, something the
system had to have been doing correctly up until the problem
started.

I
foolishly thought I could provide some information to show why
it is possible, countering Kony's claims that it wasn't. I should
have known that nothing I could say, would allow Kony (or such
as you) to see that anything but their suggestions could have merit.

I suppose if I was a virus writer, and knew the specific
board someone had, it would be possible to target them and
write something to NV memory that would allow a system to
POST but be instable, and yet, there'd be a checksum error
and if this was just an inappropriate configuration value
that prevented the board from running stabily, clearing CMOS
should resolve it. I suppose if we wanted to stretch things
even further, it might be possible to redo the entire bios
such that it passed the checksum test and then was instable,
but how many hours of work and motherboards would the hacker
have to have, identical to the target system? Some ideas
are just so "out there" that until there is one single
example, we have to ignore what is theoretically possible.

If we want to talk theory, in theory a snake might have
crawled inside and done some damage. I have actually had a
system someone brought that had a snake in it (they had
removed the snake before bringing it), and I vaguely recall
a web picture of a snake in a PSU. With these two examples
of snakes in systems, I have more reason to suggest looking
for snakes than you would to wonder about a bios flashing
virus that leaves the system posting, and yet you are
welcome to scavenge my posts to see if I've ever mentioned
snakes in systems except for this example - because it's
just not likely enough to mention.

I don't actually know if the possible explanations I provided were
the ones at work in the situation I was describing, I do know that
the actions I took and described, fixed the problem. It is certainly
possible that totally different factors were in play, that I did not
detect. I do know that what I did worked, when changing good
MBs, cases, processors, memory, and yes power supplies, did
not isolate or fix the problem.

I certainly hope that the readers don't buy into Kony's or your
position that only postings that meet your approval or that you
originate, are worth considering.

I'm sorry I got you bent into a knot Ken, the goal here is
to solve the OP's problem and if I know based on the
evidence that it can't be software or drivers, shouldn't I
have mentioned that instead of having the OP try things that
won't fix the problem? The OP made no mention of changing
drivers just prior to the onset of the problem, and the
system is now 3 years old. While we can't be guaranteed the
OP has not changed drivers, we can only go on what has been
written, and have to assume if there is no driver change
mentioned, there is no driver change... and if there is no
driver change, it should not now be locking up because of a
driver when it had ran same driver previously, UNLESS there
is some other hardware problem too... and if there is some
other hardware problem, it'll have to be solved at least
enough that the system is not locking up even at entry to
the bios before drivers can even be considered.

We don't know that there isn't a driver problem, but we do
know there is at least one problem not related to drivers,
causing an instability. It is exceedingly difficult to try
to troubleshoot OS or drivers when the hardware isn't even
stable yet before loading any software.

 

[Ken Maltby]

On Thu, 30 Aug 2007 21:37:35 -0500, "Ken Maltby"


If I had ever seen or even heard of one single virus that
flashed any semi-modern motherboard bios and this resulted
in the board posting but then locking up, I might then
consider it worth mentioned from time to time. Instead I
only mention things that have any reasonable chance of
happening without some indication of an unusual
circumstance.



There's no need to be defensive, we're all just throwing out
ideas. Nobody is going to guess right immediately, 100% of
the time, but some things are far less likely and/or
contrary to the evidence presented. Group collaboration can
reduce the magnitude of things someone troubleshooting would
need try, and hopefully reduce the time and expense to get
the system working right again.


That's an interesting statement, considering I had posted
nothing to this thread yet when you wrote:

"Have you and "Grinder" been replaced by "kony" clones?"

Your "Sock Puppet" didn't use that in his argument, just
my responses to your later posts.

While your suggestion about drivers and bios can help, maybe
even solve some seemingly related problems, in this case
there was a specific reason to believe it could be ruled
out. The only way to narrow the focus and have someone try
a few targeted things, is to also rule out the hundreds if
not thousands of things that don't need to be considered.

