Odd power behavior

  • Thread starter seriouslynonamesleft
  • Start date
S

seriouslynonamesleft

AMD Sempron 3000
WinPro DDR, 1 stick, 1gig
2 X SATA Western Digital Caviar, 80gig
Samsung dual layer DVD+RW drive
Silverstone SST-LC11 HTPC case w/ 300W PSU (also silverstone)
An extra 80mm fan I installed to improve air flow near the HDDs

In general, hitting the power button only works right after I switch
outlets. Sometimes I can get it to power on by repeatedly pushing the
power button, but this takes literally 20-30 tries.

However, when the thing powers up successfully, all hardware is
correctly detected and I can even boot Linux from a live CD.

Also worth noting: I have a Foxconn wireless PCI card. If I stick this
in the board, the machine won't start up at all.

This seems like an obvious PSU problem, but if that were the case I
would expect that the machine would die or malfunction in some way
after having been successfully powered on for some period of time. This
hasn't happened yet. I've also never heard of anyone having a PSU issue
quite like this so it makes me wonder. Any experience that any of you
could share would be greatly appreciated.
 
C

Conor

AMD Sempron 3000
WinPro DDR, 1 stick, 1gig
2 X SATA Western Digital Caviar, 80gig
Samsung dual layer DVD+RW drive
Silverstone SST-LC11 HTPC case w/ 300W PSU (also silverstone)
An extra 80mm fan I installed to improve air flow near the HDDs

In general, hitting the power button only works right after I switch
outlets. Sometimes I can get it to power on by repeatedly pushing the
power button, but this takes literally 20-30 tries.

However, when the thing powers up successfully, all hardware is
correctly detected and I can even boot Linux from a live CD.

Also worth noting: I have a Foxconn wireless PCI card. If I stick this
in the board, the machine won't start up at all.

This seems like an obvious PSU problem, but if that were the case I
would expect that the machine would die or malfunction in some way
after having been successfully powered on for some period of time. This
hasn't happened yet. I've also never heard of anyone having a PSU issue
quite like this so it makes me wonder. Any experience that any of you
could share would be greatly appreciated.
Your PSU is working at its limit and the PCI card is the straw breaking
the camels back. You'd find that if you disconnected a drive, it'd all
work OK even with the PCI card in.

Get a bigger PSU.
 
O

OSbandito

AMD Sempron 3000
WinPro DDR, 1 stick, 1gig
2 X SATA Western Digital Caviar, 80gig
Samsung dual layer DVD+RW drive
Silverstone SST-LC11 HTPC case w/ 300W PSU (also silverstone)
An extra 80mm fan I installed to improve air flow near the HDDs

In general, hitting the power button only works right after I switch
outlets. Sometimes I can get it to power on by repeatedly pushing the
power button, but this takes literally 20-30 tries.

However, when the thing powers up successfully, all hardware is
correctly detected and I can even boot Linux from a live CD.

Also worth noting: I have a Foxconn wireless PCI card. If I stick this
in the board, the machine won't start up at all.

This seems like an obvious PSU problem, but if that were the case I
would expect that the machine would die or malfunction in some way
after having been successfully powered on for some period of time. This
hasn't happened yet. I've also never heard of anyone having a PSU issue
quite like this so it makes me wonder. Any experience that any of you
could share would be greatly appreciated.
The wireless card thing sounds like a separate issue. Try simple things
first. Throw a jumper across the case start-switch. If the rig starts,
the switch is bad.
 
O

OSbandito

Conor said:
Your PSU is working at its limit and the PCI card is the straw breaking
the camels back. You'd find that if you disconnected a drive, it'd all
work OK even with the PCI card in.

Get a bigger PSU.

Yeah, Conor may well be right. I just saw the 300W rating on yer psu.
That's too little capacity for any modern box. My kid had a similar
psu--a resistor overheated, the psu fan stopped and the processor fried.
 
R

Rod Speed

(e-mail address removed) wrote
AMD Sempron 3000
WinPro DDR, 1 stick, 1gig
2 X SATA Western Digital Caviar, 80gig
Samsung dual layer DVD+RW drive
Silverstone SST-LC11 HTPC case w/ 300W PSU (also silverstone)
An extra 80mm fan I installed to improve air flow near the HDDs
In general, hitting the power button only works right after I switch
outlets. Sometimes I can get it to power on by repeatedly pushing
the power button, but this takes literally 20-30 tries.
However, when the thing powers up successfully, all hardware
is correctly detected and I can even boot Linux from a live CD.
Also worth noting: I have a Foxconn wireless PCI card. If
I stick this in the board, the machine won't start up at all.

