2 PCs not starting up properly nor shutting down

[Guest]

Greetings all.

Just when I thought I'd seen it all, this happens to me (us)...

This is going to be a fairly lengthy story; but please stick with me because
I really need some help here.

Some time on Saturday my wife's PC started acting up. Random file errors,
the occasional blue screen, etc. She tried a system restore, stuff like that.
I was at work, so I couldn't look at it myself.

On Sunday things got worse. I advised her to start making backups; but then
she couldn't even see the "cut/copy/paste" menus anymore. She tried to
reinstall XP; but things went downhill from there. Sunday was another work
day for me; I couldn't really do anything until Monday.

On Monday I took her primary drive out of her PC, and hooked it up to my
system. I made a full backup of her drive; and consequently reformatted it.
We both use XP Pro SP2, so I wasn't expecting any problems.

OK. So this is where the weird stuff starts.

I put the HD back in her PC, and got ready for a fresh installation.
However, as soon as I plugged her AC cable back in, her PC started. Fans and
lights came on, drives started spinning, etc. without even pressing the
case's power button. Nothing showed up on her screen, except the message that
the monitor didn't have a signal. I could not shut down her PC by holding the
front power button.

After messing with this for a while I found out that the only way to shut
down the PC was to use the PSU's power button. And that is now the only way
to start it up to (note that it doesn't boot).

I made sure all hardware was hooked up properly. Even reset the BIOS, etc.
Nothing really helped.

Fortunately for my wife, she has a backup PC. An older AMD Duron 1.3 GHz
system. Not lightning fast, but she keeps it as a backup, and it worked fine
when she put it in the closet.

My first thought was to swap the power supplies. Don't ask me why. When I
did I noticed the same problem: the system can only be shut down with the
power supply's power button, not the case's power button. Weird...

So I hooked the power supply back up to the Duron's motherboard, put that in
place of the newer (Sempron) system, and fired it up. To my surprise the
Duron system no showed the same problem: it can only be started/shut down
with the power supply's power button! And again: no boot.

So at this point I'm sort of losing it. What the heck is going on...

I started suspecting an uninterruptable power supply that is hooked up to my
wife's gear, so I took that offline. The same happened, with both PCs. I
verified the AC plug's wiring and voltage, no problems.

Regardless, I hooked up both PCs to an AC outlet that I know is good (and
which my own PC is hooked up to) and the same kept happening.

However, after some experimenting, I managed to reinstall XP Pro on the old
Duron, using the "good" AC outlet. So now I have an "old" slow Duron system,
fresh install, that works OK, exept for the fact that it doesn't power up the
way it should, and doesn't power down the way it should: I have to use the
power supply's power button.

The old Duron is now "working"; that is to say I have it set up with a fresh
copy of XP Pro, so my dear wife can do her classes and homework online. When
I try to shut it down through XP, both the power and HD led go out after a
while, but the screen keeps displaying "Windows is shutting down". It never
really does though, until I flip the switch on the back.

I've been messing with this for two days now. I'm left with two, non-working
(not properly anyway) PCs. The UPS has been taken offline. I have one more PC
here, my own; but to be honest I'm a little hesitant to start swapping
hardware. It's the only one in the house still functioning properly...

I haven't had time today to look at the Sempron system; and I'm going to
wait with that until the old Duron gets fixed (if it can be fixed at all).

I told you this would be a lenghty story...

What do you think went wrong? And, how can I fix it?

All suggestions and ideas are welcome!

Thanks,

Erik

 

[Guest]

With pc power,if troubles are'nt easily figured out,repeating a turn on/off
w/o using the normal operation will certainly cause other failures
quickly....
If both pcs are faulting with same problem,a power surge or similiar may have
taken place,youre best bet,unhook all from pc tower,take to local independent
pc repair shop.Most will have it figured out while you wait,usually free....

 

[Guest]

Hi Andrew.

I agree, there may have been a power surge. Despite the surge suppressors -
including the one in the UPS... Wouldn't surprise me, in this black hole of
the country.

Taking it to the "local" store is not an option, I'm afraid. There's just no
such thing...

And so that takes me back to my original question: has anyone ever seen
something like this before, multiple PCs failing with the same symptoms? And,
what turned out to be defective?

