Gateway problem - CPU fan spins momentarily on power-on

[David H. Cook]

I'm trying to repair a Gateway desktop machine (model MFATXPNT NMZ
500S).

The power-supply WAS faulty (I plugged in my 'tester' and got a RED
[BAD]
indicator, as well as ONE of the green LEDs was off, which was the -5V
light.)

So, I've now replaced the power-supply, with a new 400-watt supply.
(The NEW
power-supply now tests great...all the green LEDs lit.)

But, when I push the front-panel 'power-on' button, all that happens is
that the CPU-fan spins momentarily (from what appears to be only
about 1-seconds-worth of power.

The other fans, including the power-supply fans do NOT come on.
(This is with the power-connections to the motherboard in place.)

My first GUESS is that maybe the motherboard got 'hit' (damaged)
during the FAILURE-event of the original power-supply. Can anyone
comment about the 'odds' (frequency of times) that a failed
power-supply
is ALSO paired with motherboard-being-bad?

But, I'm ALSO wondering whether it could instead be just a
failure (damaged) front-panel-POWER-SWITCH.

(Or even something more obscure...like a 'case-intrusion' feature?!
This is NOT my machine, so I have no PRIOR knowledge of how
this machine has behaved in the past, etc.)

Or maybe even something else.

How do I 'bypass' the front-panel power-switch? Can I just remove
the two wires from the front panel-power-switch to the motherboard
and then just JUMPER those two motherboard pins (and then
use the switch on the power-supply as the (TEMPORARY)
on-off switch?

Any/all ideas on how to proceed to trouble-shoot will be
greatly appreciated!

TIA....

Dave

 

[paulmd]

I'm trying to repair a Gateway desktop machine (model MFATXPNT NMZ
500S).

The power-supply WAS faulty (I plugged in my 'tester' and got a RED
[BAD]
indicator, as well as ONE of the green LEDs was off, which was the -5V
light.)

So, I've now replaced the power-supply, with a new 400-watt supply.
(The NEW
power-supply now tests great...all the green LEDs lit.)

But, when I push the front-panel 'power-on' button, all that happens is
that the CPU-fan spins momentarily (from what appears to be only
about 1-seconds-worth of power.

Hang on. If you've got one of those atx power testers, the thing to
watch out is that many oems (hp, dell, etc) have a non-standard power
supply. For instance, no -5v line was ever installed. Look at the
diagram printed on the power supply. If you see no -5, then it's- not
bad. What's happening, i think is that your new power supply DOES have
a -5, and your motherboard doesn't like it.

I suspect that your original power supply is good.

The other fans, including the power-supply fans do NOT come on.
(This is with the power-connections to the motherboard in place.)

My first GUESS is that maybe the motherboard got 'hit' (damaged)
during the FAILURE-event of the original power-supply. Can anyone
comment about the 'odds' (frequency of times) that a failed
power-supply
is ALSO paired with motherboard-being-bad?

With a black and silver emachine, they approach 100%. Never seen that
on ANY other computer, however.

But, I'm ALSO wondering whether it could instead be just a
failure (damaged) front-panel-POWER-SWITCH.

Not if you get any response at all. You indicated that pressing the
button makes fans spin for about 1 sec.

(Or even something more obscure...like a 'case-intrusion' feature?!
This is NOT my machine, so I have no PRIOR knowledge of how
this machine has behaved in the past, etc.)

No. Never seen a chassis intrusion actually physically stop a machine
from running.

Or maybe even something else.

How do I 'bypass' the front-panel power-switch? Can I just remove
the two wires from the front panel-power-switch to the motherboard
and then just JUMPER those two motherboard pins (and then
use the switch on the power-supply as the (TEMPORARY)
on-off switch?

Use a flat blade screwdriver to short the pins.

Any/all ideas on how to proceed to trouble-shoot will be
greatly appreciated!

TIA....

Dave

First. Check the old power supply again. Make sure that it really is
bad. Proceed from there.

 

[Vanguard]

I'm trying to repair a Gateway desktop machine (model MFATXPNT NMZ
500S).

The power-supply WAS faulty (I plugged in my 'tester' and got a RED
[BAD]
indicator, as well as ONE of the green LEDs was off, which was the -5V
light.)

So, I've now replaced the power-supply, with a new 400-watt supply.
(The NEW
power-supply now tests great...all the green LEDs lit.)

But, when I push the front-panel 'power-on' button, all that happens
is
that the CPU-fan spins momentarily (from what appears to be only
about 1-seconds-worth of power.

