PC dead but MOBO light on ???????

[p.mc]

Hi Tom

A power supply that fails to provide sufficient voltage (or other
problems) will light the light, light the mains tester, and still be
defective. As noted previously, there is no cheaper and effective
solution that a 3.5 digit multimeter.

**(Is this what you were refering to)**
http://www.powercenta.co.uk/storefr....aspx?sfid=79410&i=198797736&mpid=4033&dfid=1

If so I'll buy the tester for future problems. As I said earlier I did test
for continuity across the power switch with my voltage testers which was
easy enough, but also they measure between12v and 600v and to be honest I
didn't know what pins to test between (20 pin mains) to determine if there
was a supply for the MOBO and peripherals.

Also bad is to test a power supply removed from the system. Again, a
supply that is defective can even measure correct voltages with a meter
when not connected to a large load (ie motherboard). Best test of a
power supply is to test it in the computer AND with maximum load -
motherboard and peripherals.

Now if I use the 3.5 digit multimeter, lets say I'm testing a 4 pin
connector to my HDD, the cables being;
1x red.......5v
2x black....common
1x yellow...+12v
**This is using this chart for reference:
http://www.ee.surrey.ac.uk/Workshop/advice/conns/connPSU.html

Would I put one end of the tester into a common and with the other
alternating from red to yellow to see if there is a 5v and +12v supply?
What do the "black" cables do, are they like the -negative on a battery
drawing the power through?
I thought "common" meant feed!

 

[p.mc]

I can't QUITE tell for sure (may be a trick of the light), but it
appears as if one or both of the yellow capacitors near the processor
has a VERY slight bulge to it. That would mean a bad board.

Hhmm!! they do look slightly raised. I think I'll see if that's the case
when I buy a 3.5 digit multimeter.

"Thanks Paul"

Regards
p.mc

 

[p.mc]

Hi Tom

I'll buy a 3.5 digit multimeter hopefully tomorrow and do as instructed
either tomorrow or Monday.
I'll put back and connect all hardware and start the test and post the
outcome. Unfortunately I didn't record event log, I did write down the BSOD
but mislaid it, presuming that the DVD/RW was at fault I probably binned the
info.

"Thanks to all who replied" I'm sure this thread will benefit others who
read this very instructive and informative thread.

On your info... "Don't buy a power supply only on price and watts."
Well I did want to replace it with the same PS thinking the manufacturer
"Evesham" knows best,but I think they're becoming obsolete. I also read that
a 400-450watt PS was better suited to the XP platform!! so if I do need one
I was going for this model:

http://cgi.ebay.co.uk/ws/eBayISAPI....EWA:IT&viewitem=&item=130063996276&rd=1&rd=1I must admit I chose it on price and wattage, but to be honest, besideschanging like for like I wouldnt have a clue what to buy.*...WinXP Home SP1*...ASUS A7A266 MOBO*...2x DVD/RW*...1x DVD ROM*...1x Creative Soundblaster live 5.1*...1x Nvidia Graphics*...1x Adaptech ASH 1205 SA Host Controller PCI Card*...1x Sata 180gb Maxtor HDD*...1x Sata 300gb Maxtor HDDModen & Floppy--Regardsp.mc"w_tom" <[email protected]> wrote in message In but a minute with a 3.5 digit multimeter, trim the list of> suspects down to only power supply or motherboard, or ..... That mains> testing light cannot report anything useful because numbers are> required.>> Light says power cord connects to AC mains. Light does not say any> voltage is sufficient. Only meter numbers can say that. Start by> measuring voltage on purple wire by pressing meter probes into 20 pin> nylon connector for purple wire and to any black wire (or touch> chassis). With computer powered off (and light still illuminated), that> number must exceed 4.87 (and it probably will). Next take the number> for green wire both before and when power switch is pressed. That> number should be above 2.0 volts and drop to less than 0.8 volts when> switch is pressed. What happens here is very informative.>> Next, the gray wire. It must rise to well above 2.4 volts within> second of the power switch pressed. I suspect either voltage will not> move or will rise and then fall.>> Now measure any red, orange and yellow wires when switch is pressed.> Assuming the green wire voltage did drop, then each voltage here> should rise up (and then will probably fall). Any voltage that does> not rise up, then look for a short circuit for that voltage on> motherboard. If all voltages do rise up, then within a second, each> voltage must achieve more than 3.23, 4.87, or 11.7.>> Information from those quick measurements either make it obvious> which is the suspect or post those numbers here for better analysis.>> Rather than 'buy this and replace that', a best solution is to> first take numbers. Just another reason why a 15 quid meter is a tool> as important as a screwdriver, as complex as an Ipod, and the> hands-down quickest answer of your question and to identify where to> start looking for a solution.>> Nice pictures. But visual inspection almost never identifies a> failure. Without that meter (your only useful eyes), then replace> possibly good parts until maybe something is fixed. It is called> shotgunning. Without that meter, you are essentially blind.>> Had you been recording numbers and message in the BSOD, or reviewed> information in the system (event) logs, then a suspect might have been> more quickly identified. Too late now.>> A simple rule for power supplies: if manufacturer does not provide a> comprehensive list of numerical specs, well, this is how supplies> missing essential functions are dumped in a market of electrically> naive computer assemblers. A power supply must contain a long list of> functions that were required even 30 years ago. But many supplies sell> for less (and for higher profits) by forgetting to include many> essential functions. Then some assmbler says "computer powers on;> therefore it must be fine". It's not. But the naive may need other> more expensive solutions (such has a UPS) to solve what would not exist> if using a properly constructed supply. Don't buy a power supply only> on price and watts.>> p.mc wrote:>> As I wrote in my first post, I powered the MOBO with its 21 pin supply,and>> also the CPU and tower fans were connected to the MOBO, but on pressingthe>> power button the fans didn't kick in.>>>> I'd be over the moon if the power supply was the culprit, as you said the>> unit may not be fully functional.>> I'm tempted to buy a new P.Supply unit but knowing my luck it'll be the>> MOBO... (>>>> I do have some mains testers that will show a 12v supply, but when I>> blunderingly tried poking the 2 terminals into any of the connections ofthe>> HDD feeds and others (Whilst the board was powered) I only got acontinuity>> light but as I said I'm no expert and probably not using them properly!!>

