A7N8X Deluxe - booting hard drive spinup CHAOS! :(

[Unk]

When I power up my computer on a cold start I go through this sequence:

- push power button
splash screen and stops
have to tab to get to post screen
hear hard drive spin up
at post screen computer freezes
hard drive light may be on, or off continuous
- reset
goes to splash and post screen and stops
hit delete for bios
hard drive spins up
computer freezes
- reset
goes through splash and post
dots for XP start moving
hear hard drive spin up
computer freezes,
- reset
splash,post,XP screen giving choice of boots
hit return
gets a long way into XP loading screen
hard drive spins up
computer freezes
- reset
same, let countdown choice timer run out
computer boots normally!

2800 Athlon, Thermalright sink running at 106F norm temp
Corsair 3200 1 gig ram (swapped mods back and forth)
WD 80 7200( with Windows) and 40 gig drives both less than two years old
Did fresh install of XP Pro last week
Antec 420 power supply 6 months old
ATI 9800 pro and modem cards
NEC DVD burner, Sony CD burner
Updated to newest bios last week
Bios settings all default, look okay (speed correct)
Swapped power cables from cd's on ide 1 to ide 0 hard drives
Reseated everything in computer, memory, cpu, wiring.....

Once the computer is running it will be fine for days, so I leave it on.
If it has to reset it reboots normally.

But, every so often (every couple of days maybe) the hard drive will just
'reboot', an instant cutoff followed by power spin up. It might make XP
freeze, or might not. It started this about a month or so ago, no
components added at that time, though I added a new Sony CD burner about
three months ago.

I'll be glad to start replacing things but wanted to run this by the
group to see which component should be the first in line.

Thanks for any input!!!!

Unk

 

[Rob Hemmings]

When I power up my computer on a cold start I go through this sequence:

- push power button
splash screen and stops
have to tab to get to post screen
hear hard drive spin up
at post screen computer freezes
hard drive light may be on, or off continuous
- reset
goes to splash and post screen and stops
hit delete for bios
hard drive spins up
computer freezes
- reset
goes through splash and post
dots for XP start moving
hear hard drive spin up
computer freezes,
- reset
splash,post,XP screen giving choice of boots
hit return
gets a long way into XP loading screen
hard drive spins up
computer freezes
- reset
same, let countdown choice timer run out
computer boots normally!

2800 Athlon, Thermalright sink running at 106F norm temp
Corsair 3200 1 gig ram (swapped mods back and forth)
WD 80 7200( with Windows) and 40 gig drives both less than two years old
Did fresh install of XP Pro last week
Antec 420 power supply 6 months old
ATI 9800 pro and modem cards
NEC DVD burner, Sony CD burner
Updated to newest bios last week
Bios settings all default, look okay (speed correct)
Swapped power cables from cd's on ide 1 to ide 0 hard drives
Reseated everything in computer, memory, cpu, wiring.....

Once the computer is running it will be fine for days, so I leave it on.
If it has to reset it reboots normally.

But, every so often (every couple of days maybe) the hard drive will just
'reboot', an instant cutoff followed by power spin up. It might make XP
freeze, or might not. It started this about a month or so ago, no
components added at that time, though I added a new Sony CD burner about
three months ago.

I'll be glad to start replacing things but wanted to run this by the
group to see which component should be the first in line.

Thanks for any input!!!!

Sounds typical of a failing hard drive. Try WD's DLG diagnostic
to see if it flags any errors on your HDs
http://support.wdc.com/download/index.asp#diagutils
HTH

 

[Unk]

Thanks for any input!!!!

Sounds typical of a failing hard drive. Try WD's DLG diagnostic
to see if it flags any errors on your HDs
http://support.wdc.com/download/index.asp#diagutils
HTH

I puttered a little more, changing cables around etc. Regardless of
whether it's the boot drive/slave/master or not it shuts down/starts up a
couple of times on boot. Went to WD site and it's under warranty.
Replacement being shipped, hopefully that solves it.

Thanks again!

Unk

 

[Chris Hill]

When I power up my computer on a cold start I go through this sequence:

I'll be glad to start replacing things but wanted to run this by the
group to see which component should be the first in line.

Thanks for any input!!!!

