How to troubleshoot?


D

DK

I would appreciate any suggestions:

Starting last night, the desktop spontaneously shuts down.
All signs of power failure - just sudden turn off of everything.
Happened overnight and three times today.

Opened up, cleaned everything, reset memory sticks, quickly
tested memory with MemTest86, all seems fine. No bulging
or leaky capacitors anywhere I can see.

No overheating - CPU 38C, hard drives 40 and 44C.

Restored an image from two weeks ago - same thing, so obviously
hardware issue. The machine has seen no hardware changes for
the past 2 years and was alway rock stable.

The only thing I can think about is power supply but I am not sure
how to troubleshoot it. (I have Nexus Value 430 PSU that is 5 years
old). Another guess is a motherboard but, once again, not really
sure what I can do to pinpoint the problem (Asus M4N68T-M-V2
that is 2.5 years old).

DK
 
Ad

Advertisements

P

Paul

DK said:
I would appreciate any suggestions:

Starting last night, the desktop spontaneously shuts down.
All signs of power failure - just sudden turn off of everything.
Happened overnight and three times today.

Opened up, cleaned everything, reset memory sticks, quickly
tested memory with MemTest86, all seems fine. No bulging
or leaky capacitors anywhere I can see.

No overheating - CPU 38C, hard drives 40 and 44C.

Restored an image from two weeks ago - same thing, so obviously
hardware issue. The machine has seen no hardware changes for
the past 2 years and was alway rock stable.

The only thing I can think about is power supply but I am not sure
how to troubleshoot it. (I have Nexus Value 430 PSU that is 5 years
old). Another guess is a motherboard but, once again, not really
sure what I can do to pinpoint the problem (Asus M4N68T-M-V2
that is 2.5 years old).

DK

1) Remove power supply. Undo the four screws on top of the supply.
(This will involve removing the "warranty void" stickers
on top of the screws.)

Look for leaking or bulging capacitors. *Do not* touch the
components in there. One or two capacitors in there, could have
dangerous voltages present, if the bleeder resistors are not
functioning.

In this picture, C5 and C6 are dangerous, and R2 and R3 are
safety bleeders. You work out the product of those two component
values, to determine the discharge time constant.
0.000470 * 220000 = 103.4 seconds. That's the tau for the
circuit, and assumes R2 or R3 aren't blown open. A careful
technician assumes R2 and R3 are open circuit, and a dangerous
charge remains on C5 and C6. C5 and C6 won't drain, unless R2
and R3 are working properly. *Do not* stick a screwdriver across
the terminals of C5 or C6 - the noise will deafen you, and you'll
need to wash your underwear :)

http://www.pavouk.org/hw/en_atxps.html

This is why I shorten the information to "just don't touch
anything in there". To keep you really really safe.

You can see nine capacitors in this power supply are leaking,
and considering the price of power supplies, and the difficulty
of retrofitting caps, it's time for the garbage pail. Typically
the capacitors have smaller than normal dimensions, and when
you shop for a replacement, it always ends up one size (volumetrically)
larger than the previous one. So even if you shop at the
Radio Shack for caps, then come home, your new caps will have
too large an outside diameter. Just to illustrate one of the problems
you can run into.

http://upload.wikimedia.org/wikipedia/commons/f/f5/Al-Elko-bad-caps-Wiki-07-02-17.jpg

I had an Antec supply that failed like that - leaking capacitors.
The supply would make a "sizzling" sound, as current arced inside
the damaged capacitors - the sizzling (and occasional puff of smoke),
would happen for the first 30 seconds after power-up.

2) The CPU has THERMTRIP. If there was any defect in that
circuit, a failed component close to that signal, that might
be sufficient to turn off the system. You've already checked the CPU
temperature, and so I'm going to assume it's not a motherboard
or CPU issue. While the computer is still running, I would
take the side off the computer and visually inspect the CPU
cooler for proper clamping, check that the fan is spinning, in
case the CPU temperature reading was always fixed at 38C and
never ever changes.

