New Power Supply but no power

[Zentraleinheit]

When my computer would nolonger turn on I thought my 236W PS died.

So, I got a 550W replacment.

But, after connected it to my DFI P2XBL motherboard, it still won't turn on
(and yes the outlet I used has power)?.

The best I am able to get is the green power LED to light if I use the on
switch on back of the PS. And a faint flicker of the kitchen lights.

What in the heck is wrong

I disconnect everthing, I even pull the CPU, but still no power (or sparks,
smoke, strange noises).

Am I doing something wrong? Is the CPU, MB, or even the switch dead?

And did it matter that I installed the PS "upside down" with the second fan
pointing down insted of up (it was the only way that the screws would line
up)?

Please help

 

[badgolferman]

When my computer would nolonger turn on I thought my 236W PS died.

So, I got a 550W replacment.

But, after connected it to my DFI P2XBL motherboard, it still won't
turn on (and yes the outlet I used has power)?.

The best I am able to get is the green power LED to light if I use
the on switch on back of the PS. And a faint flicker of the kitchen
lights.

What in the heck is wrong

I disconnect everthing, I even pull the CPU, but still no power (or
sparks, smoke, strange noises).

Am I doing something wrong? Is the CPU, MB, or even the switch dead?

And did it matter that I installed the PS "upside down" with the
second fan pointing down insted of up (it was the only way that the
screws would line up)?

Please help

I'm sorry, but your lights shouldn't be flickering when you turn on the
computer -- it just doesn't take that much power. There is either
something wrong with the power outlet, power cord, or motherboard. You
can check the switch with a multimeter or just short the Power ON pins
on the front panel header of the motherboard momentarily.

 

[kony]

When my computer would nolonger turn on I thought my 236W PS died.

So, I got a 550W replacment.

But, after connected it to my DFI P2XBL motherboard, it still won't turn on
(and yes the outlet I used has power)?.

The best I am able to get is the green power LED to light if I use the on
switch on back of the PS. And a faint flicker of the kitchen lights.

Do not use the power supply switch to turn it on. That
switch should always stay on.

Use the case front-panel switch. If that doesn't work,
disconnect AC, clear the motherboard CMOS, then reconnect AC
and retry the front case switch.

What in the heck is wrong

Likely the motherboard is nonfunctional. With a system that
age it might be fouled electrical contacts, burst capacitors
(examine all caps for swelling, cracked tops or leaky
residue), or even something like a fan failure on CPU
heatsink (or video?). Examine everything... all fans should
be turning even if the system won't POST, doesn't produce
video output.

I disconnect everthing, I even pull the CPU, but still no power (or sparks,
smoke, strange noises).

Connect only the following:
CPU, CPU heatsink/fan, 1 memory module, video
Clear CMOS
Turn on the motherboard by shorting the two power-on pins
(on the motherboard front panel header, where the switch
plug goes) together.

All other hardware- drives, keyboard, other cards, etc, can
and should be disconnected.

Am I doing something wrong? Is the CPU, MB, or even the switch dead?

Unless the heatsink fan is frozen up, odds are it isn't the
CPU. Are you sure the power supply's voltage switch is set
correct per your location (110 vs 220 volts)?

If this was a generic (new) power supply, there is an
outside chance it is defective. It would be good to try a
known working and viable power supply.

And did it matter that I installed the PS "upside down" with the second fan
pointing down insted of up (it was the only way that the screws would line
up)?

In a standard ATX case, the fan on the side is supposed to
be pointing down, not up. Ignore pictures that suggest
otherwise, they just turn the PSU upside down so the fan is
included in the picture.

If i were guessing, I'd guess that the capacitors around the
CPU slot have failed. That guess could easily be wrong.
System is now at an age where it may not be worth putting a
lot of $ or time into it if it still won't work. If the
power supply is generic, I would return it to get money back
now, as a generic 500W may not be suitable for a modern
system if you were to try to reuse it. "Some" generic 500W
can be suitable, but there is a large variability in the
quality, it is a gamble.

 

[Paul]

I'm sorry, but your lights shouldn't be flickering when you turn on the
computer -- it just doesn't take that much power. There is either
something wrong with the power outlet, power cord, or motherboard. You
can check the switch with a multimeter or just short the Power ON pins
on the front panel header of the motherboard momentarily.

It is perfectly OK for the lights to flicker. That flickering
is a result of the "inrush current" when the main capacitors
in the ATX power supply are charged from the line. It takes a very
short time to charge them, but a lot of current can flow during that
interval.

