Power supply, Motherboard, or something else?

One of my desktops (built a few years ago) right now not only won't boot,
but won't even get to the bios. I get no video output at all. This is a
fanless system with fanless power supply so it's a bit harder to tell what's
going on. When I hit the power button, I can hear it click on. Some type of
power is getting to the MB, since the power light on the MB is on. When I
attach a case fan to the motherboard it starts for a second and then shuts
off. If I push and hold the power button, I can hear the power supply click
off. I don't have a spare power supply right now that isn't already
installed in a computer, so it will be harder for me to test the machine
with a different PS. Since I can hear the PS turn on and off and the fan
runs briefly, I suspect that it isn't the power supply and is likely the MB.

Any ideas about how I might figure out exactly what the problem is before I
start taking this thing apart?

Thanks

Jeff

On Thu, 1 Jan 2015 03:29:57 +1000, Jeff wrote:

One of my desktops (built a few years ago) right now not only won't boot,
but won't even get to the bios. I get no video output at all. This is a
fanless system with fanless power supply so it's a bit harder to tell what's
going on. When I hit the power button, I can hear it click on. Some type of
power is getting to the MB, since the power light on the MB is on. When I
attach a case fan to the motherboard it starts for a second and then shuts
off. If I push and hold the power button, I can hear the power supply click
off. I don't have a spare power supply right now that isn't already
installed in a computer, so it will be harder for me to test the machine
with a different PS. Since I can hear the PS turn on and off and the fan
runs briefly, I suspect that it isn't the power supply and is likely the MB.

Any ideas about how I might figure out exactly what the problem is before I
start taking this thing apart?

Either you get your crystal ball out or you take a power supply out of
one of your other systems and try it in this one. I'm assuming that if
you had a meter you would have used it.

--
Faster, cheaper, quieter than HS2
and built in 5 years;
UKUltraspeed http://www.500kmh.com/

On Thu, 1 Jan 2015 03:29:57 +1000, "Jeff" wrote:

One of my desktops (built a few years ago) right now not only won't boot,
but won't even get to the bios. I get no video output at all. This is a
fanless system with fanless power supply so it's a bit harder to tell what's
going on. When I hit the power button, I can hear it click on. Some type of
power is getting to the MB, since the power light on the MB is on. When I
attach a case fan to the motherboard it starts for a second and then shuts
off. If I push and hold the power button, I can hear the power supply click
off. I don't have a spare power supply right now that isn't already
installed in a computer, so it will be harder for me to test the machine
with a different PS. Since I can hear the PS turn on and off and the fan
runs briefly, I suspect that it isn't the power supply and is likely the MB.

Any ideas about how I might figure out exactly what the problem is before I
start taking this thing apart?

Been there, done that. But, no, your presumption the PS unit is good
isn't -- the MB is just as likely good and there are other issues (not
to rule out the PS). Having another PS in another computer, thing
that's cool about that, is it starts getting easier each time such
situations arise (damn, no spare). Best to pull the good/working one
out for a test pwr-up to see what's revealed, deductively/logically
speaking. Something different is just as likely to happen. Although
you could start simply in another direction, like pulling everything
on your fanless system, one at a time, that's not necessary for a boot
-- to include cleaning, at least reseating your memory contacts, along
with doublechecking all plugs are seated properly;- swap
plugs/contacts, as well, when duplicates are optionally available. A
fault might reveal itself and pop up. I've only had one, so far, that
I reduced to a grounding-issue from case standoffs. I take them right
down to a bootable MB on a towel, with a good PS, vid and keyboard;-
then if it doesn't boot, I might worry about a MB.

Never had a MB that "lost it," so to speak, although I've worn quite a
few down to the point they'll exhibited anomalies first, not fail. Any
MB I've "worn down into a state of senility" was way long past an
update, anyway. They're some pretty tough ol' beastards, the better
made brands, in my experience. (Didn't use to, but I'm now a firm
believer in researching/buying only Top Notch PS makes.)

Hey, I've gotten so bad, I only use one maybe two "thumb-screws" to
hold in a PS;- cases, hell, screw the side opposite/parallel to the
MB, which I never bother anymore putting back on. Not a hostile
environ, I try watch what I'm doing with open computers laying around.
They do run a little cooler and it's pretty easy getting in and out
when need to "change things around" a bit.