You sure you plugged the CPU fan's power connector to the motherboard
header marked "CPU" and not to "FAN" or somewhere else (like using a
Y-adapter power cable to hook into the cabling coming from the PSU)?
The BIOS is probably configured to shutdown the system if it detects
that the CPU fan is not spinning. If you don't connect the CPU fan to
the CPU header, the BIOS sees no reading for the RPM of the CPU fan and
immediately shuts down the power (because AMDs have less protection and
can burn up in seconds and take the mobo with it whereas Intels will
reduce their duty cycle to reduce power simply by effectively running
slower).

Another possibility is that you have the wrong fan type on the CPU's
heatsink. Does it have 3 leads? One lead is the "sense" wire to tell
the BIOS what is the RPM for the fan. Did you make sure you didn't
reverse the connector on the header? Or that the fan's connector is
fully pressed onto the header? Could be the fan is defective and you
need to get a new one (i.e., no RPM output). I've bought fans that
sometimes turned out to look bad either the sense wire didn't get
connected (loose connection, reversed connection, off by 1 pin for the
connection, too low a speed reported by the sense output of the fan, or
no output on the sense wire).

What do you have connected to the PSU when you perform your power-on
test? The motherboard, system RAM, video card, and one hard drive
should be plugged into the PSU. Although you'd think that the hard
drive was superfluous, some PSU's won't power on unless there is some
amperage drain, like from a hard drive. In fact, if you have a good
tester, it should put a 10W drain on the PSU to make sure enough is
being drawn for the PSU to figure it has something to power.

Have you checked that the PS-ON pin on the mobo power header is
connecting to the cabling from the PSU? It shorts to ground and, when
the PSU is connected, tells the PSU that it is connected to a mobo.
Without shorting PS-On (pin 14, I think) to ground, the PSU figures that
it is not connected to a mobo and won't power up. Check for bent pins.

 

[Loren Pechtel]

With a black and silver emachine, they approach 100%. Never seen that
on ANY other computer, however.

I've had a flock of RaidMax power supplies that died premature deaths
and more often than not took the MB with them. I also ended up
replacing a processor off one and a stick of memory off another. Other
than that I've never seen a power supply take anything with it.

 

[Loren Pechtel]

I'm trying to repair a Gateway desktop machine (model MFATXPNT NMZ
500S).

The power-supply WAS faulty (I plugged in my 'tester' and got a RED
[BAD]
indicator, as well as ONE of the green LEDs was off, which was the -5V
light.)

So, I've now replaced the power-supply, with a new 400-watt supply.
(The NEW
power-supply now tests great...all the green LEDs lit.)

But, when I push the front-panel 'power-on' button, all that happens is
that the CPU-fan spins momentarily (from what appears to be only
about 1-seconds-worth of power.

The other fans, including the power-supply fans do NOT come on.
(This is with the power-connections to the motherboard in place.)

My first GUESS is that maybe the motherboard got 'hit' (damaged)
during the FAILURE-event of the original power-supply. Can anyone
comment about the 'odds' (frequency of times) that a failed
power-supply
is ALSO paired with motherboard-being-bad?

But, I'm ALSO wondering whether it could instead be just a
failure (damaged) front-panel-POWER-SWITCH.

(Or even something more obscure...like a 'case-intrusion' feature?!
This is NOT my machine, so I have no PRIOR knowledge of how
this machine has behaved in the past, etc.)

Or maybe even something else.

How do I 'bypass' the front-panel power-switch? Can I just remove
the two wires from the front panel-power-switch to the motherboard
and then just JUMPER those two motherboard pins (and then
use the switch on the power-supply as the (TEMPORARY)
on-off switch?

Any/all ideas on how to proceed to trouble-shoot will be
greatly appreciated!

Every time I've had a machine that did this it was a failed power
supply. The supply "works" enough to light the lights on the tester
but it can't put out the power needed to run the board. The system
starts to spin up but the voltage isn't what it's supposed to be and a
safety on the board shuts it down. (This safety exists to keep the
computer from trying to run when it can't run reliably. Better to
shut down than do something insane like issue an errant write command
to the HD.)

 

[paulmd]

What I often do is attach a few hard drives to act as a load whten i
use a tester. Or just unplug the motherboard.

 

[Loren Pechtel]

What I often do is attach a few hard drives to act as a load whten i
use a tester. Or just unplug the motherboard.

What I would love is a tester that had a power knob on it--run the
power up and see what it can actually deliver.

 

[David H. Cook]

Just for completeness. (This is the 'base note' poster
again.)

Yes, I had NOT made any of those any potential errors
that were suggested as possibilities.

The final OUTCOME was that the NEW power supply I
had replaced the old one with was fine. The MOTHERBOARD
was indeed damaged and was (later) replaced by
someone else.

Thanks for all the excellent discussion and suggestions!

Cheers...

Dave