A CPU temperature readout is only trustworthy, if you've noted it
previously had the correct dynamics (rate of change relatively
quick, if the readout is a silicon die temperature, relatively
slow rate of change if the sensor is a socket sensor with
flat Kapton film type sensor stuffed under the socket). Note
that tools like Speedfan, will occasionally display the wrong
sensor as being the one for the CPU. And that's what I'm trying
to warn you about. If your CPU is modern enough, you could double
check with CoreTemp or some other utility, just in case.

HTH,
Paul
 
J

John Doe

dk said:
I would appreciate any suggestions:

Starting last night, the desktop spontaneously shuts down. All
signs of power failure - just sudden turn off of everything.
Happened overnight and three times today.

Opened up, cleaned everything, reset memory sticks, quickly
tested memory with MemTest86, all seems fine. No bulging or
leaky capacitors anywhere I can see.

No overheating - CPU 38C, hard drives 40 and 44C.

Restored an image from two weeks ago - same thing, so obviously
hardware issue. The machine has seen no hardware changes for the
past 2 years and was alway rock stable.

The only thing I can think about is power supply but I am not
sure how to troubleshoot it. (I have Nexus Value 430 PSU that is
5 years old). Another guess is a motherboard but, once again,
not really sure what I can do to pinpoint the problem (Asus
M4N68T-M-V2 that is 2.5 years old).

Obviously the very first thing you do is check power supply
voltages. Obviously by looking in the BIOS first. Then, the next
obvious thing you would do is check from within Windows XP. You
can find a utility that shows power supply voltages from within
Windows.

If everything looks good, you might suspect your household power,
especially at that location/outlet.

Instead of pretending to be an electronics technician, for
troubleshooting the power supply you can always try a different
power supply. That's the obvious and easy way to troubleshoot the
power supply. If it proves faulty, then you can decide whether or
not you want to become an electronics technician.
 
L

Loren Pechtel

I would appreciate any suggestions:

Starting last night, the desktop spontaneously shuts down.
All signs of power failure - just sudden turn off of everything.
Happened overnight and three times today.

Opened up, cleaned everything, reset memory sticks, quickly
tested memory with MemTest86, all seems fine. No bulging
or leaky capacitors anywhere I can see.

No overheating - CPU 38C, hard drives 40 and 44C.

Restored an image from two weeks ago - same thing, so obviously
hardware issue. The machine has seen no hardware changes for
the past 2 years and was alway rock stable.

The only thing I can think about is power supply but I am not sure
how to troubleshoot it. (I have Nexus Value 430 PSU that is 5 years
old). Another guess is a motherboard but, once again, not really
sure what I can do to pinpoint the problem (Asus M4N68T-M-V2
that is 2.5 years old).

I've seen such things, it's always been a failing power supply.

To test shed whatever load you can and see if it remains up. If
that's not feasible try another power supply and see if it works.
 
L

Loren Pechtel

In this picture, C5 and C6 are dangerous, and R2 and R3 are
safety bleeders. You work out the product of those two component
values, to determine the discharge time constant.
0.000470 * 220000 = 103.4 seconds. That's the tau for the
circuit, and assumes R2 or R3 aren't blown open. A careful
technician assumes R2 and R3 are open circuit, and a dangerous
charge remains on C5 and C6. C5 and C6 won't drain, unless R2
and R3 are working properly. *Do not* stick a screwdriver across
the terminals of C5 or C6 - the noise will deafen you, and you'll
need to wash your underwear :)

While I certainly agree you don't want to touch them is it really that
big a bang? 2/3 of a joule doesn't strike me as that much energy. Or
am I doing something wrong in figuring it out?
 
P

Paul

Loren said:
While I certainly agree you don't want to touch them is it really that
big a bang? 2/3 of a joule doesn't strike me as that much energy. Or
am I doing something wrong in figuring it out?