Pull your working PS, plug it in and see what happens. If it doesn't
come up, tear it down, try again with it rebuild it on a towel instead
of inside a case. (I've a spare jumper block and switches pulled from
an old case I was probably using for assembling on a towel;- they're
logical momentary-on states, anyway, and should be able to be shorted
for a jumpstart.)

Jeff wrote:
One of my desktops (built a few years ago) right now not only won't
boot, but won't even get to the bios. I get no video output at all. This
is a fanless system with fanless power supply so it's a bit harder to
tell what's going on. When I hit the power button, I can hear it click
on. Some type of power is getting to the MB, since the power light on
the MB is on. When I attach a case fan to the motherboard it starts for
a second and then shuts off. If I push and hold the power button, I can
hear the power supply click off. I don't have a spare power supply right
now that isn't already installed in a computer, so it will be harder for
me to test the machine with a different PS. Since I can hear the PS turn
on and off and the fan runs briefly, I suspect that it isn't the power
supply and is likely the MB.

Any ideas about how I might figure out exactly what the problem is
before I start taking this thing apart?

Thanks

Jeff

You need a multimeter.

The ATX PSU "comes in two pieces". Two separate switching circuits
exist in the supply. The +5VSB powers USB ports and supervisor logic
(any logic which turns the computer on and off). The other three rails
are the main power rails, and a bigger portion of the circuit
is associated with that.

ATX PSU

AC Input ------+--- +5VSB circuit --- controlled by switch on back
|
+--- +3.3/5.0/12V main section --- controlled by PS_ON#

On an Asus motherboard (with some new recent exceptions), there
will be a green LED which runs off +5VSB. So you can tell the
switcher for +5VSB is running. The +5VSB is a "supervisor voltage"
and powers the logic that makes the rest of the machine work.

Your "click" symptom suggests it's not an issue with getting PS_ON#
to work, and turn on the main rails.

+5VSB (0.0V level +5VSB
| means "run please") |
Pullup \_ Pullup
Resistor \ Resistor
| PS_ON# |
PWR X----+---- Motherboard ---- Open -------------------+- ... control
/ logic Collector (to of main
| GND X----+ Driver ATX + PSU
| | supply) | section
(Front GND GND
Panel
Switch - normally open, momentarily close to operate)

To test the PSU, you could

1) Connect a fan directly to a Molex or SATA 15 pin connector.
This usually requires an adapter cable you might not have in
your junk drawer. I actually bought some fans once, just to get
the adapter cable that came in the box :-) Running the fan directly
off a Molex, makes a simple way to check +12V is there. It does not
verify the exact voltage. For that you need...

2) Multimeter, harbor freight, $20

The multimeter, set to 20V full scale, and with the probes
in the volt/ohm red and black holes, can be used to check
the voltages. Note that the twits who designed this particular
meter, didn't use black plastic for the right-most ground terminal :-)
The 20V DC scale is on the upper right, near the blue button. It's
possible to get a quality meter for $20 - just check the reviews
for comments about whether the thing is crappy or not.

http://www.amazon.com/Etekcity®-Digi.../dp/B00KHP6EIK

My crystal ball guess, is the fanless PSU has died. While
it "clicked", perhaps one of the rails is weak, and there
is not sufficient voltage to run things.

My first IBM PC, the power supply failed on it. The 12V output
runs at 12V when I connect one 12V 0.1A fan to it. If I connect
two fans, or a hard drive, the 12V rail drops to 6V. Which
means the outputs are a bit weak. It's only able to make
about 5% of the output power it used to make. And I tested
by grounding PS_ON# directly, with a ground wire.

The PS_ON# control signal isn't purely digital. A level
of 5V on the line, keeps the supply turned off. A level
of 0.4-0.7V or so (logic low), turns it on. The "#" in the
signal name, means the signal is "active low". Now, it is
possible to cause a power supply to have a weak output,
by feeding that logic signal a 1.5V to 2.0V level or so.
It turns the supply "half on", using half the expected voltage.
It makes the 12V weak, and unable to "hold up" a motherboard.
So a weak supply isn't always a supply fault, but that
exact set of circumstances isn't too common.

I did a walkthrough on this with poster "Adam" in a recent
thread, which is where my stick art diagrams come from.
He used a multimeter, and since his fans would not run,
the immediate suspicion was motherboard (no working PS_ON#).
And a swapped motherboard, brought things up. Adam did a separate
power supply test (grounding PS_ON#, checking for signs
of power). But you can do something like that, with a
newly purchased multimeter, and check to see if proper
voltages are there after the "click".