That makes it a bit more likely to be a motherboard problem
rather than a power supply problem particularly if you still get
that effect with the power switch with no drives plugged in.
This seems like an obvious PSU problem,

Not necessarily with the effect you get with the wireless card if its
only seen with the wireless card and it isnt just a simple load effect.
but if that were the case I would expect that the machine
would die or malfunction in some way after having been
successfully powered on for some period of time.

Nope, not if the problem is with the power supply responding
to the motherboard signal to turn on. Once its turned on, that
line isnt relevant anymore so there is no reason to malfunction.
This hasn't happened yet. I've also never heard
of anyone having a PSU issue quite like this

It is something that has been seen with some PSUs
with the exception of the wireless card effect.
so it makes me wonder. Any experience that any
of you could share would be greatly appreciated.

I doubt its the PSU myself, because of the wireless
card effect if you have described it accurately.

I'd have a look at the PS_ON# line with a CRO to see what
is actually happening there, that should tell you definitively
whether its a motherboard or PSU problem and may well
provide some useful info on the wireless card effect too.
Even a multimeter on that line may be useful.
 
R

Rod Speed

Conor said:
Your PSU is working at its limit and the PCI card is the straw
breaking the camels back. You'd find that if you disconnected a
drive, it'd all work OK even with the PCI card in.
Get a bigger PSU.

Seems a bit unlikely that if its that marginal, it
always works fine once it has started successfully.

Possible that overload protection is too sensitive in the PSU
and the wireless card is enough to push THAT over the limit.
 
A

Andy

AMD Sempron 3000
WinPro DDR, 1 stick, 1gig
2 X SATA Western Digital Caviar, 80gig
Samsung dual layer DVD+RW drive
Silverstone SST-LC11 HTPC case w/ 300W PSU (also silverstone)
An extra 80mm fan I installed to improve air flow near the HDDs

In general, hitting the power button only works right after I switch
outlets. Sometimes I can get it to power on by repeatedly pushing the
power button, but this takes literally 20-30 tries.

However, when the thing powers up successfully, all hardware is
correctly detected and I can even boot Linux from a live CD.

Also worth noting: I have a Foxconn wireless PCI card. If I stick this
in the board, the machine won't start up at all.

Sounds like the power supply standby power output +5Vsb does not have
enough capacity.
 
S

seriouslynonamesleft

Thanks for all your suggestions. Here's an update on what I've tried
(and the result):

1) Disconnected everything but the CPU and RAM from the mobo: no
difference
2) Switched out the Silverstone PSU for another 300W PSU from an old
machine: system starts successfully every time
3) Re-connected SATA drives, DVD drive, extra case fan: system starts
successfully every time
4) Put the PCI riser back in: : system starts successfully every time
5) Put the wifi card back in: system won't start for anything

This is the kind of behavior I'd expect if the PSU just didn't have
enough juice for all my components. I suspect the Silverstone PSU is
faulty or damaged.
 
R

Rod Speed

seriouslynonamesleft said:
Thanks for all your suggestions. Here's an update on what I've tried (and the result):
1) Disconnected everything but the CPU and RAM from the mobo: no difference
2) Switched out the Silverstone PSU for another 300W PSU from an old
machine: system starts successfully every time
3) Re-connected SATA drives, DVD drive, extra case fan: system starts
successfully every time
4) Put the PCI riser back in: : system starts successfully every time
5) Put the wifi card back in: system won't start for anything

That last may be an entirely separate problem. The obvious test
for that is to try with no hard drives and the wifi card. Its its a power
problem it should start fine and only be able to boot a CD etc.
This is the kind of behavior I'd expect if the PSU just
didn't have enough juice for all my components.

Yes, but that isnt the only possibility.
I suspect the Silverstone PSU is faulty or damaged.

And that you have a separate problem with the wifi card most
likely unless it starts fine with that installed and no hard drives.
 
S

seriouslynonamesleft

Sure enough, when I disconnect the HDDs and plug the wifi card in, the
machine powers on but doesn't boot. It's not the riser either. I
switched out the wifi card with another, more power-hungry PCI
component and the system boots fine.
 