I originally started out as a "screwdriver jockey", working
on tube radios (with radio unplugged), and making circuits
safe by sticking the screwdriver blade across them. The
first time I stuck the blade across something substantial,
is when I learned my lesson. I fashion bleeders now, for
anything like that. Some technicians, such as Mr. Kenney
who repaired vacuum tube equipment for the university, he
still used the screwdriver technique, for "reforming"
capacitors in old equipment. So there are times when a
screwdriver is useful, but do that with your hearing protection
on.

The time the capacitor in the microwave oven at work arced
over inside, the noise was so loud (with me standing next to
it, heating up supper), I was deaf in one ear for ten minutes.
But we'll just do the ATX power supply calc for now, rather
than the nasty microwave oven calc.

*******

230VAC is approximately 226V at the mast
226 * 1.414 = Vp = 319.6V or 320V

AC ------ bridge ------ 320VDC -----+-------------+
226VAC rectifier | |
C5 470uF R2
160VDC across bleeder resistor
| |
+-------------+
| |
C6 470uF R3
160VDC across bleeder resistor
| |
+-------------+

Energy (joules) = 1/2 * C * V^2

= 0.5 * 0.000470 * 160 * 160 = 6 Joules

I must now launder my shorts. I value my hearing.
I know someone who is deaf in one ear, from doing
too many HV discharges without hearing protection.
Your hearing doesn't go in one shot, but over the
years, you can lose it. And being an idiot of a PHD,
he ran experiments without hearing protection. He
makes a nice health and safety promotional in person,
when you have to "talk into his good ear".

On modern Active PFC circuits, I've read somewhere that
the boosting action of the PFC circuit, raises the voltage
on the caps a bit. The exact energy value in joules,
is going to depend on how big they made those capacitors.
They may actually be using different sizes than 470uF,
depending on the capacity of the design, because part
of the purpose of the caps is "holdup time" at full load
and a loss of input. By having holdup time, your crappy
"home UPS" with the glitch in the output during switchover,
you never see a thing on the computer itself. The
C5 and C6 hold enough juice to tide you over, until
the UPS inverter starts up.

Paul
 
Ad

Advertisements

P

Paul in Houston TX

Loren said:
While I certainly agree you don't want to touch them is it really that
big a bang? 2/3 of a joule doesn't strike me as that much energy. Or
am I doing something wrong in figuring it out?

Well... It's all relative. I've been into electronics for a long time.
A 470mfd charged to 200v makes a decent bang when shorted.
It's about the same as a small 3/4" firecracker going off about 18"
in front of your face. A 1/8" spherical bit of metal vapor expanding
at the speed of sound is fairly loud. If I can find my
old screwdriver I will take a pic of it.
I can also testify that discharging a similar capacitor through
your fingers will paralyze them for about 24 hours.
And it hurts quite badly.
 
B

Buffalo

"DK" wrote in message news:[email protected]
I would appreciate any suggestions:

Starting last night, the desktop spontaneously shuts down.
All signs of power failure - just sudden turn off of everything.
Happened overnight and three times today.

Opened up, cleaned everything, reset memory sticks, quickly
tested memory with MemTest86, all seems fine. No bulging
or leaky capacitors anywhere I can see.

No overheating - CPU 38C, hard drives 40 and 44C.

Restored an image from two weeks ago - same thing, so obviously
hardware issue. The machine has seen no hardware changes for
the past 2 years and was alway rock stable.

The only thing I can think about is power supply but I am not sure
how to troubleshoot it. (I have Nexus Value 430 PSU that is 5 years
old). Another guess is a motherboard but, once again, not really
sure what I can do to pinpoint the problem (Asus M4N68T-M-V2
that is 2.5 years old).

DK
Check the simple things that power your pc first.
Make sure the wall outlet is good.
If there is a voltage selector switch on the back of the PSU, switch it back
and forth a few times to make sure it is making good contact (with the
power disconnected, of course)
Try a different power cord to your PC.
Inside the PC, remove the big connector and put it back on several times to
possibly make better contact.
Run a program that will monitor your voltages and temps and record them.
Make sure your cpu cooler fan is running and that the cpu is solidly
connected to the heatsink.
There are sites online that describe how to check a PSU.
Best of luck and when you get it taken care of, please post back.
 
