newly assembled PC - won't power up

[KJ]

Hi folks,

I just put together my 3rd homebuilt PC. I am having a very irritating
problem. The PC will not power up.

I put all the components together, and turn the power switch of the
PSU on. The led on the front of the case which corresponds to the p-
led cable flashes once for a split-second, then goes off. The
motherboard light is steadily lit the whole time. Pushing the power
button on the case does nothing. Putting a screwdriver between the
power switch pins (momentarily) on the MB does nothing. No sound comes
from the speaker at any time. The cpu and heatsink are seated properly
and thermally greased according to the instructions on the Arctic
Silver site.

I have already replaced the motherboard and power supply, and I am
beginning to think they were not at fault, since the 2nd pair of
components are exhibiting the same symptoms as the first.

Any suggestions as to what may be happening? I am wondering if it's
the case somehow (which I paid a premium for - it's Lian Li brand). Or
some grounding issue. Any and all help is appreciated.

Thank you,
KJ

 

[Frank McCoy]

Hi folks,

I just put together my 3rd homebuilt PC. I am having a very irritating
problem. The PC will not power up.

I put all the components together, and turn the power switch of the
PSU on. The led on the front of the case which corresponds to the p-
led cable flashes once for a split-second, then goes off. The
motherboard light is steadily lit the whole time. Pushing the power
button on the case does nothing. Putting a screwdriver between the
power switch pins (momentarily) on the MB does nothing. No sound comes
from the speaker at any time. The cpu and heatsink are seated properly
and thermally greased according to the instructions on the Arctic
Silver site.

I have already replaced the motherboard and power supply, and I am
beginning to think they were not at fault, since the 2nd pair of
components are exhibiting the same symptoms as the first.

Any suggestions as to what may be happening? I am wondering if it's
the case somehow (which I paid a premium for - it's Lian Li brand). Or
some grounding issue. Any and all help is appreciated.

Try running the motherboard *out of the case*.
Just hook up PSU, speaker, and video-board.
If that doesn't work, try without the video.
If it does, add keyboard, etc..

Watch out for standoff problems and case contacting the bottom of the
mobo. Look around the standoffs and make sure there's *ground* all
around where the standoff connects. Sometimes you have to use plastic
standoffs in some places on badly designed boards.

 

[KJ]

Hello Frank,

I just tried the following:

Took MB out of case, connected only CPU + heatsink, PSU, video card, 1
stick RAM. I also tried with no ram, ram and video card, no video
card, etc. The same thing is happening - no POST. I also noticed that
when I shut off the power, the p-led on the case flashes for a split-
second, just as it does I turn the power on. Something is preventing
the power supply from kicking in, and I have no idea what it is.

 

[KJ]

I should also mention, there are no beeps whatsoever, and, I also
noticed that there seems to be a barely noticeable, momentary flicker
in the light bulb of this room when I push the power button or short
the pwsw pins. The wierd thing is that the lights are on a different
circuit than the outlets. The outlet is a 20A circuit, and its only
other load to speak of is a Dell PC and monitor.

 

[Frank McCoy]

Hello Frank,

I just tried the following:

Took MB out of case, connected only CPU + heatsink, PSU, video card, 1
stick RAM. I also tried with no ram, ram and video card, no video
card, etc. The same thing is happening - no POST. I also noticed that
when I shut off the power, the p-led on the case flashes for a split-
second, just as it does I turn the power on. Something is preventing
the power supply from kicking in, and I have no idea what it is.

I'd look for a match between CPU, motherboard, and jumper settings then.
I'd also make *damned* sure the CPU was properly heat-sunk.
But the jumpers first.

Are you *sure* the CPU you have is one specified for the mobo?
Also, did you clear CMOS memory?
Even on newer boards that only support a few types of CPU, often they
have at least on jumper (or switch) setting to select which of two,
three, or four types. On older boards, the combinations could get
rather complicated; and you usually have to match the settings
*exactly*.

Another possibility is the same sort of thing with the memory chips.
If the clock-rate is set too fast for either one, for example ....

Yet another reason to be sure you cleared the CMOS ... The default
setting or one left from testing might be wrong for your memory or CPU.