What is your 5v showing? If any less than about 4.9, you need a power
supply.>

 

[Unk]

What is your 5v showing? If any less than about 4.9, you need a power
supply.>

Honking torpedoes, it's 4.52. Well, I've tried Antec, Enermax, both
failed on me, any other 'good' brands?

Thank you !!!!

Unk

 

[jroc]

Honking torpedoes, it's 4.52. Well, I've tried Antec, Enermax, both
failed on me, any other 'good' brands?

Thank you !!!!

Unk

"De Oppresso Liber"
From Oppression Liberate

jroc® 7th Special Forces Ft.Bragg N.C 18/Bravo
****only thing easy in life is failure****
*******so get tough*******

 

[Chris Hill]

Honking torpedoes, it's 4.52. Well, I've tried Antec, Enermax, both
failed on me, any other 'good' brands?

Thank you !!!!

I've had good luck with enermax, but you need to check the size, not
only the brand. You may have just received a couple of bad samples,
as well. I'd grab something in the 400w range and of a decent brand.
I have a Dynapower in this machine, kind of cheap but better than the
original by .2v.

 

[Paul]

Honking torpedoes, it's 4.52. Well, I've tried Antec, Enermax, both
failed on me, any other 'good' brands?

Thank you !!!!

Unk

Some power measurements for you:

A7N8X-E,11x200 processor,3x512MB DIMMs==>[email protected] amps, [email protected], [email protected]
Add ATI9800pro [email protected] amps, [email protected]
Disk drive or CD drive, for each 5V@1 amps, [email protected]

(Note: These measurements were artificially contrived. The 16.6 amp
number was using a non-graphics CPU-intensive test. The 5.5 amp number
is measured on the graphics card Aux power connector, while doing
3DMark. The two supply currents generally cannot peak at the same
time, because as soon as the CPU talks to the video card, it stalls
more often, and the CPU current will be less than the max above.)

So, for an AthlonXP system with high end video card, the supply
needs a rough minimum of 25 amps on the +5V rail.

When the power reads low, like it is on your system, there can be
two reasons:

1) Power supply is overloaded.
2) ATX 20 pin connector is bad.

If the motherboard power connector is damaged, then changing
power supplies may not correct the low reading. In any case, you
should use a multimeter, to measure the voltage on the board, to
see whether it is indeed as low as the Power Monitor is
indicating.

Make sure the ATX 20 pin connector is fully seated, and the latch
engaged, to prevent damage to the pins by overheating. If the
contact resistance is low, the pins won't heat up. If a pin is
making poor contact, it could get hot.

There could also be some kind of fault in the motherboard itself,
which is drawing more current than normal. You say you've had
two power supplies "fail" on you, so consider the motherboard
or something on the motherboard, as a possible reason for your
problems as well.

HTH,
Paul

 

[Unk]

(e-mail address removed) (Paul) wrote in

There could also be some kind of fault in the motherboard itself,
which is drawing more current than normal. You say you've had
two power supplies "fail" on you, so consider the motherboard
or something on the motherboard, as a possible reason for your
problems as well.

HTH,
Paul

Thanks Paul, lots of stuff to digest there. No word from Antec support,
so I might try buying a quick throwdown supply from Compusa etc. to see
if that immediately alleviates the low voltage to confirm that at least
the final problem is the ps. But then I'll need to look/watch for
something else with mb possibly.

I wonder if this also is somehow related to my memory 'problem'. 2 x 512
Corsair platinum 3200's that sometimes fail memcheck, sometimes don't.
Happened through both power supplies, but only this Asus board. I guess
I'll start a little fantasizing about a nice 939 pin 64 bitter board in
the near future. My patience on these kinds of problems (elusive ones)and
money throwing grows thin quickly.

Thanks again for all the good information!

Unk

 

[FG]

Can't see why a new ANTEC power supply would
give you low readings, an even less why it would
not work at least as well as « a quick throwdown supply ».
Antec is one of the very good PSUs

 

[Paul]

(e-mail address removed) (Paul) wrote in


Thanks Paul, lots of stuff to digest there. No word from Antec support,
so I might try buying a quick throwdown supply from Compusa etc. to see
if that immediately alleviates the low voltage to confirm that at least
the final problem is the ps. But then I'll need to look/watch for
something else with mb possibly.