Thanks,

Erik

 

[PA20Pilot]

Hi Erik,

........what turned out to be defective?

I had a problem that had a couple of your symtoms, at least on this
computer. The biggest plug from the power supply to the motherboard was
not making a good connection on one of the pins. I scrapped it clean and
closed it down a little with a needle and things have been working well
since.

I was getting the no signal every now and then and the OS was freezing
every now and then if the signal was working. You might just try
cleaning everything cable wise and see what happens.

Good luck, it has a lot to do with it too.

---==X={}=X==---

Jim Self

AVIATION ANIMATION, the internet's largest depository.
http://avanimation.avsupport.com

Your only internet source for spiral staircase plans.
http://jself.com/stair/Stair.htm

Experimental Aircraft Association #140897
EAA Technical Counselor #4562

 

[w_tom]

After all that work, now tell us; what is 'known good' and what is
'known bad'? Nothing? After all that work, nothing has been
accomplished? The problem has been made exponentially worse due to
shotgunning. Trying to fix Windows when hardware integrity is unknown
and when Windows is not yet on the suspect list will only create more
unknowns; therefore exponentially complicate things. Fixing Windows
also destroyed useful facts.

Step back. Start all over again because everything is in the third
category - unknown. Move each 'component' from the unknown category
to 'definitively good' or 'definitively bad'. The world is ternary.

Everything will look bad if a power supply is failing. Fans can
spin; lights illuminate; and still a power supply can be 100%
defective. A defective supply that works in one computer can appear
fail in another. No way around using a 3.5 digit multimeter as sold
in any 'guys' store from K-mart to Lowes for maybe $20. Perform a two
minute procedure described in "When your computer dies without
warning....." starting 6 Feb 2007 in the newsgroup alt.windows-xp
at:
http://tinyurl.com/yvf9vh

With numbers both before power switch is pressed and as power switch
is pressed, then a power supply 'system' is 'definitively good' or
'definitively bad'. Post those numbers here for further insights. It
is a supply 'system' - far more than just one 'system' component - a
power supply. Just another reason why shotgunning can so complicate a
problem.

Having established the state of that 'system', never look back.
Something has been accomplished. We know that something in that
computer is 100% functional or defective. Move on.

Long ago, important facts were being stored in the system (event)
log. Windows finds problems, records them, and works around them.
This information so that you can deal with problems when you are
ready. Well, useful facts were destroyed by 'fixing' Windows without
first learning what was wrong. So move on.

All responsible computer manufactures provide comprehensive hardware
diagnostic in one complete package. Software that tests hardware
without Windows complications. Strip the problem down into parts;
then analyze only those parts. Utilities examine well beyond what
Windows can see and report back with numbers. If your computer
manufacture was not so responsible, then download each diagnostic from
the appropriate hardware provider. If the computer manufacture did
not provide that one complete hardware diagnostic, then download one
from the disk drive manufacturer, from the video controller
manufacturer, sound card providers, some memory diagnostic (maybe from
a third party such as MemTst86), etc. Establish what is and is not
good long before disconnecting or replacing anything. Don't shotgun.
First identify the suspect.

At this point, something may appear defective. If still confused,
well the better informed can only reply if you provide facts.
Diagnostics and voltages are facts and numbers then empower the more
technical informed posters. By verifying all hardware and confirming
power supply voltages, now much has been accomplished. We know what
is good; so move on to other suspects. All hardware worked just fine
without Windows. Only now suspect Windows since it is the last
'unknown'. And still we don't repair anything yet. Instead identify
the suspect.

Only replace a 'definitively bad' item. A list of what is
'definitively good' means something has been accomplished. That's the
point. After all that shotgunning, nothing has been accomplished and
the problem has been made more complex.

You have no reason to believe a power surge existed or did damage.
If suppressors were adjacent to the computer, then a surge can even be
provided more destructive paths to earth via electronics. High
reliability facilities don't use plug-in suppressors for this and
other reasons. The effective suppressor is located far from
electronics and typically 'less than 10 feet' from earth ground.
Surges that can overwhelm protection already inside all electronics
occur typically once every seven years - a number that can vary
significantly. Others without that knowledge and numbers that
automatically assume only what they understand as the reason for
failure. Manufacturing defects are a far more common source of
failures such as yours.

