Yes wrote:
While troubleshooting a new HD, I just ran into a new problem of which
I'm not sure where the problem is.
When I press the on/off button for the pc on the front of the case, I see
it start to light up but within a fraction of a second, it shuts down is
the best way I can describe it - does not continue to boot up. When I
press the on/off button again, nothing happens.
When I turn off power completely (using the on/off button on the PSU),
wait awhile, turn back on the PSU power button, and then turn on the pc
using the on/off button on the front of the pc, I get the same problem
described above - pc lights up but immediately stops.
The problem occurred while I was troubleshooting a problem with a new HD.
My testing of the HD involved frequently turning the pc on and off in
order to switch the power cords and SATA cords among the HDDs.
My first thought was that the switch on the front of the pc broke, but
that doesn't seem likely because the pc will power up however briefly
when I press the switch.
Second thought is that there's a "fuse" ?? needing to be reset.
Third thought was wondering if the CMOS battery needs replacing.
Fourth thought was worrying if I need to replace my mobo.
Fifth thought was same question but about my PSU.
Sixth thought is wondering if I corrupted BIOS somehow. I would enter
BIOS to check if the new HD was listed in the SATA info, so I was always
entering and exiting BIOS and resuming booting up.
My notes about the pc are not handy. The mobo is an ASUS AMD mobo about
4 or 5 years old- I don't have its model number handy. The CPU, PSU, DVD/
CD device and graphics card are the same age. The HDDs vary in age, all
SATA. I have four. IIRC, the PSU is rated to handle 600W.
Thanks,
John
There is a protective fuse, but it's a slow blow. If
it was gone, the supply sections wouldn't come on at all.
In particular, with the fuse open, the +5VSB would not be present
before you push the button on the front. The fuse F1 is on
the upper left of the schematic below.
You have sufficient +5VSB, that it is powering the motherboard
supervisor circuitry and when you push the front button, it
asserts (grounds) the open-collector PS_ON# signal.
A secondary way to operate a supply, is to ground PS_ON# yourself.
This is what we do when making home-made power supply testers,
is use a switch to ground PS_ON#. I don't recommend leaving
hard drives you don't have backups of, on the SATA portion
while messing around. If you want to jiggle PS_ON# your own
self, boot with a floppy or CD. At most then, you could
blow up a $20 DVD drive.
After PS_ON# is asserted, the power supply has the internal
overcurrent detection disabled for the first 35 milliseconds.
This gives time for the output stage capacitors on the PSU
to charge up. (C30, upper right and so on).
http://www.pavouk.org/hw/en_atxps.html
After the 35 millisecond interval, the PSU starts checking
for overload. If something on the motherboard draws too much
current, the PSU shuts off. It uses +5VSB power to "record"
the failure internally, inside the PSU. Attempts to push the
button on the front at this point, are likely to fail. As soon
as you turn off the power at the back, and turn it on again,
that resets the recording of an overcurrent, and you can then
try the front button again.
Some old supplies, can become "weak" and produce hardly any
output at all (I have one here that will run a 12V 100mA fan,
but if you add a second fan, it causes the PSU to croak). It's
not clear why the overcurrent isn't working properly in
that case. The power supply can be "weak", if the primary side
switching wasn't working properly. Q1 and Q2 alternate applying
300VDC to the primary side of T3, and all the outputs
are tied via turns ratio, to producing some output values.
The diodes on the output side of T3, rectify the AC waveform
coming from transformer T3. The large output caps (like C30)
filter the ripple from the rectification action, making DC
voltages (12V, 5V, 3.3V and so on). The PSU design uses
"isolation" and transformers or optoisolators, to isolate
sections of the PSU for safety. The drive signals to run
the switching transistors, are transformer isolated. The idea
is to leave fewer paths that can electrocute the user (a
noble objective).
So what we know at the moment, is the motherboard did manage
to activate PS_ON#. We don't know if it's a good solid logic
low. The PS_ON# driver on the motherboard is notorious for
fouling up, even though the input spec on the PSU only
requires a couple milliamps of ground sinking to work.
Some of the motherboards from the year 2000 era were
using 74F series logic with 48mA drive for some reason,
"as if the motherboard makers knew something". You see,
the power supply maker couldn't give a damn about
making a "logic gate" to analyze PS_ON# and they use
various discrete circuits to check the logic level.
And there must be some failure conditions, that takes
it out on the motherboard PS_ON# driving chip.
I would agree with the previous posters, to swap in a
spare supply and re-test for symptoms. If you had a power
supply tester (one where you could put a decent load
on each output), then you could get some idea how
healthy the PSU is. Instead of that, simply buying
another supply and testing, is a step you can carry
out.
If you have a multimeter and know how to use it, you
can check the logic level on PS_ON#. Before pushing
the front button on the PC, PS_ON# should be the
same voltage as +5VSB (i.e. 5.0 volts). Once you
press the front button, you want a good solid
logic low to be seen on PS_ON#. Since your
supply only stays on for 35 milliseconds, it's
going to be pretty difficult to get that reading :-/
Only a digital storage scope would work for a job
like that.
*******
Here are the three generations of ATX PSU specs I know
of. If your supply has a 20 pin connector, then the
middle spec is the one you want. If the supply has
a 24 pin (I have a supply now with a solid 24 pin and
it isn't even split any more into 20+4 sections), then
you want the bottom spec. The first spec is just
to prove that at one time (~year 2000) there was
a -5V on the supply. The bottom two specs removed
-5V, and the motherboard only needs -12V if it has
an RS232 serial port on the motherboard (something
a TI 75232 uses).
Spec 1.1 has -5V on it. 20 pin connector.
http://web.archive.org/web/200304240...12V_PS_1_1.pdf
Spec 1.3 has -5V removed. 20 pin connector.
http://www.formfactors.org/developer...X12V_1_3dg.pdf
Spec 2.2 introduces... 24 pin connector.
http://www.formfactors.org/developer...public_br2.pdf
Some of the information in there, helps explain how
the interface works. Power_Good for example, is a way
for the PSU to tell the motherboard that the rails
are stable, and the motherboard can come out of
reset. You're not there yet, as at 35 milliseconds,
the supply winks out, and is never going to deliver
Power_Good logic 1 level. During the 35 millisecond
interval, the power supply isn't yet ready to assert
Power Good. That comes a bit later. There is a
timing diagram in the spec that might illustrate the
details of that.
While the motherboard VCore can latch off, when it detects
a failure, the fans would still be spinning. And I think
your computer case cooling fans are only "twitching" a bit
before they go off and stay off. So it's not a motherboard
VCore problem.
Swap in a spare PSU and retest.
Good luck,
Paul