Loud Click then hanging - Power Supply?

[Robert Heiling]

This happened to me about 3 weeks ago and I thought at the time that I must have
lost one of my hard drives, but I hadn't and they've all been fine and passed
testing. What happens is that I will hear a loud click come from the box and
unpredictable events will follow such as programs not responding or mouse-cursor
stuck. More clicks can follow until the system is totally hung. The disk
activity light is on solid when this is happening. In retrospect, it is the same
loud click I hear when I shut down the system, so I assume the power supply is
doing this.

The system is a Compaq Presario bought at Fry's as a refurb and placed in
service Dec 2004. I think the "refurb" bit was simply a legal sales technicality
and that the system was unused.
CPU Type Intel Celeron D 325, 2533 MHz (19 x 133)
Motherboard Name MSI Gamila/Giovani/Neon Series
Motherboard Chipset Intel Brookdale-G i845GEV
System Memory 759 MB (PC2700 DDR SDRAM)
BIOS Type Award (08/05/04)
IDE Controller Intel(r) 82801DB Ultra ATA Storage Controller-24CB
SCSI/RAID Controller Tekram DC-395U/UW or DC-315/U PCI Ultra SCSI Host Adapter
V3.02
Floppy Drive Floppy disk drive
Disk Drive IDE1-Sl ST340015A (40 GB, 5400 RPM, Ultra-ATA/100)
Disk Drive IDE0-Ma WDC WD1200JB-00GVA0 (111 GB, IDE)
Disk Drive IDE0-Sl WDC WD2500JB-00GVA0 (232 GB, IDE)
Optical Drive DVDRW IDE 16X
SMART Hard Disks Status OK
Power Supply Liteon PS5251-08HT 250W

I was able to catch it while in the Bios and had it set to detect one of the
drives and was pressing return so that it would do so. There was a click, I
pressed return to detect, and it didn't work that time and the system hung. In
the past, the power connection on the drive at IDE-Sl has been flaky, no matter
what power connector I use and that might be part of this problem as I opened
the case late last night to remove a front panel section and get more airflow. I
might have upset that connector although I had been trying hard to avoid it. In
any case, what would let it run for such a long time (1 hour+) before failing
and what would be the mechanism in the PS for selectively dropping power?

Bob

 

[saturnlee]

Loud Click is the sound of exploding capacitors. Some of the capacitors
failed and exploded. You can take a look at the motherboard and power
supply and see if something got blow up.

you can also check www.badcap.net/forum
and see if this has happened before.

 

[saturnlee]

I think Loud Click is from a dying hard drive. The chip on the
controller board or bearing got overheated, so it fails after 1 hours
+.
I saw it couple of times.
Sometimes the hard drive is good temporary after it 's in a freezer for
30min-1 hr(in a bag). This is the trick i use to rescue the data.

 

[Robert Heiling]

I think Loud Click is from a dying hard drive. The chip on the
controller board or bearing got overheated, so it fails after 1 hours
+.
I saw it couple of times.
Sometimes the hard drive is good temporary after it 's in a freezer for
30min-1 hr(in a bag). This is the trick i use to rescue the data.

The capacitors & hard drives are just fine and I've been running everything just
fine since I reseated that HD power plug this morning. It will happen again
though the next time I go inside the case and do anything. I'm looking for
advice from anyone who understands how power supplies operate and why I was
getting power selectively stopped for different devices while remaining on for
others. At least that's what it looks like was happening.

Bob

 

[Andy]

The capacitors & hard drives are just fine and I've been running everything just
fine since I reseated that HD power plug this morning. It will happen again
though the next time I go inside the case and do anything. I'm looking for
advice from anyone who understands how power supplies operate and why I was
getting power selectively stopped for different devices while remaining on for
others. At least that's what it looks like was happening.

Bob

Sounds like some of the socket contacts in the drive power connectors
may be too loose, causing intermittent connection. I recently bought
an Antec SP-350 power supply. When I plugged the connectors into the
hard drives, I noticed they went in too easily. So I removed the
contacts from the plastic housing, and when I plugged the individual
contacts onto the power pins on the drive, many of them were so loose
that they slipped on without any physical resistance. I used pliers to
tighten up the socket contacts, and the power supply is working fine.