Currently, we have near zero facts to answer why PCs are failing.
Too much shotgunning has exponentially complicated the problem -
severely clouded the water. Shotgunning can even cause damage to a
good computer - just another reason we avoid shotgunning. To obtain a
useful reply, your replies will only be as good as the information you
provide. That means 'definitive' answers, numbers, and working step
by step to establish subsystem and component integrity. Using
information provided, every answer would only be wild speculation -
"it could be this or could be that". Get numbers. Get
diagnostics. Know "this is good, that is bad, and those are still
unknown". It is a ternary world.

 

[bud--]

You have no reason to believe a power surge existed or did damage.
If suppressors were adjacent to the computer, then a surge can even be
provided more destructive paths to earth via electronics. High
reliability facilities don't use plug-in suppressors for this and
other reasons. The effective suppressor is located far from
electronics and typically 'less than 10 feet' from earth ground.

Excellent information on surges and surge protection is in an IEEE guide at:
http://omegaps.com/Lightning Guide_FINALpublishedversion_May051.pdf
And one from the NIST at:
http://www.nist.gov/public_affairs/practiceguides/surgesfnl.pdf

Both say plug-in suppressors are effective.

Note that all interconnected equipment needs to be connected to the same
plug-in suppressor, or interconnecting wires need to go through the
suppressor. External connections, like phone, also need to go through
the suppressor.

 

[w_tom]

Excellent information on surges and surge protection is in an IEEE guide at:http://omegaps.com/Lightning Guide_FINALpublishedversion_May051.pdf
And one from the NIST at:http://www.nist.gov/public_affairs/practiceguides/surgesfnl.pdf

Both say plug-in suppressors are effective.

Bud promotes for plug-in protector manufacturers. He is not honest
enough to admit it. Bud's citation http://omegaps.com/Lightning Guide_FINALpublishedversion_May051.pdf
even demonstrates what happens when a protector is adjacent to
electronics. Page 42 Figure 8 - the surge is earthed 8000 volts
destructively through the adjacent TV. If the computer was damaged by
a surge, well, look what happens when the protector is not earthed; is
instead too close to electronics - Page 42 Figure 8.

Bud hopes you don't read those pamphlets carefully. A protector not
at earth ground only protects from a type of surge that typically does
not damage computers. The pamphlets warn why a plug-in protector does
not provide protection - can only be effective with an earthing
connection. Worse, protector circuits may earth the typically
destructive type of surge 8000 volts destructively through the wife's
computer.

Bud's post is irrelevant to the OP's problem. Computer has what
appears to be a manufacturing defect. Bud is my troll - follows me
everywhere. Ignore the troll. He posts nothing that explains or
solves problems with the wife's computer. If Bud was honest or was
posting something relevant, then the wife has no computer problems.
Damage even though it had an adjacent plug-in protector circuit.
Could not happen according to Bud.

Provided are where to begin fixing that computer, including and not
limited to "When your computer dies without warning....." starting 6
Feb 2007 in the newsgroup alt.windows-xp at:
http://tinyurl.com/yvf9vh

 

[Guest]

Hi all.

My apologies for the late reply; I have the most "fantastic" work hours...
But today I had some time off.

I managed to test the drives of the first computer that failed, today; and
they all seem to work fine. I connected those to my own PC, one by one, did
some testing (with HD tune, to give them a good workout, and played a DVD in
the optical drive). No problems at all. (I just realized I forgot to test the
disk drive...)

Other hardware in the PC that failed first were an AGP card, and a wireless
network card. The wireless card was installed in the backup PC a few days
ago, and works great. The AGP card, unfortunately, I can't test. The backup
PC has no AGP slot, and my own PC has a PCIe slot...

I have not managed to do anything on the backup PC; my wife needed it for
her online classes today.

OK, I did the voltage test as described in your article, on the first PC.

I disconnected the ATX connector from the MB for the first part of the test;
because the CPU heatsink obscures part of it, and is hard to reach with the
probes.

The purple wire measures 5.20 V. After hooking up the ATX connector to the
MB the voltage goes down to about 5.10 V.