 

[Leythos]

Hhmm!! they do look slightly raised. I think I'll see if that's the case
when I buy a 3.5 digit multimeter.

If this is an older machine, you don't need a multimeter, if the caps
are bulged then your problem has already been identified. The larger
caps should have a X on the top of them, it should be perfectly flat or
a little sunken in some designs, if it's raised at all then you've got
the curse of the stolen cap formula.

I had a customer with a Dell workstation that had worked for many years,
having only been powered off a couple times, and when they came back
from Christmas vacation it would only randomly turn on - having spare
PSU, Drives, etc.... It was quicker to put the spares in than dig
through all the wiring, than to use a multimeter, etc... Nothing worked,
so, with all parts disconnected or replaced (memory and PSU), I noticed
7 caps that had bulged, and sure enough, when I replaced those the
system fired right up. Keep in mind, replacing CAPS on a multilayer
board is not for the novice, and you need a quality soldering iron and a
good solder sucker.

 

[databaseben]

i don't think thats a good idea because:

1) if you dont have any training with electronics
and the equipment. So you can seriously hurt yourself
by electricity.

2) your mobo was designed prior to 2001
and outdated. Neither your mobo and p.s.
can handle new devices like dvd's, etc....

you might succeed with simply buying
a new p.s. if that is the problem. but
it will likely overload your old mobo
and you will find yourself in the same pickle.

of course, you can do as you wish. But don't
beleive yourself to be compentant with
diagnosing equipment, especially out dated
and ready for the dumpster electronics.

I have reason to beleive that if you were
then you would obtain the schematics for
the mobo and for the p.s., have experience
with multimeters and o-scopes, solder, soldering
irons, etc....

do yourself a favor and save time, money and possibly your own life by
getting a cheaply priced pc, that is
more powerful than your old 500 mhz pc.

good luck....

 

[w_tom]

Excellent price on a 3.5 digit meter that does all functions
required.

Power switch tells power supply controller (on motherboard) to turn
on power. Power supply controller needs that switch connection and
sufficient voltage on the purple wire to respond. Power supply
controller uses green wire to order power supply on. First two wire
voltages checked are purple wire (before switch is pressed) and green
wire (before and when switch is pressed). With proper voltages, both
switch and power supply controller were confirmed - and without
disconnecting any wires. This paragraph simply summarizes why purple
and green wires are measured in that previous post. All that
confirmation in two quick measurements.

Purple wire is power for power supply controller. Red, orange, and
yellow wires are power for rest of motherboard. If power supply
controller is telling power supply to turn on (green wire), then red,
orange, and yellow wires will provide power. Power supply controller
is told when power supply is OK via gray wire.