A power supply does not need a relay to work. But some
of the supplies with no on/off switch at the back of the
computer, they use a relay to apply mains power to themselves.
That's a typical Apple trick, and a few Dells maybe, have that.
The unit "clicks" a couple seconds after being plugged in,
after which the green LED (motherboard power monitor LED),
may be visible, and proof that the supervisor voltage
is available. The +5VSB is also used to charge tablets
and the like, via the USB port, when the computer is
soft off.

This site has plenty of pictures and tables, so you can
probe stuff and check for voltages. You can even probe
the main ATX PSU connector when it is plugged in. You connect
the black wire, to an I/O screw on the back of the computer.
As the metal around the I/O area is grounded. You can then use
the red probe, and poke where the wire goes into the plastic
shell of the connector. Enough exposed metal exists in there, to
take electrical readings off each wire. While the PC is running.
By only having to hold the red probe in your hand, you're
less likely to short stuff out.

http://www.playtool.com/pages/psucon...onnectors.html

And these tests are a lot easier to do, if the motherboard
and PSU are pulled out of the box, and tested on your
kitchen table. With appropriate precautions being taken
so the video card doesn't fall over, or get pulled out of
the slot. Be very careful with the plugin cards, as they
can easily get pulled from a slot and damage things in the
process. Having the electronics on the table, makes it easier
to get meter readings, but also makes it easier to
damage a card or motherboard.

Some computer cases, make it virtually impossible to pull
a populated motherboard out of the case. There is a stiffener
bar with rivets, to hold the chassis square, which prevents
easy removal. On my latest build, I was able to lower the
whole thing into place, in spite of one of those bars. So
sometimes, you get lucky.

Paul

On Wed, 31 Dec 2014 15:11:03 -0500, Paul wrote:

One of my desktops (built a few years ago) right now not only won't
boot, but won't even get to the bios. I get no video output at all. This
is a fanless system with fanless power supply so it's a bit harder to
tell what's going on. When I hit the power button, I can hear it click
on. Some type of power is getting to the MB, since the power light on
the MB is on. When I attach a case fan to the motherboard it starts for
a second and then shuts off. If I push and hold the power button, I can
hear the power supply click off. I don't have a spare power supply right
now that isn't already installed in a computer, so it will be harder for
me to test the machine with a different PS. Since I can hear the PS turn
on and off and the fan runs briefly, I suspect that it isn't the power
supply and is likely the MB.

Any ideas about how I might figure out exactly what the problem is
before I start taking this thing apart?

Thanks

Jeff

You need a multimeter.

To be honest I don't think a meter will help in this situation. We know
the 5VSB is there because the led on the motherboard is on and we know
it's sufficient to turn on the PSU because the fan starts and then
stops and I don't think that will give time for a meter reading.

The OP could unplug the disc drive, I've seen them pull a psu down and
give these symptoms, a one week old 2TB Seagate drive earlier this year
(replaced under warranty).

--
Faster, cheaper, quieter than HS2
and built in 5 years;
UKUltraspeed http://www.500kmh.com/

Rodney Pont wrote:
On Wed, 31 Dec 2014 15:11:03 -0500, Paul wrote:

One of my desktops (built a few years ago) right now not only won't
boot, but won't even get to the bios. I get no video output at all. This
is a fanless system with fanless power supply so it's a bit harder to
tell what's going on. When I hit the power button, I can hear it click
on. Some type of power is getting to the MB, since the power light on
the MB is on. When I attach a case fan to the motherboard it starts for
a second and then shuts off. If I push and hold the power button, I can
hear the power supply click off. I don't have a spare power supply right
now that isn't already installed in a computer, so it will be harder for
me to test the machine with a different PS. Since I can hear the PS turn
on and off and the fan runs briefly, I suspect that it isn't the power
supply and is likely the MB.

Any ideas about how I might figure out exactly what the problem is
before I start taking this thing apart?

Thanks

Jeff
You need a multimeter.

To be honest I don't think a meter will help in this situation. We know
the 5VSB is there because the led on the motherboard is on and we know
it's sufficient to turn on the PSU because the fan starts and then
stops and I don't think that will give time for a meter reading.

The OP could unplug the disc drive, I've seen them pull a psu down and
give these symptoms, a one week old 2TB Seagate drive earlier this year
(replaced under warranty).


Sure. There's nothing wrong with an ad-hoc "try stuff"
approach. On a computer, this is called "simplification"
for want of a better word. Try removing stuff, a bit at
a time, and look for a change in symptoms.