I wonder if this also is somehow related to my memory 'problem'. 2 x 512
Corsair platinum 3200's that sometimes fail memcheck, sometimes don't.
Happened through both power supplies, but only this Asus board. I guess
I'll start a little fantasizing about a nice 939 pin 64 bitter board in
the near future. My patience on these kinds of problems (elusive ones)and
money throwing grows thin quickly.

Thanks again for all the good information!

Unk

My experience is only a "sample size of one", but I was
getting "CPU failed overclock" as long as I used what turned
out to be bad RAM. The RAM passed memtest86 (from memtest.org),
but then I decided to run Prime95 (mersenne.org), and it could
only run for 30 minutes before it got an error. The error was
always in the same block of memory, so I knew it had to be the
DIMMs and not the motherboard. The computer was only stable at
180-185MHz when using the bad RAM.

I used a Knoppix boot disk, and ran multiple copies of Prime95,
each one using a custom memory size. In this way, I knew the
area in memory each would be testing, and the same copy of Prime
failed each time. You can probably find other read-only CD based
Linux systems, to do the same kind of test.

By using a CD based test environment, I don't need to have a
Windows install on a HD, to test the new motherboard. Now that
the processor and memory testing are finished, via the CD, I
can run Windows from the HD without having to worry.

You should set up the memory timing parameters manually, as
there have been a few DIMMs made where the SPD programming
was incorrect for the memory. See if you can dig up some
original advert info, with the numbers for the RAM, then
enter those manually in the BIOS.

http://corsairmicro.com/corsair/xms.html

As for power supplies, it is really hard to recommend a brand,
as every company that makes them is tempted to cut corners.

In your case, when two supplies fail, I would be suspicious
of the cause.

When buying them, the best anyone can do, is ask to read
the label on the side of the supply, to see how many amps
it has for each voltage. Some high power supplies are not
generous on every output, so those supplies will cook more
often than other supplies.

If you want less trouble, buy a pcpowerandcooling supply.
Costs big bucks, but would be worth it if you are tired of
problems with supplies. Antec and Enermax are OK, but not
completely immune to problems.

Once all of that is tested (new PSU, CPU, memory, CDROM),
you can go back to testing the hard drive, to make sure
the original or its replacement are OK. As Rob mentioned,
the manufacturer's diagnostic is something you can download
for free.

I make my current measurements with a $400CDN DC clamp-on
ammeter, which is something that not too many sane people
would bother buying It comes in handy for figuring out
whether there is an overload or not, and you don't have to
cut or touch any wires to use it. There are jaws that close
around the wire and the instrument senses the magnetic field
around the wire caused by the current flow. I can even
measure the AC current coming from the wall, but I need a
"cheater" cord, that breaks out the individual wires in
the cord, as if you clamp around the two AC wires at the
same time, the magnetic fields cancel out.

I've even used it for debugging my central air conditioning.

HTH,
Paul

 

[Unk]

(e-mail address removed) (Paul) wrote in

My experience is only a "sample size of one", but I was
getting "CPU failed overclock" as long as I used what turned
out to be bad RAM. The RAM passed memtest86 (from memtest.org),
but then I decided to run Prime95 (mersenne.org),

I used a Knoppix boot disk, and ran multiple copies of Prime95,
each one using a custom memory size.

Downloaded it and have piddled with Knoppix a bit last month, but I'm not
a Linux whiz and have to figure out some of the parameter stuff and
setting things up if needed. Will try Windows prime first.

Corsair Ramguy reply says it's not the ram if it's giving intermittent
errors, so either he's right and something else will fix things, or I'll
be coughing up more money for ram also. Argh.

You should set up the memory timing parameters manually, as
there have been a few DIMMs made where the SPD programming
was incorrect for the memory. See if you can dig up some
original advert info, with the numbers for the RAM, then
enter those manually in the BIOS.
http://corsairmicro.com/corsair/xms.html

I'll try relaxed timings first for it, 73332.5, I haven't done that yet,
all on 'auto'.

As for power supplies, it is really hard to recommend a brand,
as every company that makes them is tempted to cut corners.
In your case, when two supplies fail, I would be suspicious
of the cause.

My luck there's a loose screw I dropped while assembling something
shorting things out.

If you want less trouble, buy a pcpowerandcooling supply.
Costs big bucks, but would be worth it if you are tired of
problems with supplies. Antec and Enermax are OK, but not
completely immune to problems.