L

larrymoencurly

Starting last night, the desktop spontaneously shuts down.
All signs of power failure - just sudden turn off of everything.
Happened overnight and three times today.

Opened up, cleaned everything, reset memory sticks, quickly
tested memory with MemTest86, all seems fine. No bulging
or leaky capacitors anywhere I can see.

No overheating - CPU 38C, hard drives 40 and 44C.

Restored an image from two weeks ago - same thing, so obviously
hardware issue. The machine has seen no hardware changes for
the past 2 years and was always rock stable.

The only thing I can think about is power supply but I am not sure
how to troubleshoot it. (I have Nexus Value 430 PSU that is 5 years
old). Another guess is a motherboard but, once again, not really
sure what I can do to pinpoint the problem (Asus M4N68T-M-V2
that is 2.5 years old).

I suggest you unplug the power connections to the hard disks
while testing because a bad computer can overwrite the disks
with junk.

There may be a bad power connection, including cracked solder
around one of the socket pins on the motherboard, a marginal
short between the motherboard and case (especially around the
mounting holes -- check both top and bottom of each) and at
the corners.

I'd try another PSU or at least measure the voltages of the
current one with a digital multimeter, while the computer is
running because you need to test with a realistic load.

Don't open up a PSU except when its AC power cord is unplugged
because a lot of areas inside have high voltage on them, often
including one of the big heatsinks. Fortunately the components
can be tested without the PSU running, by measuring them with
a multimeter's diode check or ohms function. However electrolytic
capacitors (the cylindrical ones) can't be thoroughly tested
except by removing them and using an ESR meter. Reinstall the
PSU cover completely (tighten all screws) before reattaching
the AC power cord.

It seems the Asus M4N68T-M-V2 motherboard is made with all
solid polymer capacitors (those silver cylinders), rather
than regular wet electrolytic caps, meaning the caps should
last practically forever.

Despite the seemingly fake UL approval ("RU") for the
Nexus PSU, it actually looks well built and got a glowing
review from HardwareSecrets, which also dissected it:

www.hardwaresecrets.com/printpage/Nexus-Value-430-Power-Supply-Review/1544

However it was made with Teapo brand capacitors (notice the 3-slice
pie pattern stamped into each top, with a dot in each slice). Low
voltage Teapos made for high frequency operation are not known for
lasting that long, and after 5 years there's a good chance some
have failed, but caps don't always bulge or leak when they do.
BadCaps.net and its forums have much more information.
 
B

BillW50

In Bill in Co typed:
They don't use any isolation transformers in these power supplies? I
wonder how they avoid some potential grounding issues? (I can still
remember working on those old All American Five vacuum tube radios
which also ran directly off of the mains, with NO mains transformer,
and were a bit of a pain to work with (in reference to potential
grounding and/or shock issues).

I always use an isolation transformer to solve those problems when
working on electronics on the bench. I haven't bought one in years, so I
don't know what they would cost nowadays.
 
Ad

Advertisements

B

BillW50

In DK typed:
No overheating - CPU 38C, hard drives 40 and 44C.

Here is one interesting thing I found on one Gateway MX6124. I have
three of them and one of them reads about 10F lower at idle and 40F
lower than normal with the CPU maxed out. That machine is super clean
inside and I wasn't sure what to make of it? I did push the CPU hard for
about 6 weeks straight and then the CPU cooked itself to death.

Dropped in a new CPU and that one also has the same lower readings. So
at least it wasn't the CPUs fault anyway. The machine never runs the fan
on high since it thinks the CPU is running so cool anyway. It still
works today (many years later) and I just don't use it for heavy CPU
tasks and it works fine that way.

Now yours running at 38C (100F) is pretty low for a CPU. What kind of
CPU is this? And when you run a task that max out the CPU, how high does
it go? In my experience, CPU temp reading incorrectly is very rare. But
so is a CPU running at 38C is also pretty rare. I don't know about for
desktops (I quit using desktops 10 years ago), but for laptops and
tablets 38C isn't normal.
 