You need not measure power at disk drive. Those same voltages are
measured in that 20 pin connector between power supply and motherboard.

Black wire is common (or negative side of battery). Switch meter
selector to DC volts. Connect black meter probe to any black wire
contact inside the 20 pin nylon connector or connect black meter probe
to computer chassis. Then touch red meter probe to each voltage
(purple, green, gray, orange, yellow, and red wires) to get a voltage
number.

No reason to measure voltages on disk drive connector. Those same
red and yellow wires are also the red and yellow wires that connect
power supply directly to motherboard.

What I believe: power supply controller (green wire) is not telling
power supply to power on. If true, then one simple measurement where
power supply switch connects to motherboard would provide a final
puzzle piece. But first, let's see the numbers.

http://www.ee.surrey.ac.uk/Workshop/advice/conns/connPSU.html are
pinouts of the 20 pin nylon connector between power supply and
motherboard. Purple wire is often called +5VSB or Five volt standby
power. Gray wire is often called Power OK. Green wire is often called
PowerOn (with word overlined), !PowerOn, or PowerOn#. Each means power
turns on when voltage goes to less than 0.8 volts.

In each case, just any number is not sufficient. Voltages must
exceed minimum values. Lights cannot report that. Testing a power
supply removed from a system often will not detect a lower voltage that
only occurs when connected to motherboard.

 

[Leythos]

i don't think thats a good idea because:

1) if you dont have any training with electronics
and the equipment. So you can seriously hurt yourself
by electricity.

Yes, but most power in the computer, that you can reach, is 12v or less,
and unless you're overly sensitive, there is little chance to hurt
yourself. You would have to be playing with an older power switch that
actually switched the 120V line, and that not been used for a LONG time.

2) your mobo was designed prior to 2001
and outdated. Neither your mobo and p.s.
can handle new devices like dvd's, etc....

You don't know that, I have several P1's and P2's that work just fine
with the new DVD readers.

you might succeed with simply buying
a new p.s. if that is the problem. but
it will likely overload your old mobo
and you will find yourself in the same pickle.

You can't overload the motherboard.

[snip]

do yourself a favor and save time, money and possibly your own life by
getting a cheaply priced pc, that is
more powerful than your old 500 mhz pc.

As it appears to be a bad CAP (swelling) problem, I agree, Dell has
several under $400 that would work well.

 

[databaseben]

the following is a new pc on the low
end. And if you compare the specifications
to what you have now, you will see that it
is much more powerful.

Your current cpu is rated at 500 mhz, that
is if it is not actually a 250mhz that has a clock doubler.

However, the new one that is on the low end
of the market is over 3 ghz. (yours is .5 ghz)

http://www.tigerdirect.com/applications/SearchTools/item-details.asp?EdpNo=1221657&sku=BTO 970811

 

[w_tom]

i don't think thats a good idea because:
...

1) Anything dangerous in a power supply is inside a box so that human
cannot touch it. Only precaution (as demonstrated by voltage on purple
wire) is to pull AC mains power cord before connecting or disconnecting
components inside a computer.

2) Power supply will not and cannot overload motherboard. That is even
known to electronic project hobbyists.

3) Those with basic computer experience know schematics are not
available for motherboards.

4) If your 2001 motherboard is ready for the dumpster, then why am I
still using early 1990 computer on the internet (not all internet
functions, but even they work just fine)?

Take some measurements. Know what is and is not working. And finish
having learned something useful for the future. Too many computer
experts even fear the meter - even believe myths that a power supply
can "overload your old mobo".

 

[databaseben]

unfortunately, your
foolishness doesn't impress
me.

And if the o.p.
gets hurt, then her survivors
as well as the authorities
will come after all the persons
involved...

i don't think thats a good idea because:

1) if you dont have any training with electronics
and the equipment. So you can seriously hurt yourself
by electricity.

Yes, but most power in the computer, that you can reach, is 12v or less,
and unless you're overly sensitive, there is little chance to hurt
yourself. You would have to be playing with an older power switch that
actually switched the 120V line, and that not been used for a LONG time.

2) your mobo was designed prior to 2001
and outdated. Neither your mobo and p.s.
can handle new devices like dvd's, etc....

You don't know that, I have several P1's and P2's that work just fine
with the new DVD readers.

you might succeed with simply buying
a new p.s. if that is the problem. but
it will likely overload your old mobo
and you will find yourself in the same pickle.

You can't overload the motherboard.