You can also listen for beep codes, assuming the computer
has something connected to the SPKR front panel pins. Even
with pulling video card and RAM sticks, if you get a beep code
it tells you the CPU is getting power and the CPU is running
BIOS code. And that's half the motherboard tested right there.

So there are a ton of ad-hoc tests, and interesting results
to examine, to go further. It's all a question of whether
a person wants to write out a flow chart, every time this
happens :-)

And if you do volunteer a flow chart, it needs a lot of details.
For example, once I suggested to someone, they pull the CMOS
CR2032 battery and test it. And because I didn't give details
on how to get the battery out, they managed to ruin the battery
socket. That means I have to modify my suggested procedures
a bit, like specify the purchase of a multimeter, probe the
top surface of the CMOS battery and get a reading off it. As
that is less dangerous, and a person won't snap off the
battery socket while working on it. At least, at first.

Paul

It appears that it is the power supply. I was able to find a spare and after
plugging it in, the fan I plugged in stays on and so does the case's power
light. The problem I will have is that the machine is a discontinued Zalman
TNN300 totally noiseless with built-in power supply. I ran into one other
post on-line where someone with more electrical experience than I have was
speaking about attempting to repair the PS, but nothing else. I've emailed
the merchant who sold me the case to see whether they know of an option.
From what I gather, there was an external power supply for desktops made at
one point, but it looks like that was discontinued also.


Any ideas?
(hard for me to believe that someone couldn't figure out how to remove a
cmos battery! - sounds like someone who should even be opening the case)



"Paul" wrote in message ...

Rodney Pont wrote:
On Wed, 31 Dec 2014 15:11:03 -0500, Paul wrote:

One of my desktops (built a few years ago) right now not only won't
boot, but won't even get to the bios. I get no video output at all. This
is a fanless system with fanless power supply so it's a bit harder to
tell what's going on. When I hit the power button, I can hear it click
on. Some type of power is getting to the MB, since the power light on
the MB is on. When I attach a case fan to the motherboard it starts for
a second and then shuts off. If I push and hold the power button, I can
hear the power supply click off. I don't have a spare power supply right
now that isn't already installed in a computer, so it will be harder for
me to test the machine with a different PS. Since I can hear the PS turn
on and off and the fan runs briefly, I suspect that it isn't the power
supply and is likely the MB.

Any ideas about how I might figure out exactly what the problem is
before I start taking this thing apart?

Thanks

Jeff
You need a multimeter.

To be honest I don't think a meter will help in this situation. We know
the 5VSB is there because the led on the motherboard is on and we know
it's sufficient to turn on the PSU because the fan starts and then
stops and I don't think that will give time for a meter reading.

The OP could unplug the disc drive, I've seen them pull a psu down and
give these symptoms, a one week old 2TB Seagate drive earlier this year
(replaced under warranty).


Sure. There's nothing wrong with an ad-hoc "try stuff"
approach. On a computer, this is called "simplification"
for want of a better word. Try removing stuff, a bit at
a time, and look for a change in symptoms.

You can also listen for beep codes, assuming the computer
has something connected to the SPKR front panel pins. Even
with pulling video card and RAM sticks, if you get a beep code
it tells you the CPU is getting power and the CPU is running
BIOS code. And that's half the motherboard tested right there.

So there are a ton of ad-hoc tests, and interesting results
to examine, to go further. It's all a question of whether
a person wants to write out a flow chart, every time this
happens :-)

And if you do volunteer a flow chart, it needs a lot of details.
For example, once I suggested to someone, they pull the CMOS
CR2032 battery and test it. And because I didn't give details
on how to get the battery out, they managed to ruin the battery
socket. That means I have to modify my suggested procedures
a bit, like specify the purchase of a multimeter, probe the
top surface of the CMOS battery and get a reading off it. As
that is less dangerous, and a person won't snap off the
battery socket while working on it. At least, at first.

Paul

On Thu, 1 Jan 2015 03:29:57 +1000, "Jeff" wrote:

One of my desktops (built a few years ago) right now not only won't boot,
but won't even get to the bios. I get no video output at all. This is a
fanless system with fanless power supply so it's a bit harder to tell what's
going on. When I hit the power button, I can hear it click on. Some type of
power is getting to the MB, since the power light on the MB is on. When I
attach a case fan to the motherboard it starts for a second and then shuts
off. If I push and hold the power button, I can hear the power supply click
off. I don't have a spare power supply right now that isn't already
installed in a computer, so it will be harder for me to test the machine
with a different PS. Since I can hear the PS turn on and off and the fan
runs briefly, I suspect that it isn't the power supply and is likely the MB.