There is a schematic for an (old) ATX power supply here:

http://www.pavouk.comp.cz/hw/en_atxps.html

C5 and C6 (470uF) are the main caps. This example power supply doesn't
show a power switch in the circuit, so the power supply is
ready as soon as it is plugged in. The component "NTCR1" appears
to be a component added as an inrush current limiter, as without
some resistance in the circuit when the power supply first
makes contact with the mains, the bridge rectifier diodes
(D21,D22,D23,D24) would be blown. The fuse F1 doesn't blow,
because the inrush current interval is so short.

You should not see any flickering when pushing the power
button on the front of the computer case, because the amount of
power the computer uses is small compared to the power capability
of the house wiring feeding it (i.e. 2 amps on a 15 amp circuit).
When you push the power button on the front of the computer, the
main capacitors are already charged, so no additional flickering
should be seen at that instant.

For the power supply to come on, the PS_ON# signal must be
pulled to the same potential as one of the COM pins. Those
two signals are on the main 20 or 24 pin cable of the ATX
power supply. (In the sample schematic above, the PS_ON# signal
is on the left hand side of the drawing, and says "0V RUN
+5V STOP" next to the signal.) The motherboard converts the
momentary contact seen on the power switch on the front of
the computer, into a steady level on the PS_ON# signal. If
the motherboard logic has a problem, that can prevent the user's
command to start, from making PS_ON# work.

Paul

 

[Ken Maltby]

It is perfectly OK for the lights to flicker. That flickering
is a result of the "inrush current" when the main capacitors
in the ATX power supply are charged from the line. It takes a very
short time to charge them, but a lot of current can flow during that
interval.

There is a schematic for an (old) ATX power supply here:

http://www.pavouk.comp.cz/hw/en_atxps.html

C5 and C6 (470uF) are the main caps. This example power supply doesn't
show a power switch in the circuit, so the power supply is
ready as soon as it is plugged in. The component "NTCR1" appears
to be a component added as an inrush current limiter, as without
some resistance in the circuit when the power supply first
makes contact with the mains, the bridge rectifier diodes
(D21,D22,D23,D24) would be blown. The fuse F1 doesn't blow,
because the inrush current interval is so short.

You should not see any flickering when pushing the power
button on the front of the computer case, because the amount of
power the computer uses is small compared to the power capability
of the house wiring feeding it (i.e. 2 amps on a 15 amp circuit).
When you push the power button on the front of the computer, the
main capacitors are already charged, so no additional flickering
should be seen at that instant.

For the power supply to come on, the PS_ON# signal must be
pulled to the same potential as one of the COM pins. Those
two signals are on the main 20 or 24 pin cable of the ATX
power supply. (In the sample schematic above, the PS_ON# signal
is on the left hand side of the drawing, and says "0V RUN
+5V STOP" next to the signal.) The motherboard converts the
momentary contact seen on the power switch on the front of
the computer, into a steady level on the PS_ON# signal. If
the motherboard logic has a problem, that can prevent the user's
command to start, from making PS_ON# work.

Paul

Of course if a loose screw or something were shorting the MB
to the case, the lights might flicker. The green light indicates the
+5V that is always on when the PSU is plugged in, and if it has
a switch (on the PSU) that it is turned on. That the PSU is
functional at that point, might indicate that the screw isn't shorting
that circuit. Do you still have the green light after you try to turn
the rest of the computer on? If you don't then I would look for
loose screws. (Or shorts in general.)

Luck;
Ken

 

[w_tom]

You speculated and then tried to fix it rather than first
discover what was defective. Kony defined a list of possible
reasons for failure. Many 'power supply system' problems on
that list would not be corrected by your 'shotgun' solution.
Maybe first learn what is wrong before fixing anything?

First collect facts. The difference between an educated
problem solver verses a wild speculator - the shotgunner.
Some basic concepts: take but 2 minutes to identify the
problem as summarized in previous posts in "Computer doesnt
start at all" in alt.comp.hardware on 10 Jan 2004 at
http://tinyurl.com/2t69q and
"I think my power supply is dead" in alt.comp.hardware on 5
Feb 2004 at
http://www.tinyurl.com/2musa

In your case, a solution begins with numbers from the purple
wire. Then followed by what happens on green wire before and
after power switch is pressed. Yes, numbers. Because without
numbers, informed posters tend to ignore your post. Numbers
- what those who post wild speculation need not obtain.

Its called a power supply 'system'. Which 'system'
component is defective? Wildly changing things will only make
a solution more difficult. Where are the numbers - basic
facts? One makes changes only after important facts are first
learned. Wildly replacing a power supply is but a classic
example of shotgunning; which even a car mechanic will do only
if insufficiently educated.