 

[Robert Heiling]

Sounds like some of the socket contacts in the drive power connectors
may be too loose, causing intermittent connection. I recently bought
an Antec SP-350 power supply. When I plugged the connectors into the
hard drives, I noticed they went in too easily. So I removed the
contacts from the plastic housing, and when I plugged the individual
contacts onto the power pins on the drive, many of them were so loose
that they slipped on without any physical resistance. I used pliers to
tighten up the socket contacts, and the power supply is working fine.

Sounds like a good fix for that. My situation seems to be the converse of that
as the connectors work fine with other drives and it doesn't matter which
connector I use with that one specific drive. There's something funny about the
power connector on that WD250. It's all still running fine since I reseated it
yesterday morning, but I'd really like to understand what it is that the PS does
when the drive is in trouble mode.

Bob

 

[larry moe 'n curly]

Sounds like a good fix for that. My situation seems to be the converse of that
as the connectors work fine with other drives and it doesn't matter which
connector I use with that one specific drive. There's something funny about the
power connector on that WD250. It's all still running fine since I reseated it
yesterday morning, but I'd really like to understand what it is that the PS does
when the drive is in trouble mode.

I experienced the same problem with one HD. You can tighten the PSU
connectors without removing them from the plastic shell by prying
between each metal pin and the plastic around it. Notice that those
pins are tubes split down one side, and it's best to pry 90 degrees
from that split.

Liteon PSUs are really good. Maybe the loose connector prevented the
HD from getting enough +5V power. In my case the HD would actually
stop spinning.

 

[Robert Heiling]

I experienced the same problem with one HD. You can tighten the PSU
connectors without removing them from the plastic shell by prying
between each metal pin and the plastic around it. Notice that those
pins are tubes split down one side, and it's best to pry 90 degrees
from that split.

But the PS's connectors are not a problem with any other drive I've used them
with. It's just that one drive that I finally narrowed it down to and it would
have to have smaller diameter (faulty?) power pins to be loose and for that
trick to be necessary. It could be that there is something wrong/loose with the
drive itself at that point and I'll have to pull it to really check it out, but
it's so easy to let it go when it's only happened twice and runs fine until I go
in the case to do something else and it gets wiggled by accident.

Liteon PSUs are really good. Maybe the loose connector prevented the
HD from getting enough +5V power. In my case the HD would actually
stop spinning.

That may be what's happening to mine and causes selective windows tasks to hang.
It's just that the PS seems to be doing something intelligent in terms of
selectively turning various power off and I've never read anything about PS's
doing that sort of thing. That click I hear is the same as at shutdown.

Bob

 

[mxh]

But the PS's connectors are not a problem with any other drive I've used
them
with. It's just that one drive that I finally narrowed it down to and it
would
have to have smaller diameter (faulty?) power pins to be loose and for
that
trick to be necessary. It could be that there is something wrong/loose
with the
drive itself at that point and I'll have to pull it to really check it
out, but
it's so easy to let it go when it's only happened twice and runs fine
until I go
in the case to do something else and it gets wiggled by accident.


That may be what's happening to mine and causes selective windows tasks to
hang.
It's just that the PS seems to be doing something intelligent in terms of
selectively turning various power off and I've never read anything about
PS's
doing that sort of thing. That click I hear is the same as at shutdown.

Just a thought, but do you have your HDs set to go to sleep in power
management? If so, perhaps an experiment would be to set power management to
'none' to see if the issue clears up.

mxh

 

[Robert Heiling]

Just a thought, but do you have your HDs set to go to sleep in power
management? If so, perhaps an experiment would be to set power management to
'none' to see if the issue clears up.

That *is* an interesting thought as it could account for the seemingly
intelligent PS actions. It's not the way it was meant to work, but that could
explain the click part and what is probably loss of power at the same time.

This has only happened while running Win2k and I have all those options turned
off both in the bios and in Win2k - no hibernate, no nothing. What was a strange
coincidence 2 weeks ago when this first happened (and my memory of exactly what
happened when and in what order cannot be trusted) was that I finally got the
Fedora Core installation DVD to run by passing it "linux acpi=off". All other
option tries had resulted in kernel panic and I hadn't been able to make the
installation.