The green wire measures 4.66 V. (again, before being plugged into the MB).
After I plug it in to the MB, and turning on the PSU's switch, it goes to
0.78 V.

The gray wire measures 5.15 V. ("well above 2.4 V", as you say in your other
message, but maybe too high?)

I measured 3.43 V, 5.15 V and 12.10 V on the orange, red and yellow wires,
respectively.

I hope you can tell me a little more from these voltages. Please remember
that I couldn't perform the test exactly as you describe; because pressing
the power button on this PC doesn't do anything at all.

If it is significant: I noticed that the power supply's fan does not come on
when the ATX connector is not hooked up to the MB (even though the PSU is
switched on in the back).

Does this mean anything to you?

Thanks,

Erik

 

[w_tom]

I'm confused why you are disconnecting power supply from
motherboard - or disconnecting anything. Disconnecting means more
possible damage, consumes time, and makes numbers less useful.
Measure both before and as the power switch is pressed - regardless of
whether computer and fan appears to start.

Hopefully the AC power cord was disconnected from wall receptacle
before disconnecting anything. Otherwise more damage may have been
created. Another reason why repair starts by disconnecting or
removing anything.

Purple wire is OK - the only useful number being when supply was
connected to motherboard.

Green wire at 0.78 volts is too high - a defect. But apparenly the
power supply is turning on anyway as indicated by voltages on red,
orange, and yellow wires? (An example of how a defective power supply
may still power on a computer.)

Gray wire voltage is OK when power supply is on. However not
provided is the same number before power switch is pressed and what
happens as power switch is pressed. A delay should have been noticed
between power on and voltage rising about 2.4 volts. Moving on.

Important are what those numbers do and do not do when power switch
is pressed. Does not matter if computer does nothing when swtich is
pressed. Important was what each wire does before and as switch is
pressed. For example - purple wire was at 5.1 volts before switch was
pressed. What was it as switch was pressed?

If power supply does nothing when switch is pressed, then where did
the 3.43, 5.15, and 12.1 come from? Were these when the power supply
was connected to motherboard and only when and after power switch was
pressed? I am a little concerned with the 3.3 volt number - it is
border line high. Not excessive, but very close to exceeding
maximum. Would not cause a computer failure, but a useful fact that
may or may not be useful with other information. Currently only a
concern and not a problem.

Assuming these numbers were all when everything (including video
card and disk drive) were connected, then the power supply is good
except for the green wire. Power supply was telling computer to power
on. Except for green wire, the power supply system is OK.

That green wire problem could be either a motherboard defect or a
power supply defect. Another test later can determine which is
marginal (defective). This defect could cause a system to rarely
(intermittently) shutoff for no apparent reason.

Before moving on, first confirm for me the assumptions I made here
to claim the power supply as (mostly) good.

 

[bud--]

. http://www.nist.gov/public_affairs/practiceguides/surgesfnl.pdf

Bud promotes for plug-in protector manufacturers. He is not honest
enough to admit it.

To quote w_ "It is an old political trick. When facts cannot be
challenged technically, then attack the messenger." My only association
with surge protectors is I have some.

w_ has never answered if he has/had connections to surge protection
equipment manufacturers?

Bud's citation http://omegaps.com/Lightning Guide_FINALpublishedversion_May051.pdf
even demonstrates what happens when a protector is adjacent to
electronics. Page 42 Figure 8 - the surge is earthed 8000 volts
destructively through the adjacent TV. If the computer was damaged by
a surge, well, look what happens when the protector is not earthed; is
instead too close to electronics - Page 42 Figure 8.

The illustration in the IEEE guide has a surge coming in on a cable
service. There are 2 TVs, one is on a plug-in suppressor. The plug-in
suppressor protects TV1, connected to it.

Without the plug-in suppressor the surge voltage at TV2 is 10,000V. With
the suppressor at TV1 the voltage at TV2 is 8,000V. The plug-in
suppressor at TV1 does not in any way contribute to the damage at TV2.

The point of the illustration for the IEEE, and anyone who can think, is
"to protect TV2, a second multiport protector located at TV2 is required."

w_ says suppressors must only be at the service panel. In this example a
service panel protector would provide absolutely *NO* protection. The
problem is the wire connecting the cable entry block to the power
service ‘ground’ is too long. The IEEE guide says in that case "the only
effective way of protecting the equipment is to use a multiport protector."