P

Paul

BillW50 said:
In Bill in Co typed:

I always use an isolation transformer to solve those problems when
working on electronics on the bench. I haven't bought one in years, so I
don't know what they would cost nowadays.


The ATX supply has input to output isolation. All the
paths have a transformer in the way, as you can see in
my favorite tutorial page. The transformers form a dividing
line, between input side and output side. One transformer
is for "base drive". The other transformer is for combined
outputs. A few obscure ATX power supplies, used more
transformers on output. This schematic basically
shows how the simplest, cheapest ones are built.

http://www.pavouk.org/hw/en_atxps.html

And on the exterior of the ATX supply, you can see a
"Hipot tested" sticker, which means a machine tested
for a minimum level of isolation between input and
output pins. To prove it really is isolated, for each
unit tested. The Hipot test voltage is higher than
what the line can provide.

In North America, the input safety ground is connected to
the output side ground. And this can be an issue with
getting "hum" when connecting other line-powered audio
devices. Removing the safety ground connection, provides
no place for front end filter leakage current to go,
and will cause the chassis to be electrified. So if you
use a two wire cord with a modern ATX supply, leaving the
chassis metal "floating", the front end filter (common
node between C2 and C3) will see to it that AC leakage will
appear on the chassis. As a wakeup call... :)
A modern computer is only completely trouble
free, when plugged into a proper three pronged outlet.
Otherwise, you may "jump" to the wrong conclusion :)

Paul
 
P

Paul

BillW50 said:
In DK typed:

Here is one interesting thing I found on one Gateway MX6124. I have
three of them and one of them reads about 10F lower at idle and 40F
lower than normal with the CPU maxed out. That machine is super clean
inside and I wasn't sure what to make of it? I did push the CPU hard for
about 6 weeks straight and then the CPU cooked itself to death.

Dropped in a new CPU and that one also has the same lower readings. So
at least it wasn't the CPUs fault anyway. The machine never runs the fan
on high since it thinks the CPU is running so cool anyway. It still
works today (many years later) and I just don't use it for heavy CPU
tasks and it works fine that way.

Now yours running at 38C (100F) is pretty low for a CPU. What kind of
CPU is this? And when you run a task that max out the CPU, how high does
it go? In my experience, CPU temp reading incorrectly is very rare. But
so is a CPU running at 38C is also pretty rare. I don't know about for
desktops (I quit using desktops 10 years ago), but for laptops and
tablets 38C isn't normal.

We were warned in a web article, to verify the CPU sensor
really was the CPU sensor.

If you go into the BIOS, and use the hardware monitor page,
there's a good chance they'll get it right in there.

Once booted into the OS, programs like MBM5 or Speedfan, could
be referencing the wrong channel as the CPU value.

And the recommended test procedure, was to use load
test software. Then examine the dynamics of the
temperature change.

The CPU should heat up fast. The sensor value should
increase rapidly as well. If the sensor is an on-die sensor,
it should track the rapidly rising temperature very well.
If the sensor is underneath the CPU socket, it'll take a
lot longer to warm up. On modern CPUs, there's a good chance
you can use CoreTemp and get the silicon die temperature
that way.

For sensors which rise slowly, those could be monitoring
computer case air temperature. For example, if I had a sensor
located near the Southbridge chip on the motherboard, the intention
may be case air temperature monitoring. It might take
ten minutes for the case temperature to be roughly
stable at the new operating point. And the slow rise, to
a value like 38C, then tells you it's actually a case
sensor.

As for fan speed control, some Intel fans had thermal sensing
in the hub. And the Intel branded fan could speed up as
the case air temperature rose. This was a form of
compensation for the heatsinks being less effective,
if the case air is getting hot.