However, this is all complicated by the fact that I was installing that FC4 on
the WD250 HD in question and had set things up so that it was the only HD
running so that I could let the installation install grub on the mbr. When I
reconnected all drives again, I figured that I hadn't gotten that power
connector back on right. That was the first of two times I've had this click
problem.

Bob

 

[Mechbook]

Yup.. it sounds like a faulty connection to me. Have you trie
switching out the power connection with another device? If you don
use your floppy drive much, switch power plugs out with that

 

[mxh]

That *is* an interesting thought as it could account for the seemingly
intelligent PS actions. It's not the way it was meant to work, but that
could
explain the click part and what is probably loss of power at the same
time.

This has only happened while running Win2k and I have all those options
turned
off both in the bios and in Win2k - no hibernate, no nothing. What was a
strange
coincidence 2 weeks ago when this first happened (and my memory of exactly
what
happened when and in what order cannot be trusted) was that I finally got
the
Fedora Core installation DVD to run by passing it "linux acpi=off". All
other
option tries had resulted in kernel panic and I hadn't been able to make
the
installation.

However, this is all complicated by the fact that I was installing that
FC4 on
the WD250 HD in question and had set things up so that it was the only HD
running so that I could let the installation install grub on the mbr. When
I
reconnected all drives again, I figured that I hadn't gotten that power
connector back on right. That was the first of two times I've had this
click
problem.

Bob

I had a few other thoughts, but after going back through the thread, they
are rendered irrelevant. The fact that it has only happened after you've
been in the case does tend to make one think that the power connector is
relevant, or, as you've implied, the PS is using its intelligent abilities
for evil.

However, since this only occurs with this drive, one could surmise that it
is indeed an issue with the power connection on the drive itself (perhaps
some internal loose connection in the drive that, when disturbed,
misbehaves). If you have a spare drive, you could try cloning the problem
drive to the spare to see if the same problems result. At least that would
remove or confirm the possiblity of it being an HD issue (if it only happens
on the drive in question...). Barring that, perhaps WD would RMA it and send
you a new one (maybe). My apologies for being able to submit only wild
guesses, but you do seem to have an elusive issue here.

mxh

 

[kony]

You need to pull the drive and check it.



Yes this is what seems to be happening based on the info
provided.


It is very unlikely that the PSU is doing anything
intelligent, you had assumed (??) that was the click but
typically it is the HDD when it shuts off, which it would do
when power is disconnected (for whatever reason, a fault or
system shutdown).

For the PSU to do something intelligent, it would have to
gather data. It has no feedback for this from a HDD, at
most it can monitor voltage and if the voltage were too bad,
it would simply shut off (and the whole system does too).

That *is* an interesting thought as it could account for the seemingly
intelligent PS actions. It's not the way it was meant to work, but that could
explain the click part and what is probably loss of power at the same time.

This has only happened while running Win2k and I have all those options turned
off both in the bios and in Win2k - no hibernate, no nothing. What was a strange
coincidence 2 weeks ago when this first happened (and my memory of exactly what
happened when and in what order cannot be trusted) was that I finally got the
Fedora Core installation DVD to run by passing it "linux acpi=off". All other
option tries had resulted in kernel panic and I hadn't been able to make the
installation.

However, this is all complicated by the fact that I was installing that FC4 on
the WD250 HD in question and had set things up so that it was the only HD
running so that I could let the installation install grub on the mbr. When I
reconnected all drives again, I figured that I hadn't gotten that power
connector back on right. That was the first of two times I've had this click
problem.

Odds are that either the PSU connector isn't sitting tight
on the power pins, the drive PCB is damaged somehow, or
isn't making a good (typically pressure, spring contacts)
connection to the contacts on the drive housing. A close
examination of these factors will tend to require pulling
the drive out.

 

[Robert Heiling]

I had a few other thoughts, but after going back through the thread, they
are rendered irrelevant. The fact that it has only happened after you've
been in the case does tend to make one think that the power connector is
relevant, or, as you've implied, the PS is using its intelligent abilities
for evil.

In reviewing the above, I see that I had talked about that connector being flaky
in the past, but hadn't mentioned why I thought that and it was that the drive
was not appearing in the bios. What would happen (probably after swapping in
another replacement DVD drive for the umpteenth time) is that the system failed
to boot and I would discover that neither drive on IDE0 was showing in the bios.
Reseating or wiggling that power connector would solve that problem until the
next time I went into the case.