Bud hopes you don't read those pamphlets carefully.

If you are interested in surge protection I do hope you will read the
guides carefully - excellent information from reliable sources.

And both guides say plug-in suppressors are effective.

 

[Guest]

OK, so here are some more numbers. Power supply (including 4 pin 12 V
connector) hooked up to MB, graphics card and 1 drive installed.

Power supply switch in ON position. Power supply fan comes on, CPU fan comes
on, graphics card fan comes on, HD starts spinning.

Without touching the case's power switch:

Purple: 5.17 V
Green: 0.78 V
Gray: 5.15 V
Orange: 3.43 V
Red: 5.15 V
Yellow: 12.10 V

Pressing the case's power switch:

Purple: 5.17 V
Green: 0.78 V
Gray: 5.15 V
Orange: 3.43 V
Red: 5.15 V
Yellow: 12.10 V

As you can see, the voltages are the same.

I checked the operation of the case's power switch with the Ohm meter, and
the switch is OK. The same for the reset switch.

In the message that you referred to in your first reply, you say that the
voltage on the green wire should be "well below 0.8 V when the power switch
is pressed". I take it my 0.78 V is not low enough then?

I don't have much more time to test anything today; as I go to work in less
than an hour. However, your thoughts would be very much appreciated (and I
will read them today at work); so please don't hesitate to send me any other
tips about things that I can test; and I'll get to that tomorrow.

Thanks in advance,

Erik

 

[w_tom]

Those voltages listed as "without touching case's power switch" and
"pressing the case's power switch" are measuring same thing. Wires
have voltages even before power switch is pressed: "when computer is
powered off". A turned off computer is still powered. But what are
those voltages when power is supposed to be off - before power switch
is pressed - when green wire is above 2 volts? A secondary question
since we are in pursuit of a primary suspect.

When green wire is below 0.8 volts, then both lists are same: "when
computer is powered on". Sidenote: that green wire voltage should
be well below 0.8 volts - typically 0.2 volts. A secondary
(stability) problem maybe addressed later. Onward in pursuit of the
primary suspect.

Important voltages are fine and perfect. Power supply system
(including the gray wire) is telling computer to operate. But computer
does not.

Normally a tech might plug in a diagnostic board to read a BIOS
error code. An error message might also appear on the monitor. But
your monitor is not displaying anything. Most people have no
diagnostic board. No problem. What is minimally required to power up?

Remove AC power cord. Remove everything from computer except
motherboard, speaker, CPU, power switch, and power supply. That means
no keyboard, no memory, no video controller, no mouse, no drives, etc.
Everything else disconnected from motherboard. Now reconnect AC power
cable.

Before power switch is pressed, that green wire will measure
something above 2 volts. When power switch is pressed, the green wire
will drop to near zero and gray wire will rise above 2.4 volts. Then
motherboard's BIOS (program) should execute. It will find no memory
and beep an error code on the speaker (or beeper soldered to
motherboard).

We knew power switch and power supply system are functional (even
with a marginal green wire voltage) from previous and current
measurements. If speaker does not beep, then suspect only motherboard
or CPU. CPUs are almost never damaged since so many motherboard
functions protect it. CPU damage typically means motherboard caused
its failure. Primary suspect identified.

If speaker does beep, then motherboard and CPU are functional.
Suspects are those other peripherals - a most likely suspect being the
video controller.

Remove AC power cord. Install only the video controller. Reconnect
power cord. Perform same test. Does speaker still beep? If not,
primary suspect identified. If yes, repeat for other peripherals (but
no memory) until something causes that speaker to not beep.

 

[w_tom]

And both guides say plug-in suppressors are effective.

Both Bud's guides also say a protector without that earthing
connection cannot *divert it to ground*. Page 6 (Adobe page 8 of 24)
from Bud's NIST citation:

What these protective devices do is neither suppress
nor arrest a surge, but simply divert it to ground,
where it can do no harm.