Finally, look *close* at the jumper pins for the reset, speaker, and all
the other connectors. Best bet is to run the mobo on a bare wood floor
with *nothing* connected to any of them, and nothing in the mobo except
the CPU. No memory, no speaker, no video, no LEDs, nothing except PSU,
CPU, and motherboard.

See if the CPU fan starts running when you short the power-switch
connector ... and make *triply* sure that's the two pins you are
shorting, not the reset pins or speaker pins or something else.
Generally they're labled; both in the manual AND on the board itself.

If the CPU fan *does* start, then add, checking each one, in this order:
power-switch
speaker
RAM
video-board
power-LED
HD-LED
Reset-switch

Any of those, wrongly connected, shorted, or otherwise screwing things
up can stop he PSU from running.

If the CPU fan doesn't run with just CPU, PSU, and motherboard, then
you'll have to independently test each one of them in a "known good"
configuration. A power-supply-tester (about $20 from most computer
stores or BEST BUY) is a handy gadget; though not definitive. It won't
tell you for sure that a PSU is good ... but if it doesn't light up then
you're pretty damned sure the PSU is BAD. (Not always; but usually.)

 

[Frank McCoy]

I'd look for a match between CPU, motherboard, and jumper settings then.
I'd also make *damned* sure the CPU was properly heat-sunk.
But the jumpers first.

Are you *sure* the CPU you have is one specified for the mobo?
Also, did you clear CMOS memory?

Just thought: Make *sure* the CMOS jumper isn't permanently in the
"CLEAR" position, or removed from the board so it can't be set in the
"normal" position.

 

[Frank McCoy]

I should also mention, there are no beeps whatsoever, and, I also
noticed that there seems to be a barely noticeable, momentary flicker
in the light bulb of this room when I push the power button or short
the pwsw pins. The wierd thing is that the lights are on a different
circuit than the outlets. The outlet is a 20A circuit, and its only
other load to speak of is a Dell PC and monitor.

See my other reply.
The flicker (seems to me) might be normal as the PSU draws momentary
power to supply the board before the board tells it to shut down;
whether from short, overload, or just wrong configuration. It takes a
bit of a surge to charge up those big capacitors in the PSU as the power
is first applied.

 

[KJ]

Hi Frank,

I just swapped out the CPU with a known working CPU, same model -
result is the same. I cleared CMOS before doing this. The mother board
in question has a clear CMOS button, rather than a jumper. The MB is
an MSI P6N SLI Platinum, CPU is an Intel Core 2 duo. RAM is Wintec.

I tried just only MB, PSU, CPU. Still, nothing. At least the light
bulb above flickers for an instant.

Thanks for keeping with me on this one. I feel I may be entering a
hell-realm at this point.

 

[w_tom]

I just swapped out the CPU with a known working CPU, same model -
result is the same. I cleared CMOS before doing this. The mother board
in question has a clear CMOS button, rather than a jumper. The MB is
an MSI P6N SLI Platinum, CPU is an Intel Core 2 duo. RAM is Wintec.

I tried just only MB, PSU, CPU. Still, nothing. At least the light
bulb above flickers for an instant.

Now is time to first see a problem; to stop shotgunning. Currently
no useful information makes any useful decisions. Get a 3.5 digit
multimeter. Two minutes with this procedure will either report what
is wrong OR provide numbers so that others can provided something
useful.

Why does the light flash? We have no idea - just numerous
speculation. For example, power supply powers up. A function detects
a problem. Then power supply is shut down. Why or what is causing
that? No one knows without minimal data such as those numbers. And
still, even that above suggestion is nothing more than one 'wild
speculation' guess. Guess is all anyone can do until you first have
numbers AND stop disconnecting anything.

There was zero reasons to suspect CPU, memory, heatsink, CMOS, etc.
In fact, doing those things could only exponentially complicate the
problem; it's called shotgunning. What does the light flash report?
That power cord is connected to a life AC wall receptacle; nothing
more.

Get the meter. Take two minutes to get those numbers. Only then
will replies be useful.

 

[KJ]

I do have a multimeter (ohmmeter) which has dial-in settings for
various voltages. Would you mind giving me some instructions on what
to do? I am not experienced in this area.

p.s. I will not be able to reply to this thread for the next 12 hours.