However, since this only occurs with this drive, one could surmise that it
is indeed an issue with the power connection on the drive itself (perhaps
some internal loose connection in the drive that, when disturbed,
misbehaves).

It must be just that drive and if it were a simple case of power/no power as it
had been in the past, it was a nuisance, but livable. The new power supply
clicking situation added a new dimension.

If you have a spare drive, you could try cloning the problem
drive to the spare to see if the same problems result. At least that would
remove or confirm the possiblity of it being an HD issue (if it only happens
on the drive in question...).

That's probably already been covered by swapping around the PS leads and using
different ones for that same drive. The clicking problem hasn't been
intentionally reproducible and has only happened twice, so it would be hard to
know what worked and what didn't within any reasonable timeframe.

Barring that, perhaps WD would RMA it and send you a new one (maybe).

I just checked the record and it went out of warranty almost exactly a month

My apologies for being able to submit only wild
guesses, but you do seem to have an elusive issue here.

Don't feel bad, as making me think about it more gave me a possible solution as
I was sitting here writing. As I said, that connector issue was/is an
aggravation but I can live with it. It's also not the problem if my new theory
is correct!

That drive has been running in this box for 13 months, but that clicking has
only happened twice and that's within the past 3 weeks. Fedora Core installation
is the likely culprit! It leaves power management hardware in some funny state
that eventually gets cleared up somehow. I've had the FC4 DVD for months after I
gave up on it. Since FC5 is almost out, I burned the 1st CD of a betatest
version and booted it to see if that problem had been fixed. Still the same
problem with that kernel not wanting to boot with this motherboard.

To make a long story not quite as long, I discovered that acpi=off made the
installation go. That was 3 weeks ago when I had the first clicking epsisode.
Then just a few days ago, I reinstalled FC4 after some repartitioning. I haven't
kept a log because I didn't foresee a need for one, but I'd be pretty certain
that the 2nd and most recent clicking episode came after that FC4 installation.
It hasn't happened at all since Thursday morning when I started this thread.

Problem not solved nor analyzed in depth, but cause now known!

Thanks!
Bob

 

[Robert Heiling]

You need to pull the drive and check it.


Yes this is what seems to be happening based on the info
provided.


It is very unlikely that the PSU is doing anything
intelligent, you had assumed (??) that was the click but
typically it is the HDD when it shuts off, which it would do
when power is disconnected (for whatever reason, a fault or
system shutdown).

That's good to know. The system makes various sounds during startup and shutdown
and I haven't identified them all. I only knew that *something* made a click and
I sort of shudder at the thought of sounds like that coming from a HD. I guess
there would be 3 clicks from the 3 HD's, but they would all click together at
shutdown and blend into one? And then again, maybe it's just the Seagate since
the Western Digital that I have here externally makes no appreciable sound
either at powerup or powerdown. That's more than likely why I never suspected a
HD and figured it was the PS.

For the PSU to do something intelligent, it would have to
gather data. It has no feedback for this from a HDD, at
most it can monitor voltage and if the voltage were too bad,
it would simply shut off (and the whole system does too).

But the drives can be powered down by power management??

Odds are that either the PSU connector isn't sitting tight
on the power pins, the drive PCB is damaged somehow, or
isn't making a good (typically pressure, spring contacts)
connection to the contacts on the drive housing. A close
examination of these factors will tend to require pulling
the drive out.

It was a multi-problem situation which makes it harder to diagnose. That HD is
likely not to blame, but was suspected because of the record. I've posted my
diagnosis in another post that I had beeen composing when you posted this one.

Bob

 

[kony]

But the drives can be powered down by power management??

Yes that is left to the drive, the PSU is still supplying
the power (when all is working as intended).

 

[Robert Heiling]

Yes that is left to the drive, the PSU is still supplying
the power (when all is working as intended).

A ha! Interesting! So the PS really is as dumb as I had thought it was up until
this happened. Would it be correct to say that a drive would lose memory of
being in that suspended? (don't know the right term) state if all power was cut
off (power cord unplugged)? , but remain in that state if the power cord was
left in with the power led on the case staying on after a shutdown? Note: I've
never made any use or trials of power management as it's of no use to me.