Any ideas about how I might figure out exactly what the problem is before I
start taking this thing apart?

A power supply that doesn't put out enough power or not good power
will show exactly this failure.

There's a short time that it's allowed to put out an inadequate
voltage but once that time is up there's a deadman switch that kills
it if the voltages are unacceptable.

It puts out enough power to run the fan but one or more rails aren't
to up the correct voltage when the timer runs out and the deadman
kills it.

(The purpose of the deadman is to kill it before the voltage gets far
enough off spec that the computer might do errant things--like write
crap to the HD.)

On Thu, 1 Jan 2015 08:59:15 +1000, "Jeff" wrote:

(hard for me to believe that someone couldn't figure out how to remove a
cmos battery! - sounds like someone who should even be opening the case)

I can't see the connection to the thread.

I have had the experience, though--in a laptop. Swapping it would be
a triviality, finding how to get to it is decidedly non-trivial. (The
service manual doesn't even say where it's hiding.)

Jeff wrote:
It appears that it is the power supply. I was able to find a spare and
after plugging it in, the fan I plugged in stays on and so does the
case's power light. The problem I will have is that the machine is a
discontinued Zalman TNN300 totally noiseless with built-in power supply.
I ran into one other post on-line where someone with more electrical
experience than I have was speaking about attempting to repair the PS,
but nothing else. I've emailed the merchant who sold me the case to see
whether they know of an option. From what I gather, there was an
external power supply for desktops made at one point, but it looks like
that was discontinued also.


Any ideas?

The sad part is, the state of the Zalman business right now.
They were bought by some other corporation, then the
other corporation had financial trouble. It's expected Zalman could
survive, but I don't know in the interim, what it might mean
for customer service. And whether you could still contact them
for suggestions.

That power supply has a fairly unique form factor. It could be
that some of the power components are mated to one side of the
supply, so the heat can flow into the case wall.

There are (or were) some fanless PSUs, up to around 400W. But
at least some of these, they're probably relying on airflow
from remaining fans in the computer case, to help the supply
meet the power rating. If the supply was put into a tight box,
it would likely overheat. Whereas the Zalman solution, is more
likely to be using conduction rather than convection.

Someone on the badcaps forum tried to work on a Zalman supply,
and couldn't figure it out. It appeared in their case, that
the Active PFC front end burned up. That could happen, if
the TNN300 was powered by a non-sine wave UPS. There have
been cases from when Active PFC first came out, where the
kind of UPS used, influenced the health of the power supply.
One of the side effects of Active PFC, is it places more
DC on the high voltage side of the supply. So instead of
320V on the hot side of the main cap, it might be 380V or so.
This is all part of how the active PFC stage is able to adjust
the phase angle of the current the supply draws. The naive
PFC design relies on the input waveform always being a sine
wave (as the PFC strives to draw a current waveform, which
matches the shape of the incoming voltage waveform). When a
square wave UPS is connected to one of those supplies, then the PFC
is trying to make the current draw look like a matching square
wave. Which to my way of thinking, could lead to some
interesting results (because there is still a filter on
the input stage, which has to eat the exotic waveform
coming from the PFC).

Now you know why Paul is careful not to buy Active PFC supplies.
Not because they're not good supplies, but Paul knows his
UPS is one of the bad kind :-) I have to make do with my
current contingent of supplies, because it would be
relatively hard to find one without some PFC considerations.
And I plan to get a few more years out of my $300 UPS.

Another source of power would be a Pico supply, but they're
not powerful enough for anything but small projects. You
would need a laptop load, or a mini-ITX motherboard, to be
in range of the power capabilities of one of these.

http://www.mini-itx.com/store/?c=10#picoPSU-160-XT

You could also place a regular ATX outside the Zalman, and
run 24 pin extender cables. A pathetic solution, but, it's
another way to do it.

http://www.amazon.com/StarTech-8-Inc.../dp/B000FL60AI

Fanless Seasonic, with modular cabling. Always check the
reviews on the fanless ones, to get some idea the kind
of electrical load they've been tested with (by te
reviewers). And yes, this is active PFC. I can't imagine
them not doing that now.

http://www.newegg.com/Product/Produc...82E16817151097

Paul