Thanks,
KJ

 

[w_tom]

I do have a multimeter (ohmmeter) which has dial-in settings for
various voltages. Would you mind giving me some instructions on what
to do? I am not experienced in this area.

Procedure to collect numbers for analysis and for posting were in
"When your computer dies without warning....." starting 6 Feb 2007 in
the newsgroup alt.windows-xp at:
http://tinyurl.com/yvf9vh

First measurement is on purple wire - when computer is not powered
on but is connected to AC wall receptacle. Put the meter in 20
Volts DC range. Black probe (from meter) can connect to any black
wire on that 'power supply to motherboard' connector, OR can be
clipped to metal chassis.

Red probe is pushed inside white (nylon) connector to touch purple
wire contact. Now read voltage on meter. Record that number. Then
go to next ste in procedure. Green wire(?). Touch red probe to green
wire to record that voltage before power switch is pressed. Then
monitor on meter what happens when power switch is pressed. Voltage
might drop down and then remain near zero. Those numbers - useful
facts - inform all what your power supply 'system' is doing. Just
follow that procedure and report back.

 

[Frank McCoy]

Hi Frank,

I just swapped out the CPU with a known working CPU, same model -
result is the same. I cleared CMOS before doing this. The mother board
in question has a clear CMOS button, rather than a jumper. The MB is
an MSI P6N SLI Platinum, CPU is an Intel Core 2 duo. RAM is Wintec.

I tried just only MB, PSU, CPU. Still, nothing. At least the light
bulb above flickers for an instant.

Then likely, as I said (and you don't have a known-good power-supply to
try ... No, not another one of the same type; but one that's been
*working* in another computer, which is what I'd try *first*) then I'd
go on to configuration issues.

So ... I went out and looked at the manual for the board ...
First: some stupid things:

1. You *did* remove the plastic cover and leave it off when installing
the heatsink and fan?
2. You *did* connect the fan supply to CPUFAN1?
3. You *did* use heatsink compound between CPU and heatsink?
4. You *did* turn the motherboard over, after installing heatsink and
fan, and made *sure* that all four clip-ends or hold-downs for the
heatsink made it through the motherboard and out the other side?
5. YOU *did* press the clear-CMOS switch while the power-supply was not
only off, but either disconnected or the plug pulled?

Next:
Does the orange LED next to the clear-CMOS button light up when the
motherboard is connected to the PSU?

Try connecting the diagnostic LED panel (part of the second USB header)
to JBD1, and see if any of the LEDs light up. If so, read the manual to
see what the board is complaining about. (That's page 2-21 in the
manual.)

Try shorting the power-switch terminals (pressing the power-switch) four
times; holding it closed for about a second each time; but then waiting
about fifteen seconds between shorts or presses to go to default and
fail-safe clock speeds on boot. This to be done with the PSU connected.

 

[Paul]

I should also mention, there are no beeps whatsoever, and, I also
noticed that there seems to be a barely noticeable, momentary flicker
in the light bulb of this room when I push the power button or short
the pwsw pins. The wierd thing is that the lights are on a different
circuit than the outlets. The outlet is a 20A circuit, and its only
other load to speak of is a Dell PC and monitor.

Does the power supply have a 115V/230V switch on it ?
If you are using a 115V wall outlet and the switch is in the
230V position, that might cause "weak" behavior. If on the
other hand, you were in a country with 230V outlets, and
the switch was set an 115V, the caps inside the supply should
blow up.

Is the supply an Antec NEO-HE ? There are some issues with those.

Paul

 

[KJ]

Thank you so much for looking up the manual. I have tried 1-5 below,
which are basic checkpoints, but worthy of double checking anyway. The
orange MB LED is steadily lit while the PSU is switched on.

Unfortunately, the JBD1 connector does not light at all, and I think
this is because the MB is sending a kill signal to the PSU immediately
after shorting the pwsw pins or (or switching it on).

I will try the on-four-times trick in about 12 hours, when I return
home.

Thanks again.

 

[KJ]

The PSU is a 500 watt Antec "Earthwatts" model. I will check the
230/115 setting tonight when I return home. Thanks for responding.

 

[Mike T.]