As a matter of fact I did that about 5 years ago or so, the 'little' 250
watt supply is long gone, but my computer is still in the original PCP&C
case! A bit modified as the front panel had virtually no cooling vent
holes to speak of so I did some judicious drilling.

Once all of that is tested (new PSU, CPU, memory, CDROM),
you can go back to testing the hard drive, to make sure
the original or its replacement are OK. As Rob mentioned,
the manufacturer's diagnostic is something you can download
for free.

The replacement is already shipped from WD, who I guess I have to
recommend highly as even if the drives go bad, the customer service is
extremely good.

I've even used it for debugging my central air conditioning.
HTH,
Paul

This will keep me busy for the next couple of days, I'm going to run the
memory test after I reseat the main ps plug. One question, does the 5v
for the drives run through the ribbon, or the power plug? I have a volt
meter so I can measure the plug easily.

Thanks so much again!

John

 

[Unk]

This will keep me busy for the next couple of days, I'm going to run
the memory test after I reseat the main ps plug. One question, does
the 5v for the drives run through the ribbon, or the power plug? I
have a volt meter so I can measure the plug easily.

Thanks so much again!

John

Update, my 12+v this morning is now 11.67v!

Sigh.

Unk John

 

[Paul]

This will keep me busy for the next couple of days, I'm going to run the
memory test after I reseat the main ps plug. One question, does the 5v
for the drives run through the ribbon, or the power plug? I have a volt
meter so I can measure the plug easily.

Thanks so much again!

John

On an AthlonXP board, the +5V is typically used to power the
Vcore circuit. In this case, the circuit is two phase. Each
phase has a toroid, and the toroid "donut" is on the output
side of the switching circuit. There is usually a vertical
cylinder shaped inductor (coil of wire), and that is an
input side inductor.

The circuit would look like this:

+--- Measure here, on the lead of the wire cylinder
| on the side that is away from the several
| several capacitors. This is the real voltage
| the motherboard is using.
v (Vcore)
+5V --- Wire coil --Several-----MOSFETS |
Cylinder Capacitors (two per --Donut coil--More
| | | phase) of wire caps
GND GND GND (one per |
phase) GND

To see where to measure, flip to page 17 of the motherboard
manual and look at a picture of the motherboard. The
cylindrical wire coil is next to the two white three pin
fan headers. You want to probe the side closest to the AGP
connector.

http://www.asus.com/pub/ASUS/mb/socka/nforce2/a7n8x/e1170_a7n8x.pdf

If you grab an unused disk drive power connector, and probe the
+5V on that wire, that reads out the "unloaded" value of +5V.
The voltage there is equal to the +5V voltage as delivered
raw from inside the power supply.

When you measure the +5V on the motherboard itself, while the
motherboard is running, and there is some computing load,
you are observing the effects of voltage loss in the ATX20
pin connector and wires. I.e. The difference between the voltage
seen on an unused drive connector, and the voltage seen at
inductor L22, is a measure of how much voltage is
being dropped across the ATX20 pin connector and the
wiring harness of the PSU.

My experience is, the disk drive connector will read pretty close
to the correct value.

Here is an example of what happens when too much power is lost
inside the ATX20 pin connector. This particular motherboard was
poorly designed, so this fault was a common one on this Tyan
board:

http://students.washington.edu/vladis/atx_burn/

HTH,
Paul

 

[Unk]

(e-mail address removed) (Paul) wrote in

Here is an example of what happens when too much power is lost
inside the ATX20 pin connector. This particular motherboard was
poorly designed, so this fault was a common one on this Tyan
board:

http://students.washington.edu/vladis/atx_burn/

HTH,
Paul

Good link, when I checked the connector no sign of poor connection, etc.

Got in the new PCP&C 470 watt silencer supply and installed. Computer
boots normally, so it was indeed the power supply. But the new one
reports 4.89v on the 5v line, hopefully I didn't get a bum one of these
either, but so far so good.

Thanks for all the information and assistance!!

John

 

[Chris Hill]

(e-mail address removed) (Paul) wrote in


Good link, when I checked the connector no sign of poor connection, etc.

Got in the new PCP&C 470 watt silencer supply and installed. Computer
boots normally, so it was indeed the power supply. But the new one
reports 4.89v on the 5v line, hopefully I didn't get a bum one of these
either, but so far so good.

That may be good enough. I'd check a drive connector with a vom just
to get a second opinion.