Will Bud's plug-in protector stop or absorb surges? It must. Where
does that surge energy go? Surge energy must be connected to earth
ground "where it can do no harm". What does a not earthed protector
do? Page 42 Figure 8: surge earthed 8000 volts destructively through
an adjacent TV. Or the fire threat demonstrated in those 'scary
pictures'? After all, with no earth ground then where does surge
energy get dissipated?
http://www.hanford.gov/rl/?page=556&parent=554
http://www.westwhitelandfire.com/Articles/Surge Protectors.pdf
http://www.ddxg.net/old/surge_protectors.htm
http://www.zerosurge.com/HTML/movs.html :
http://tinyurl.com/3x73ol or
http://www.esdjournal.com/techpapr/Pharr/INVESTIGATING SURGE SUPPRESSOR FIRES.doc
http://www3.cw56.com/news/articles/local/BO63312/

From IEEE Standards, another quote that Bud routinely avoids to
promote ineffective (not earthed) protectors. IEEE Standard 141 (IEEE
makes their recommendations in Standards):

... lightning protection is achieve by the process of interception
of lightning produced surges, diverting them to ground, ...

How does a protector without that 'less than 10 foot' earthing
connection "divert them to ground" ? Bud claims protection from
a surge that typically does not damage. That is called effective
protection?

IEEE Standard 1100 - another place where recommendations exist:

It is important to ensure that low-impedance grounding and
bonding connections exist among the telephone and data
equipment, the ac power system's electrical safety-grounding
system, and the building grounding electrode system. ...
Failure to observe any part of this grounding requirement
may result in hazardous potential being developed between
the telephone (data) equipment and other grounded items ...

How curious. IEEE Standard 1100 describes a plug-in protector on
Page 42 Figure 8: earthing a surge 8000 volts destructively through
the adjacent TV. Bud says that damage is effective protection? No
earth ground means no effective protection.

No wonder responsible manufacturers such as GE, Siemens, Cutler-
Hammer, etc instead provide a 'whole house' protector with that
essential and dedicated earthing wire. Why would one spend $2000 or
$3000 - tens of times more money - for Bud's protectors that do not
even provide effective protection? Because a retail salesman or Bud
recommended it?

Bud claims his protectors will somehow stop or absorb what three
miles of sky could not. With no earthing wire to *divert* the surge,
then a protector must somehow block or absorb it. Somehow that tiny
plug-in protector will stop or block what three miles of sky could
not? Reality: no earth ground means no effective protection. Those
'scary pictures' of current technology plug-in protectors ... that will
somehow absorb all that surge energy.

 

[bud--]

Both Bud's guides also say a protector without that earthing
connection cannot *divert it to ground*. Page 6 (Adobe page 8 of 24)
from Bud's NIST citation:

What does the NIST guide really say about plug-in suppressors?
They are "the easiest solution".

Will Bud's plug-in protector stop or absorb surges? It must.

The IEEE guide explains plug-in suppressors work by CLAMPING the voltage
on all wires (signal and power) to the common ground at the suppressor.
Plug-in suppressors do not work primarily by earthing (or stopping or
absorbing). The guide explains earthing occurs elsewhere. (Read the
guide starting pdf page 40).

What does a not earthed protector
do? Page 42 Figure 8: surge earthed 8000 volts destructively through
an adjacent TV.

The lie repeated.

Or the fire threat demonstrated in those 'scary
pictures'? After all, with no earth ground then where does surge
energy get dissipated?
http://www.hanford.gov/rl/?page=556&parent=554

w_ can't understand his own hanford link. It is about "some older
model" power strips and says overheating was fixed with a revision to
UL1449 that required thermal disconnects. That was 1998. There is no
reason to believe from any of these links that there is a problem with
suppressors produced under the UL standard that has been in effect since
1998.
But with no valid technical arguments all w_ has is pathetic scare tactics.

From IEEE Standards, another quote that Bud routinely avoids to
promote ineffective (not earthed) protectors. IEEE Standard 141 (IEEE
makes their recommendations in Standards):

The IEEE Emerald book ("IEEE Recommended Practice for Powering and
Grounding Sensitive Electronic Equipment"), an IEEE standard, recognizes
plug-in suppressors as an effective protection device. This is the most
appropriate IEEE standard for protecting electronics.