[snip]

do yourself a favor and save time, money and possibly your own life by
getting a cheaply priced pc, that is
more powerful than your old 500 mhz pc.

As it appears to be a bad CAP (swelling) problem, I agree, Dell has
several under $400 that would work well.

 

[Shenan Stanley]

<snipped thread>
Want to see the whole thing?

http://groups.google.com/group/micr...05f574907e9?lnk=st&q=&rnum=1#1a5e505f574907e9

I was hoping Ian would reply to confirm whether that was the way to
test a PS, maybe he'll reply tomorrow. As you said I do want to
make sure that the PS is 100% faulty, and also you were right I
didn't mention I had the 2 pin power switch connected to the board,
I did test the switch with my mains testers and it was working.

<snip>

I hate to point this out, but...

There has been more talk on this than I would have ever given to even a new
system - given the component mostly being talked about ranges from $33-$120
on a bad day. =)

Not to mention you could go get one at a local 'big name' store - bring it
home, plug it in for 30 seconds to see if the thing powers up - if it
does... You're good to go.. If it doesn't - return the thing and be out no
more than the trip and time.

Speaking from a time is money point of view - the time invested in this has
greatly exceeded the cost of any power supply you would have gotten. And
with systems that offer 1.8GHz with 512MB memory/100GB hard disk drives
going for well under $400 with a 15 month warranty (less than $300 after
rebates) -- I think you might be putting way too much time into a dead
horse. ;=)

 

[Leythos]

unfortunately, your
foolishness doesn't impress
me.

And if the o.p.
gets hurt, then her survivors
as well as the authorities
will come after all the persons
involved...

My guess would be that you've not worked inside any computers for a very
long time, if ever, or you would not have made the comments above.

In the last 10 years, and I've been doing this since the 70's, I've not
seen a single workstation that has any high-voltage power that the user
can be exposed to, outside the power supply case. On a couple servers,
years ago, I found a power switch that had 120v on it, but even that was
a bad design replaced the next model.

As the OP has already stated, swelling cap's, it's a done deal - new
motherboard or a new computer.

 

[Shenan Stanley]

1) Anything dangerous in a power supply is inside a box so that
human cannot touch it. Only precaution (as demonstrated by voltage
on purple wire) is to pull AC mains power cord before connecting or
disconnecting components inside a computer.

Humans make mistakes - it's part of being human.

2) Power supply will not and cannot overload motherboard. That is
even known to electronic project hobbyists.

Okay. Probably true.

3) Those with basic computer experience know schematics are not
available for motherboards.

Most.. Most...

4) If your 2001 motherboard is ready for the dumpster, then why am I
still using early 1990 computer on the internet (not all internet
functions, but even they work just fine)?

Because you are stubborn and are happy with whatever you have or because you
don't have the money to upgrade.

No one said it wouldn't work - there is a time vs. money factor in anything.
I can spend 2 weeks getting a $35 part to work or I can replace it in 30
minutes (if I have a long drive to get the part.)

Take some measurements. Know what is and is not working. And
finish having learned something useful for the future. Too many
computer experts even fear the meter - even believe myths that a
power supply can "overload your old mobo".

If you are not an electrician and don't strive to be an electrician or
electrical hobbyist - then there is no point in learning to be one. While I
am the first to say, "You'll never learn any younger" and "If you don't try,
you'll never know" and "Teach a man to fish..." - there are some things you
just have to weigh the benefits on.

$35 part you could get at a big name store and return if it turned out not
to resolve your problem versus all the testing and jazz that does the same
thing as buying the $35 working part and plugging it in -- tells you whether
the power supply is the problem or not.

Each their own... If the OP enjoys testing and fiddling with such things -
fantastic - go on. However - there are other options - faster and possibly
less expensive in the long run. Presenting those options is a valid point -
just as presenting the testing options is a valid point.

 

[databaseben]

that is correct.
you don't know how much i know
nor do i care how much you know.

the point is, you should know better
than to encourage or manipulate anyone who is not
properly trained, to do what we are able
to.

unless you or any other respondant has
direct oversight of the o.p and the equipment,
don't think that there cannot be any consequences
from the ill faded advices being provided...

THERE WILL BE NO FURTHER DISCUSSION
ON THIS MATTER FROM ME.

 

[w_tom]

$35 part you could get at a big name store and return if it turned out not
to resolve your problem versus all the testing and jazz that does the same
thing as buying the $35 working part and plugging it in -- tells you whether
the power supply is the problem or not.