If the cooling system varies the fan speed, based
on measured temperature (i.e. coordinated with
the hardware monitor on the motherboard), then the
cooling is adjusted according to a nominated
sensor. If you set things up to respond to the
wrong sensor, they yes, silly things will happen.
I think Speedfan has the ability to bind an
arbitrary measurement channel, to the fan speed.

So there is some room for experimentation
and verification that all is well in
hardware monitor land. And you haven't
done something wrong by accident when setting
things up.

An even quicker check, is to open the case
(take the side panel off), and just verify the
fan is still spinning on the CPU, and the heatsink
clamps are still secure. That's a lot faster
to do, than any amount of Hardware Monitor work.
You might easily catch a perilous condition,
with a visual check. The plastic tabs that hold
some CPU coolers, have been known to snap off.

Paul
 
J

J. P. Gilliver (John)

Bill in Co said:
They don't use any isolation transformers in these power supplies? I wonder
how they avoid some potential grounding issues? (I can still remember
working on those old All American Five vacuum tube radios which also ran
directly off of the mains, with NO mains transformer, and were a bit of a
pain to work with (in reference to potential grounding and/or shock issues).
[]
There is an isolation transformer, but it runs at the switching rate of
the power supply, not 50 (or 60 in USA) hertz, so is ferrite rather than
laminated iron core. The mains side of the switcher does indeed run
non-isolated like those old half-live wireless and TV chassis; that's
why the PSU block is somewhat better sealed than the rest of the PC. The
5, 12, etc. volt circuitry _is_ isolated from the mains (US: line).

IME, the inside edges of the PSU block are also more likely to have
sharp edges, as they're not supposed to be user-serviced, but things
might have improved lately - I haven't been into one for a while.

Some _very_ early ones _did_ have a conventional mains
transformer/rectifier/regulator for the _standby_ supply, though I think
that was probably back in the says of AT rather than ATX supplies, and
only early ones at that. (In fact conventional mains-frequency
transformers are becoming a rarity in anything - I can't remember the
last time I saw _any_ new gear with one in. Possibly in hi-fi purist
kit, or some radio kit where switchers would interfere.)
 
M

Michael Black

They don't use any isolation transformers in these power supplies? I wonder
how they avoid some potential grounding issues? (I can still remember
working on those old All American Five vacuum tube radios which also ran
directly off of the mains, with NO mains transformer, and were a bit of a
pain to work with (in reference to potential grounding and/or shock issues).
The worry being discussed is that there is high voltage, an isolation
transformer won't help one bit.

Switching supplies of this type rectify the incoming AC line voltage,
filter it with some capacitors, and that is used to feed an oscillator
that feeds a toroid transformer. Since the oscillator runs at a much
higher frequency than 60Hz out of the line, the transformer can be much
smaller for the same amount of current. Once you are on the secondary of
the switching supply, you are isolated from the line. They generally use
optoisolators to send feedback to the primary.

Now, if one is actually going to power up the switching supply and poke
around on the primary, not only is there the issue of high voltage, but
lack of isolation from the AC line. But if you aren't careful, you'll get
shocked from the high voltage before you'll notice the lack of isolation.

Where the isolation transformer might help here is if you need to attach
an oscilliscope or VTVM to the primary, since otherwise connecting the
ground to the power supply primary will cause a short to ground.

But for most people, they aren't tracing circuitry or doing testing under
power, they will replace some components and hope that fixes it, or give
up and buy a new power supply.

Michael
 
Ad

Advertisements

L

Loren Pechtel

I originally started out as a "screwdriver jockey", working
on tube radios (with radio unplugged), and making circuits
safe by sticking the screwdriver blade across them. The
first time I stuck the blade across something substantial,
is when I learned my lesson. I fashion bleeders now, for
anything like that. Some technicians, such as Mr. Kenney
who repaired vacuum tube equipment for the university, he
still used the screwdriver technique, for "reforming"
capacitors in old equipment. So there are times when a
screwdriver is useful, but do that with your hearing protection
on.

I understand what's going on, I was just being surprised at the
volume.
The time the capacitor in the microwave oven at work arced
over inside, the noise was so loud (with me standing next to
it, heating up supper), I was deaf in one ear for ten minutes.
But we'll just do the ATX power supply calc for now, rather
than the nasty microwave oven calc.