The PSU is a 500 watt Antec "Earthwatts" model. I will check the
230/115 setting tonight when I return home. Thanks for responding.

OK, that's a start. Now what is the exact make, model AND REVISION of
motherboard? From reading this whole thread, I suspect your power supply is
not compatible with your mainboard, even if all components are working as
designed. -Dave

 

[Frank McCoy]

First: some stupid things:

1. You *did* remove the plastic cover and leave it off when installing
the heatsink and fan?
2. You *did* connect the fan supply to CPUFAN1?

Just thought of something else:
The CPUFAN connecter is a 4-pin header that will accept 3-pin fans.
If you have a 3-pin fan, did you make sure it's connected on the correct
end of the fan-header? Like the other things, seems stupid; and it's
keyed, but ....

Also, are you *sure* you're shorting the power-switch pins, not the
reset? (It doesn't hurt to try shorting both, separately, for a second.)

And last stupid thing:
Have you connected the 12-volt supply from the PSU for the CPU?
(Your power-supply *does* have that extra connector, not just the old
standard ATX pinout, doesn't it?)
Checking ... Yeah, it does have an 8-pin 12-volt cable.

Overall, nice supply and nice motherboard.
Since the orange LED lights on the motherboard, there's probably
something silly and still stupid that I'm not thinking about to try.

I'm running out of guesses.
Having a different working supply to try would be nice ...

 

[Paul]

The PSU is a 500 watt Antec "Earthwatts" model. I will check the
230/115 setting tonight when I return home. Thanks for responding.

OK, checking the description on Newegg, it is an autoranging power
supply, with active power factor correction, and does not have a
switch to set the voltage.

Looking at a picture of the label on the EA-500, it has a minimum
current consumption. The 5V loading will be met, by having at
least one hard drive connected. The 12V minimum might be met by
the processor, but it is hard to say. I don't know if underloading
that supply, could be causing it to behave that way or not. It
almost sounds like shades of NeoHE all over again.

So, here is what I want you to try next. Find two old hard drives.
These will function as dummy loads. You don't have to connect the
ribbon cable interface, or the SATA interface if they have one.
Just connect whatever is used to provide power to the drives.
The two disk drives will help draw a little bit more power from
the supply, and may allow it to stay on. If the supply starts, you
should hear both drives spinning.

The two reviews for the supply, on Newegg, give it a passing grade.
There aren't enough reviews to see a trend yet.

It is reviewed here, and is apparently made under contract by
Seasonic. The bumps in the waveforms JonnyGuru recorded, might
be line noise leaking through (I calculate 56Hz, but should the
line frequency be that far off ?).

http://www.jonnyguru.com/review_details.php?id=51&page_num=3

Paul

 

[KJ]

Friends,

I am so happy. I just swapped out the PSU with another (older) Antec
Smart Power 500W PSU, and, voila, powers up instantly!

Now, may I offer a hearty thanks to Paul, Mike, Frank, and Tom for
their troubleshooting tips. I learned a heck of a lot from you guys on
this thread.

The only thing I don't understand is *why* the new PSU fails to work.
Maybe it's just a "copmpatibilty" issue. Perhaps it will work with my
older (last year's) computer, or maybe the good people at CompUSA will
take it back even though it's pretty scratched up on the side from
being taken in and out of the case 3 or 4 times. If I only had the
energy now to put the whole thing back together and install the OS...

Be well.

-KJ

 

[Paul]

Friends,

I am so happy. I just swapped out the PSU with another (older) Antec
Smart Power 500W PSU, and, voila, powers up instantly!

Now, may I offer a hearty thanks to Paul, Mike, Frank, and Tom for
their troubleshooting tips. I learned a heck of a lot from you guys on
this thread.

The only thing I don't understand is *why* the new PSU fails to work.
Maybe it's just a "copmpatibilty" issue. Perhaps it will work with my
older (last year's) computer, or maybe the good people at CompUSA will
take it back even though it's pretty scratched up on the side from
being taken in and out of the case 3 or 4 times. If I only had the
energy now to put the whole thing back together and install the OS...

Be well.

-KJ

I recommend researching NeoHE issues, and see if the symptoms are the
same. Somehow, this all sounds too familiar.

Paul