J

JAD

seriouslynonamesleft said:
Sure enough, when I disconnect the HDDs and plug the wifi card in, the
machine powers on but doesn't boot.

doesn't 'boot' or do you mean POST? its not going to boot if the HD is
disconnected.
Leave EVERYTHING completely unpluged except memory- video- cpu -PSU, do you
POST?
Yes? Use XP CD and connect one optical drive, set to boot from CD rom first,
can you boot?


It's not the riser either. I
 
S

seriouslynonamesleft

You're right. I meant to say it doesn't POST. However it is possible to
boot an OS with no HDDs connected using a LiveCD that loads entirely in
RAM (e.g. Puppy Linux).

Thanks for your suggestion. This was actually the first thing I tried
(see step 1 in my earlier post). I left out two details: a) my mobo has
an integrated video card and that's all I'm using b) the mobo was still
being powered by the Silverstone PSU which I believed to be defective.
 
J

johns

You have bad electrolytics in the start circuit
of your power supply. I've replaced them in
8 psupplies this year. Getting to be a real PITA.

johns
 
J

JAD

seriouslynonamesleft said:
You're right. I meant to say it doesn't POST. However it is possible to
boot an OS with no HDDs connected using a LiveCD that loads entirely in
RAM (e.g. Puppy Linux).

Thanks for your suggestion. This was actually the first thing I tried
(see step 1 in my earlier post). I left out two details: a) my mobo has
an integrated video card and that's all I'm using b) the mobo was still
being powered by the Silverstone PSU which I believed to be defective.

90% probability its the PSU, if that meets your criteria to make the
purchase, I would.
seriouslynonamesleft said:
Sure enough, when I disconnect the HDDs and plug the wifi card in, the
machine powers on but doesn't boot.

doesn't 'boot' or do you mean POST? its not going to boot if the HD is
disconnected.
Leave EVERYTHING completely unpluged except memory- video- cpu -PSU, do you
POST?
Yes? Use XP CD and connect one optical drive, set to boot from CD rom first,
can you boot?


It's not the riser either. I
switched out the wifi card with another, more power-hungry PCI
component and the system boots fine.

Rod Speed wrote:

Thanks for all your suggestions. Here's an update on what I've
tried
(and the result):
1) Disconnected everything but the CPU and RAM from the mobo: no difference
2) Switched out the Silverstone PSU for another 300W PSU from an old
machine: system starts successfully every time
3) Re-connected SATA drives, DVD drive, extra case fan: system starts
successfully every time
4) Put the PCI riser back in: : system starts successfully every time
5) Put the wifi card back in: system won't start for anything

That last may be an entirely separate problem. The obvious test
for that is to try with no hard drives and the wifi card. Its its a power
problem it should start fine and only be able to boot a CD etc.

This is the kind of behavior I'd expect if the PSU just
didn't have enough juice for all my components.

Yes, but that isnt the only possibility.

I suspect the Silverstone PSU is faulty or damaged.

And that you have a separate problem with the wifi card most
likely unless it starts fine with that installed and no hard drives.


Rod Speed wrote:
 
S

seriouslynonamesleft

I did some research based on your diagnosis and found this:

http://en.wikipedia.org/wiki/Capacitor_plague

"The most common method of identifying capacitors which have failed
because of bad electrolyte is visual inspection. Such a capacitor will
show one or more of these symptoms: [...] Not turning on all the time;
having to hit reset or try turning the computer on again."

And that's exactly what's happening.

While there isn't any crusty brown residue (as described under
"Symptoms" on the page above) inside the PSU but there is a white
substance which looks to have leaked from somewhere. Could this be
electrolyte?

This is all academic now since I'm obviously going to have to replace
the PSU. I'm just curious. Thanks again for all the good responses!
 
J

JAD

seriouslynonamesleft said:
I did some research based on your diagnosis and found this:

http://en.wikipedia.org/wiki/Capacitor_plague

"The most common method of identifying capacitors which have failed
because of bad electrolyte is visual inspection. Such a capacitor will
show one or more of these symptoms: [...] Not turning on all the time;
having to hit reset or try turning the computer on again."

And that's exactly what's happening.

While there isn't any crusty brown residue (as described under
"Symptoms" on the page above) inside the PSU but there is a white
substance which looks to have leaked from somewhere. Could this be
electrolyte?

This is all academic now since I'm obviously going to have to replace
the PSU. I'm just curious. Thanks again for all the good responses!


once the fluid has leaked I replace the hardware, I like to catch cap
problems at the 'buldging' stage. I'm more confident that they have not been
too far out of spec and less damage further down the circuit.
 

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

Similar Threads


Top