*******

230VAC is approximately 226V at the mast
226 * 1.414 = Vp = 319.6V or 320V

Oh, I see--I live in 110V territory. That's a lot more energy than I
was thinking.
Energy (joules) = 1/2 * C * V^2

= 0.5 * 0.000470 * 160 * 160 = 6 Joules

And I must have done something more wrong. You're getting 10x the
bang I was thinking.
 
P

Paul

Loren said:
Oh, I see--I live in 110V territory. That's a lot more energy than I
was thinking.


And I must have done something more wrong. You're getting 10x the
bang I was thinking.

The Pavouk schematic has a slide switch on it, near the
bridge rectifier. That switch changed the operating mode
of the capacitor charging circuit. The end result is,
the caps end up charged to roughly the same voltage, whether
in one type of country or the other. That slide switch corresponds
to the red slide switch with window that shows you've selected
115 or 230V operation.

The author of this article, calls what happens to the caps,
a "voltage doubler circuit".

http://www.hardwaresecrets.com/printpage/Anatomy-of-Switching-Power-Supplies/327

Paul
 
P

Paul

Bill said:
I have to say that after studying that schematic, and tidying up the
author's English translation/explanation of the overall circuit operation,
the design of the circuit seems a bit difficult for me to grasp. Am I the
only one who finds the design a bit difficult to completely understand (I
mean above the somewhat superficial explanation provided)?

You could try one of the links at the bottom of that page, which
shows a simple inverter and works up from there.

https://translate.google.com/transl.../www.belza.cz/swmodeps/compow1.htm&edit-text=

Paul
 
Ad

Advertisements

D

DK

I would appreciate any suggestions:

Starting last night, the desktop spontaneously shuts down.
All signs of power failure - just sudden turn off of everything.
Happened overnight and three times today. ...
The only thing I can think about is power supply but I am not sure
how to troubleshoot it. (I have Nexus Value 430 PSU that is 5 years
old). Another guess is a motherboard but, once again, not really
sure what I can do to pinpoint the problem (Asus M4N68T-M-V2
that is 2.5 years old).

Thanks, everyone who replied!

Oh well, it does seem like it was a PSU issue. Pretty disappointing
considering that 5 years is not that old and Nexus does have stellar
reputation (which I why I bought it to begin with).

Same day, went to Bestbuy and picked up the only decent PSU they
had, Corsair CX750M. I don't think I will ever need to 750 W and I definitely
overpaid ($104 vs $80 from Amazon). Plus, the reviews are somewhat
mixed. So far so good, running Prfime95 for two hours without any issues.
Will let it go overnight. I should probably return this and get some fanless
SeaSonic for the same money.

Any recommendations/suggestions fro an alternative PSU?

Now the fun part (and a reason why I haven't posted this note earlier):

Changing the PSU, I killed my CPU and it took a while to get the
replacement (surprisingly, theose Phenom 4X are not that common
these days). I do feel someone stupid about it but here is how it went:

Needed to take off the big CPU cooler in order to slide in/out power
supply. Unscrewed fully, pulled it up - no go, seemed to have glued
to the CPU or something. The design of Noctua NH-C12P SE14
cooler precluded easy prying it off. So, figuring that the CPU is locked
in the socket with that lever arm, I pulled stronger... The sad result:

The CPU was ripped out from the socket, still being stuck to the cooler.
Few pins were bent and one somehow came completely off the CPU.
Try as I might, I wasn't able to solder it back and created a whole
mess of nearby pins instead. In the end, the CPU heated up but no
POST. Sigh.

My own fault, I know. Still, I wish that thermal paste supplied with the
cooler did not have this wonderful glue-like quality. In any case, the
new CPU is a little faster (Phenom II 4X 955 BE vs old 840 plain), so
for an extra $80 and some hassle I got myself a little upgrade. For
now, all is good.

DK
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Top