And I'm still wondering where the clicks are coming from. Here are my internal
drives as posted in my original post. I have an identical 120 GB sitting here
attached to usb and an independent power supply. It makes no sound (within
reason) during power up/down. From that, I conclude that the 2 WD's in the box
don't either unless there is a special "click" when they go from running into
suspend mode. I heard multiple clicks from the case when the problem was
occurring.

IDE1-Sl ST340015A (40 GB, 5400 RPM, Ultra-ATA/100)
IDE0-Ma WDC WD1200JB-00GVA0 (111 GB, IDE)
IDE0-Sl WDC WD2500JB-00GVA0 (232 GB, IDE)

Bob

 

[kony]

A ha! Interesting! So the PS really is as dumb as I had thought it was up until
this happened.

Well it's a power supply, it's plenty smart enough at what
it's supposed to do, given the variability of the systems
and the construction cost.

Would it be correct to say that a drive would lose memory of
being in that suspended? (don't know the right term) state if all power was cut
off (power cord unplugged)?

Yes, if power is cut then returned a drive will turn on
fully.

, but remain in that state if the power cord was
left in with the power led on the case staying on after a shutdown? Note: I've
never made any use or trials of power management as it's of no use to me.

Power led on case means nothing towards the drive... The
issue is only of whether the power at the drive plug is kept
continuous or not. IOW, whether any LED stays on (on the
case or elsewhere) is a matter of how the motherboard was
designed, and what features are taken advantage of (if it
has the 5VSB powered LEDs, whether they're on the board
itself and/or a pin header for the case).

And I'm still wondering where the clicks are coming from. Here are my internal
drives as posted in my original post. I have an identical 120 GB sitting here
attached to usb and an independent power supply. It makes no sound (within
reason) during power up/down. From that, I conclude that the 2 WD's in the box
don't either unless there is a special "click" when they go from running into
suspend mode. I heard multiple clicks from the case when the problem was
occurring.

IDE1-Sl ST340015A (40 GB, 5400 RPM, Ultra-ATA/100)
IDE0-Ma WDC WD1200JB-00GVA0 (111 GB, IDE)
IDE0-Sl WDC WD2500JB-00GVA0 (232 GB, IDE)

Bob

Maybe it's not the drives, I can't hear the click but drives
can make such a sound. Take a piece of hose to your ear and
direct it around the system while trying to reproduce the
sound. I'm not saying it's impossible for a click to come
from a PSU, but rather one working properly will not change
a state to turn off a drive, it will supply the same power
at the drive plug continuously and the drive itself decides
what to do.

 

[CBFalconer]

.... snip ...

Maybe it's not the drives, I can't hear the click but drives
can make such a sound. Take a piece of hose to your ear and
direct it around the system while trying to reproduce the
sound. I'm not saying it's impossible for a click to come
from a PSU, but rather one working properly will not change
a state to turn off a drive, it will supply the same power
at the drive plug continuously and the drive itself decides
what to do.

I have mine set to spin down after 15 mins idle. This seems to
produce the odd click at spin down time, and the expected delay
(about 2 seconds) on access.

Since the timer control is in Windoze, I am not sure what is
determining the actual spin down. It may be the drive proper, and
the Windozery has just sent it suitable instructions, or it could
be in the hands of the driver. It would be nice to know which,
especially when co-ordinating with Linux systems on the same
machine.

--
"If you want to post a followup via groups.google.com, don't use
the broken "Reply" link at the bottom of the article. Click on
"show options" at the top of the article, then click on the
"Reply" at the bottom of the article headers." - Keith Thompson
More details at: <http://cfaj.freeshell.org/google/>
Also see <http://www.safalra.com/special/googlegroupsreply/>

 

[kony]

I have mine set to spin down after 15 mins idle. This seems to
produce the odd click at spin down time, and the expected delay
(about 2 seconds) on access.

Since the timer control is in Windoze, I am not sure what is
determining the actual spin down. It may be the drive proper, and
the Windozery has just sent it suitable instructions, or it could
be in the hands of the driver. It would be nice to know which,
especially when co-ordinating with Linux systems on the same
machine.

I believe the command is issued by the OS and the driver
must work properly to pass it to the drive.