Fixing your own equipment is how we learn.

 

[kony]

It is perfectly OK for the lights to flicker. That flickering
is a result of the "inrush current" when the main capacitors
in the ATX power supply are charged from the line. It takes a very
short time to charge them, but a lot of current can flow during that
interval.

No, it is not "OK". The amount of current required to
charge 2 x 470 uF caps is relatively trivial, the duration
very brief, it should not be dimming lights enough to be
noteworthy. There is either a problem with the power supply
or the household wiring for this to happen.

 

[Paul]

No, it is not "OK". The amount of current required to
charge 2 x 470 uF caps is relatively trivial, the duration
very brief, it should not be dimming lights enough to be
noteworthy. There is either a problem with the power supply
or the household wiring for this to happen.

On page 3 here, it says:
http://www.antec-inc.com/pdf/manuals/truepower.pdf

INRUSH CURRENT
115V/50A(max.), 230V/80A(max.) at 25°C cold start

That 50A is what makes the lights blink, albeit not
everyone is quick enough to see it. The "25C cold start"
part tells you they are using an NTC resistor to limit
the current, and the colder the supply is when it starts,
the larger the current will be.

Paul

 

[Paul]

On page 3 here, it says:
http://www.antec-inc.com/pdf/manuals/truepower.pdf

INRUSH CURRENT
115V/50A(max.), 230V/80A(max.) at 25°C cold start

That 50A is what makes the lights blink, albeit not
everyone is quick enough to see it. The "25C cold start"
part tells you they are using an NTC resistor to limit
the current, and the colder the supply is when it starts,
the larger the current will be.

Paul

Now that I think about it for a bit, it could be that
the hotter the supply is, the worse the inrush current
will be - the NTC resistor would have a lower resistance
at high temperature. Not that I've ever tested this or
attempted to measure it.

Paul

 

[kony]

On page 3 here, it says:
http://www.antec-inc.com/pdf/manuals/truepower.pdf

INRUSH CURRENT
115V/50A(max.), 230V/80A(max.) at 25°C cold start

That 50A is what makes the lights blink, albeit not
everyone is quick enough to see it. The "25C cold start"
part tells you they are using an NTC resistor to limit
the current, and the colder the supply is when it starts,
the larger the current will be.

I don't recall if they use an NTC or not, but I would not
consider the above as any evidence of it. It is likely not
the case as 50A surge WITH an NTC is rather high.

Plugging in a power supply does not generally make lights
blink. I can't even guess how many times I"ve done it
(plugged in many, many different PC PSU), no blink. Right
now I'll do it... there was no blink.

 

[Zentraleinheit]

Well, I tired what was surgested:

....disconnect AC, clear the motherboard CMOS, then reconnect AC
and retry the front case switch.

And when that did not work:

Connect only the following: CPU, CPU heatsink/fan, 1 memory module, video
Clear CMOS Turn on the motherboard by shorting the two power-on pins
(on the motherboard front panel header, where the switch plug goes) together

Still nothing!

So I started looking for fouled electrical contacts, burst capacitors, fan
failure on CPU heatsink.

I think I found something!

Three of the capacitors next to where the memory slots are have instead of
a flat shinny top have instead a slight bulging top with small brown
discoloration.

(I wish there was a way to post the pictures I made of them so you can see
what I was talking about)

So, if this is the (in this case) the "smoking capacitor" then my
motherboard has met its demise ;-(

Now the only question is do I only get a new motherboard/CPU or a new
case/motherbord/CPU?

And I don't have any one to buy me christmas gifts!

P.S. I should have known something like this was going to happen when my
computer began to go into endless reboots last summer. I thought I fixed it
with a new CPU heat sink and fan, but, as long as I was using applications
that needed high CPU power it still kept happening untill finally the
motherboard just gave up.

 

[Zentraleinheit]

You speculated and then tried to fix it rather than first
discover what was defective

I did not go off and just replace the power supply at the first sign of
trouble.

I posted my problems to my regular group of troubleshooters at AumHa and it
was THEM who came up with the replace the power supply answer!

In the end I was following what was at the time the best advice I had.

And like most advice - it should have been taken with a large pinch of salt

 

[w_tom]

They said to do something, but did not say why. That should
have raised red flags. What you have done (hopefully) is
learn how to see the difference between wild speculation
verses good advise.