And the IEEE guide, which was published by the IEEE, says plug-in
suppressors are effective.

Bud claims his protectors will somehow stop or absorb what three
miles of sky could not.

w_ is halucinating.
And the IEEE guide explains plug-in suppressors work by clamping, not
stopping or absorbing

Reality: no earth ground means no effective protection.

w_ has a religious belief (immune from challenge) that surge protection
must use earthing. Thus in his view plug-in suppressors (which are not
well earthed) can not possibly work. The IEEE guide explains they work
by clamping, not earthing.

Everyone is in favor of earthing. The only question is whether plug-in
suppressors work. Both the IEEE and NIST guides say plug-in suppressors
are effective. Read the sources.

There are 98,615,938 other web sites, including 13,843,032 by lunatics,
and w_ can't find another lunatic that says plug-in suppressors are NOT
effective. All you have is w_'s opinions based on his religious belief
in earthing.

w_ never answers questions:
- Why do the only 2 examples of protection in the IEEE guide use plug-in
suppressors?
- Why does the NIST guide says plug-in suppressors are "the easiest
solution"?
- How would a service panel suppressor provide any protection in the
IEEE example, pdf page 42?
- Why does the IEEE Emerald book include plug-in suppressors as an
effective surge protection device.
– What are w_’s connections to surge protection equipment manufacturers?

 

[Guest]

Hello again.

First, let me say that it seems we're not exactly on the same page anymore.
For clarification: when I said that the power supply is off (and consequently
the PC), I meant that I turned it off with the switch on the power supply.
Since that is rigged to the mains there are absolutely no voltages going to
the PC.

Since CPU and case fans, drives etc. come on as soon as I flick that mains
switch, when the cables are hooked up to the MB, the only way I can test the
"before pressing case power switch" voltages is by disconnecting the cables
from the MB, because pressing the case's ON/OFF button doesn't do a thing. I
gave you those readings in my first voltage test; here they are again:

The purple wire measures 5.20 V. (After hooking up the ATX connector to the
MB the voltage goes down to about 5.10 V.)
The green wire measures 4.66 V. (when it's plugged into the MB it goes to
0.78 V.)
The gray wire measures 5.15 V. (and reads the same after I plug it in to the
MB)

Before power switch is pressed, that green wire will measure
something above 2 volts. When power switch is pressed, the green wire
will drop to near zero and gray wire will rise above 2.4 volts.

Again, pressing the case's power switch does nothing. System gets powered
right away after AC is applied.

Then
motherboard's BIOS (program) should execute. It will find no memory
and beep an error code on the speaker (or beeper soldered to
motherboard).

No beeps heard...

We knew power switch and power supply system are functional (even
with a marginal green wire voltage) from previous and current
measurements. If speaker does not beep, then suspect only motherboard
or CPU. CPUs are almost never damaged since so many motherboard
functions protect it. CPU damage typically means motherboard caused
its failure. Primary suspect identified.

In other words: the motherboard would be defective?

If speaker does beep, then motherboard and CPU are functional.
Suspects are those other peripherals - a most likely suspect being the
video controller.

Remove AC power cord. Install only the video controller. Reconnect
power cord. Perform same test. Does speaker still beep? If not,
primary suspect identified. If yes, repeat for other peripherals (but
no memory) until something causes that speaker to not beep.

I skipped the last 2 paragraphs, as they don't seem to apply (I get no beeps
at all).

Any other suggestions?

Thanks,

Erik

 

[w_tom]

First, let me say that it seems we're not exactly on the same page anymore.
For clarification: when I said that the power supply is off (and consequently
the PC), I meant that I turned it off with the switch on the power supply.
Since that is rigged to the mains there are absolutely no voltages going to
the PC.

Since CPU and case fans, drives etc. come on as soon as I flick that mains
switch, when the cables are hooked up to the MB, the only way I can test the
"before pressing case power switch" voltages is by disconnecting the cables
from the MB, because pressing the case's ON/OFF button doesn't do a thing. I
gave you those readings in my first voltage test; here they are again:

That switch mounted on the power supply must never be changed.
Computer must power on and power off with that switch unchanged.
Power supply controller always needs the 5VSB from purple wire so that
it can turn on and turn off power supply. Power supply controller
controls power supply using the green wire. Power supply controller
is told to turn on power supply when front panel switch is pressed for
less than a second.