When a cold-off system locks up before one can even get into
the bios menu, which is necessarily before it has ever tried
to read anything from the hard drive, when the only thing
running as of yet is the decompressed bios, we know there is
no software issue because there is no software running yet,
nor any chance a volatile register held the wrong values -
because it was a clean boot from cold off, something the
system had to have been doing correctly up until the problem
started.




I suppose if I was a virus writer, and knew the specific
board someone had, it would be possible to target them and
write something to NV memory that would allow a system to
POST but be instable, and yet, there'd be a checksum error
and if this was just an inappropriate configuration value
that prevented the board from running stabily, clearing CMOS
should resolve it. I suppose if we wanted to stretch things
even further, it might be possible to redo the entire bios
such that it passed the checksum test and then was instable,
but how many hours of work and motherboards would the hacker
have to have, identical to the target system? Some ideas
are just so "out there" that until there is one single
example, we have to ignore what is theoretically possible.

Now lets add back in a portion of my post that you felt
shouldn't be quoted here:

" No one suggested that a virus was the cause of the OP's
problem. I was the only one to even mention a virus and
that was just as an example of how code/software can
effect the startup process. You talk of your adherence
to the scientific principle/requirement for numeric data, but
you deliberately misinterpreted statements or totally invent
them to try and make your point."

Then you just ignore what was posted, so you can put
words in my mouth. There was no virus involved in the
startup problem that was cured by replacing the keyboard
and drivers. The point was that some errant or corrupted
code could effect the startup process, and that it could
explain the events I observed, in fixing a startup problem
by replacing the keyboard and its drivers.

 

[kony]

Your "Sock Puppet" didn't use that in his argument, just
my responses to your later posts.

It seems this conversation is degrading pretty fast.

Do we have anything useful to cover now that might help the
OP with this system or are we both just wasting time?

Now lets add back in a portion of my post that you felt
shouldn't be quoted here:

" No one suggested that a virus was the cause of the OP's
problem. I was the only one to even mention a virus and
that was just as an example of how code/software can
effect the startup process. You talk of your adherence
to the scientific principle/requirement for numeric data, but
you deliberately misinterpreted statements or totally invent
them to try and make your point."

"Shouldn't" be quoted? It was left out because it is
non-applicable to the OP's primary problem.

Then you just ignore what was posted,

Well, if you say so.

 

[Plato]

Looking for a good newsgroup to help me diagnose a weird hardware problem with
my 3-year old desktop.

If this is not it, could someone recommend one, please?

(Inconsistent symptoms, but machine always locks up within 4 minutes of cold
power up, regardless of which operating system is being booted.

Look for heat related problems. ie clogged up cpu fansinks, not enough
case fans, etc.

 

[Guest]

Maurice Batey wrote:

Look for heat related problems. ie clogged up cpu fansinks, not enough
case fans, etc.

It's highly logical of you to blame lack of heat for a heat problem

 

[Maurice Batey]

I unplugged and replugged all the power connections in the PC
(apart from the big one plugged into the motherboard, which I simply could not
get out).

After powering up from cold it has been running normally (WinXP, Kubuntu,
Mandriva) -

No more lockups, but since that day I have been using the
following procedure, just to be on the safer side, after powering
off overnight.


(1) Half hour before wanting to use PC, boot up the PC and let
it sit at the prompt for BIOS password (with everything else
still powered off).T
(2) After half an hour, power on monitor, router, etc, and
boot some system up.

Later on I might just try without the half hour warm up ...

Been OK for sevetral days now. Fingers crossed...

 

[Maurice Batey]

I have been using the following procedure, just to be on the safer side, after powering
off overnight.

(1) Half hour before wanting to use PC, boot up the PC and let
it sit at the prompt for BIOS password (with everything else
still powered off).T
(2) After half an hour, power on monitor, router, etc, and
boot some system up.

Gradually reduced the warm-up time to 5 minutes from power up,
and have had no lockups for 2 months now...