If you are paying $35 for the power supply, well, there is part of
the problem right there. Many know so little about electricity as to
confuse simple hobbyist knowledge with being a licensed electrician.
Much about electricity is common knowledge. One fact that any computer
assemble should know. To sell a power supply for $35, the manufacturer
'forgets' to include necessary functions. But then they are not
selling to informed customers. Informed customers also want to see a
long, numeric spec sheet that says the power supply does contain these
functions.

Instead those who also fear to use a meter and who shotgun, also buy
power supplies only on watts and dollars. Then they need more
equipment (ie UPS) to fix what was missing inside that power supply.

The power supply is inside a box sealed with screws and warning
labels. Yes, even the sky could fall as Chicken Little proclaimed.
But a power supply is not dangerous. Many do not first learn
technology before hyping fears and recommendations. Many even fear a
six quid meter - but would spend only $35 (full retail) on a power
supply. Somehow that is called 'smarter'.

 

[Shenan Stanley]

If you are paying $35 for the power supply, well, there is part of
the problem right there. Many know so little about electricity as
to confuse simple hobbyist knowledge with being a licensed
electrician. Much about electricity is common knowledge. One fact
that any computer assemble should know. To sell a power supply for
$35, the manufacturer 'forgets' to include necessary functions.
But then they are not selling to informed customers. Informed
customers also want to see a long, numeric spec sheet that says the
power supply does contain these functions.

Instead those who also fear to use a meter and who shotgun, also
buy power supplies only on watts and dollars. Then they need more
equipment (ie UPS) to fix what was missing inside that power supply.

The power supply is inside a box sealed with screws and warning
labels. Yes, even the sky could fall as Chicken Little proclaimed.
But a power supply is not dangerous. Many do not first learn
technology before hyping fears and recommendations. Many even fear
a six quid meter - but would spend only $35 (full retail) on a power
supply. Somehow that is called 'smarter'.

$35 to test - if it works - you have found the problem. Return it and get a
$120 coolermaster power supply after that if you wish.

I don't know about you - but I get paid more than $35/hour net.
Not everyone makes that much - that's a given - which would make it even
more logical to spend as little time fiddling with it and more time doing
something pproductive to make money off of.

This problem has been going on for almost 2 days. That's a lot more than
even the most expensive power supply you could buy for a normal home
computer system (ATX). Cut out all you want, recommend all you want - I
argued nothing electrical with you - just common sense.

 

[McSpreader]

... each voltage must achieve more than 3.23, 4.87, or 11.7.

Wrong, wrong, wrong!

Just for the record: In the ATX PSU spec from Intel ATX Design Guide*
you'll find the minimum values for the 3.3V, 5V and 12V supplies are
quoted as 3.14V, 4.75V and 11.4v, respectively.

*Section 3.2.1 of <www.formfactors.org/developer/specs/ATX12V_PSDG_2_
2_public_br2.pdf>

w_tom still quotes these incorrect values for reasons he can't/won't
explain or back up with facts and references.

 

[Leythos]

THERE WILL BE NO FURTHER DISCUSSION
ON THIS MATTER FROM ME.

You should take your medication and apply the above to all Usenet
threads.

In case you missed it, I suggested that the motherboard was a loss and
there is no need to test anything, as the swelled CAP's always indicate
a loss of the board.

 

[Uncle Grumpy]

THERE WILL BE NO FURTHER DISCUSSION
ON THIS MATTER FROM ME.

I didn't think it would be so easy to get you to shut the f*** up!

 

[p.mc]

<snipped thread>
Want to see the whole thing?

http://groups.google.com/group/micr...05f574907e9?lnk=st&q=&rnum=1#1a5e505f574907e9



<snip>

I hate to point this out, but...

There has been more talk on this than I would have ever given to even a
new system - given the component mostly being talked about ranges from
$33-$120 on a bad day. =)

Not to mention you could go get one at a local 'big name' store - bring it
home, plug it in for 30 seconds to see if the thing powers up - if it
does... You're good to go.. If it doesn't - return the thing and be out no
more than the trip and time.

Speaking from a time is money point of view - the time invested in this
has greatly exceeded the cost of any power supply you would have gotten.
And with systems that offer 1.8GHz with 512MB memory/100GB hard disk
drives going for well under $400 with a 15 month warranty (less than $300
after rebates) -- I think you might be putting way too much time into a
dead horse. ;=)

To be honest you may be right, but this is not just about saving a few bob
now, I am genuinely interested in the fault finding aspect.