Also, heat as described was a diagnostic tool - not a reason
for failure. A computer with a heat problem in a 70 degree
room should then be executed in a hotter room to learn what
happens, make the problem more obvious, and then to eliminate
what may be a hard problem. Again, what the boss (and that is
what you are doing) does when asking for advise from his
subordinates: he never just accepts advise without supporting
facts - the reasons why.

Demonstrated is why fixing things yourself is so useful; so
educational. You are not saving money. Fix things to learn
how the world works - in this case what is required of those
who provide advise.

There is a difference between just replacing a power supply
on recommended whims verses what you have now. Bulging
electrolytic capacitors - especially if leaking material - are
good reasons for failure. And yet that answer still is not
sufficient. You should then ask why. Being as your computer
was becoming less stable (a characteristic of failing
electrolytic), that electrolytics are designed to fail as
described, AND that this is a widespread and recent
electronics problem; three good reasons to suspect the
motherboard.

Now for solutions. One is to replace those electrolytics.
They are cheap and can be purchased even overnight through
electronic supply houses. You end up with what you had
previously which means no other changes or corrections would
be required (ie repairing Windows XP for the new hardware).
That is both good and bad. You now have an opportunity to
upgrade motherboard and processor to a faster system. The
latter solution would be quicker, may cost more money, would
create complications elsewhere, and yet result in a 'better'
system.

BTW, is motherboard still in warranty?

Returning to another statement that others have been
answering. No, kitchen lights should never flicker when power
was applied to computer. To have this happen, one may have
two failures. First the new power supply - if it was built
only on price - would be missing the inrush current limiter.
Still that alone should not cause kitchen light flicker. A
second problem would also exist such as a poorly connected
wire inside one electrical box. For example, if household
wire was back stabbed rather than wrapped around a receptacle
screw, then this poor workmanship in combination with a
missing inrush current limiter could cause a light flicker. Or
maybe a wire to the circuit breaker is loose.

This is certain. Powering on a power supply should never
cause light flickering. That flicker suggests (one
possibility) a minor wiring problem and a power supply missing
an essential function. Does the old power supply also cause
light flicker? Neither should.

Why would I look into that light flicker? In your case,
number one reason why: to learn. Learn about basics of
household wiring AND learn about why many power supplies are
sold without essential internal functions (a lesson about free
market economics).

 

[DaveW]

Apparently your motherboard has failed, since replacing the power supply
unit didn't fix it.

 

[kony]

I think I found something!

Three of the capacitors next to where the memory slots are have instead of
a flat shinny top have instead a slight bulging top with small brown
discoloration.

(I wish there was a way to post the pictures I made of them so you can see
what I was talking about)

That sounds like it has to be failed caps.

So, if this is the (in this case) the "smoking capacitor" then my
motherboard has met its demise ;-(

Now the only question is do I only get a new motherboard/CPU or a new
case/motherbord/CPU?

Most cases of same era as your system, dont' have sufficient
cooling for modern parts. Some do though, you will have to
assess your case. Assuming it is std. ATX, it is compatible
with most modern motherboard beyond the cooling issue.

And I don't have any one to buy me christmas gifts!

P.S. I should have known something like this was going to happen when my
computer began to go into endless reboots last summer. I thought I fixed it
with a new CPU heat sink and fan, but, as long as I was using applications
that needed high CPU power it still kept happening untill finally the
motherboard just gave up.

It was common for the lower-end boards of the P3 era to have
caps that failed. Consider yourself fortunate to have
gotten as much life out of it as you have.

 

[jameshanley39]

Well, I tired what was surgested:

...disconnect AC, clear the motherboard CMOS, then reconnect AC
and retry the front case switch.

And when that did not work:

Connect only the following: CPU, CPU heatsink/fan, 1 memory module, video
Clear CMOS Turn on the motherboard by shorting the two power-on pins
(on the motherboard front panel header, where the switch plug goes) together

Still nothing!

So I started looking for fouled electrical contacts, burst capacitors, fan
failure on CPU heatsink.

I think I found something!

Three of the capacitors next to where the memory slots are have instead of
a flat shinny top have instead a slight bulging top with small brown
discoloration.

(I wish there was a way to post the pictures I made of them so you can see
what I was talking about)

So, if this is the (in this case) the "smoking capacitor" then my
motherboard has met its demise ;-(

George Macdonald once wrote

"Check the electrolytic capacitors on the mbrd for swelling tops or
leakage
near the base. There was a slew of mbrds a couple of years ago which
had
bad caps: http://www.niccomp.com/taiwanlowesr.htm. For examples:
http://www.badcaps.net/ & http://www.motherboardrepair.com/. "

those links have pictures and discussions.

just get a new mbrd, return that one.