Apparently, power supply controller on motherboard is defective. It
tells power supply to be powered on constantly. Three options. First
is to trace green wire connector, along printed circuit trace, to the
chip that tells power supply to power on. Find something that is
shorting printed circuit trace causing power supply to remain on
constantly. Suspects include a metallic standoff that is somehow
shorting to a printed circuit trace.

Second option: replace motherboard because power supply controller
circuit is defective. More likely, something like a semiconductor for
that controller circuit has failed. Only solution to a failed
semiconductcor is a new motherboard.
If a mechanical short to the motherboard printed circuits is not
found, then motherboard replacement is the only solution.

Third option - the bios is somehow set to always power on power
supply (extremely unlikely especially if CMOS is reset). A computer
must execute the BIOS program to change that setting. Without a
working motherboard (as also indicated by no beeps), then that option
cannot be executed. Third option only mentioned to provide a complete
answer.

Primary suspect is motherboard.

 

[Guest]

Thanks, that sounds like a good explanation.

I will look for shorts tomorrow, but I'm afraid that may be a long shot.
Remember, I had two PCs fail, and what are the chances that both of them
failed because of a short. It's not like I have them installed in a metal
work shop, filled with metal filings

So I suppose we're going to have to shop around for at least one replacement
motherboard...

Thanks for all your time and help; I really appreciate it!

Erik

 

[w_tom]

Thanks, that sounds like a good explanation.

I will look for shorts tomorrow, but I'm afraid that may be a long shot.
Remember, I had two PCs fail, and what are the chances that both of them
failed because of a short.

A short is not a likely reason for damage. Short typically only
creates a temporary failure. Digital IC outputs can be completely
shorted to ground and still not be damaged as was standard even before
CMOS transistors - when ICs were made with bipolar transistors.
Permanent failures are created by manufacturing defects that may
appear months or years later, static electric discharges, and other
human mistakes such as removing or disconnecting something with power
cord still connected to AC. Heat, shorts, and other things so often
blamed in myths typically do not create permanent damage.

If this diagnosis had proceeded without part removal or swapping,
then a list of possible problems would have been significantly
limited. Symptoms observed before complete failure occurred would
have merit. Currently, the motherboard could have two or more
problems since part swapping has a bad habit of creating new problems.
With so many additional possibilities, then the original intermittent
symptoms provide too little useful information.

Sometimes a defective part tested elsewhere can cause that
elsewhere machine to become defective. When constructing
$multimillion custom computer controlled workstations, part swapping
between machines to find a problem was forbodden - in one case almost
causing employment termination. Once one defect becomes two, then
the problem becomes exponentially more complex.

Same procedure should be executed on the second machine without
swapping or disconnecting any parts. IOW, like in CSI, "follow the
evidence". And do not shotgun. Perform the same diagnostic procedure
on the second machine.

Returning to the original machine. CPU would not execute BIOS. But
CPU will not even clock a single machine cycle until the power supply
controller issues a command that says, "OK - ready to work". Power
supply controller is known to have at least one defective. Quite
possible is that motherboard failure is a single point problem limited
to power supply controller. Therefore CPU also would never get
permission to execute software.

One step at a time for that first system. Replace motherboard
(which might mean a new CPU). Then establish that other components
are working by first confirming those power supply voltages are
correct. With correct voltages, then power supply system works -
something accomplished. With only motherboard, speaker, power switch,
CPU, power supply, and nothing else, then motherboard should beep.
Another accomplishment: CPU, some motherboard functions, etc are
working. Add memory and video controller. More accomplished if
BIOS displays a video message. Slowly removing everything from the
'unknown' state.

Repeat same for the second system. Slowly get each component or
subsystem moved to 'definitively good' or definitively' bad. Do as on
CSI - "follow the evidence". Break that massive unknown second
computer into components "definitively known good" or "definitively
bad". It may appear tedious. But each step or test means
accomplishment every time a component is removed from an "unknown"
state.

Break each computer down into parts, then analyze only each part to
get an answer that includes the word 'definitively'. IOW do to the
second computer what we did to the first.