CPU Core Voltage Too Low -> Crash?

[Bob Dietz]

NOT 0.1 - 0.4 volts, but 0.01 - 0.04 volts. And the DA converters that
MbMontior and AIDA32 are reading generally don't have that degree of accuracy.
Accuracy would something like 1.46V to 1.49V plus or minus 0.1V at best.

If you want some what accurate measurements, test with a digital multi-meter
at the ATX power connector. You cannot test the core voltage directly at that
connector, but the core voltage is derived from the +3.3 volt line. You'd need
special ($$$) equipment to accurately measure core voltage.



His Gateway motherboard is somewhat lacking in the bios configuration department.

Your MB uses RDRAM which is used in pairs.
Call the modules A1,A2 and B1,B2.
Call the slots they are in 1,2,3 and 4.
Swap A1,A2 from slots 1,2 to slots 3,4
and B1,B2 from slots 3,4 to slots 1,2.

While you're swapping the sticks (especially if you or someone in your
household is a smoker!), use SOFT (not for #1 pencil or ink) eraser to
clean the contacts.

Re-run Memtest86. If Memtest finds an error, replace the appropriate ram
module(s). If no errors are found, run the computer normally to see if
cleaning the contacts has fixed the intermitant spontaneous reboots.


Bob

 

[Don Cohen]

Hi Doug,

Thanks for the reply and advice.

I presume that you feel the voltage measurements are sufficiently accurate
with "Motherboard Monitor" and/or AIDA32 to conclude that the VCore is
genuinely lower than it should be?

And if so, you also feel this small but significant difference could be
creating the system instability?

I will go ahead and give your plan a try tomorrow, and report back.

Thanks much!

Best regards,

Don

Don,

The voltages for the various components (RAM, CPU and etc.) are controlled
by voltage regulators on the motherboard. Granted, it could be a different
component that is having issues, and causing a voltage drop across the
entire system. It could also simply be the voltage regulator that supplies
the CPU Core Voltage failing due to age.

Try removing all add-on cards, and anything but your primary hard disk (you
will need to leave the video card in) and retest the system. If the voltage
is normal, then add the hardware back one piece at a time until you see the
voltage drop. If you still get the voltage difference, even with all extra
hardware removed, then you're most likely looking at a motherboard
replacement.

--
Doug Knox, MS-MVP Windows XP/ Windows Smart Display
Win 95/98/Me/XP Tweaks and Fixes
http://www.dougknox.com

 

[Don Cohen]

Hi Bob,

NOT 0.1 - 0.4 volts, but 0.01 - 0.04 volts. And the DA converters that
MbMontior and AIDA32 are reading generally don't have that degree of accuracy.
Accuracy would something like 1.46V to 1.49V plus or minus 0.1V at best.

If you want some what accurate measurements, test with a digital multi-meter
at the ATX power connector. You cannot test the core voltage directly at that
connector, but the core voltage is derived from the +3.3 volt line. You'd need
special ($$$) equipment to accurately measure core voltage.

A bit beyond my capabilities at the moment ;-)

His Gateway motherboard is somewhat lacking in the bios configuration

department.

So I've found...

Your MB uses RDRAM which is used in pairs.
Call the modules A1,A2 and B1,B2.
Call the slots they are in 1,2,3 and 4.
Swap A1,A2 from slots 1,2 to slots 3,4
and B1,B2 from slots 3,4 to slots 1,2.

While you're swapping the sticks (especially if you or someone in your
household is a smoker!), use SOFT (not for #1 pencil or ink) eraser to
clean the contacts.

Re-run Memtest86. If Memtest finds an error, replace the appropriate ram
module(s). If no errors are found, run the computer normally to see if
cleaning the contacts has fixed the intermitant spontaneous reboots.

Will do.

Thanks!

Don

 

[Don Cohen]

Sorry to but in on this fishing expedition but it seems to me the only thing
common to all your events is the CD Rom you were using with the Windows
disk. Try unplugging it and see if problems continue. A long shot I know but
cheap to try.

No problem - join the party!

I'll add this to the list of things to try tomorrow.

Hopefully I'll come up with the answer sooner or later!

Don

 

[Doug Knox MS-MVP]

Its possible, Don. Its also possible that the monitoring circuit is what's flaky, but that shouldn't cause a reboot, unless your BIOS would support turning off the computer to prevent damage, but then it should shut down. When you're dealing with voltages that low, there's not a lot of room for error.

 

[Don Cohen]

Thanks, Doug.

I'll report back tomorrow on what happens with the experiment you suggest.

I went ahead and switched the RAM modules as one of the other posters
suggested, and will run memtest86 again overnight.

Somehow, I have to fix this darn thing!

All the best,

Don

Its possible, Don. Its also possible that the monitoring circuit is what's
flaky, but that shouldn't cause a reboot, unless your BIOS would support
turning off the computer to prevent damage, but then it should shut down.
When you're dealing with voltages that low, there's not a lot of room for
error.

 

[Jym]

Your voltages seem to be right where they are suppose to be with the
processor that you have ,and well within the working variance of the
specifications. Researching your stop error seem to point toward system
ram , taking into account the problems that you have posted . Although this
common stop error could be masked by underlying causes. Try one stick at a
time , then replacing more with successful operation. I too believe the best
way to troubleshoot is to pull out all the nonessential hardware and
software and replace one at a time. Good Luck . Jym

 

[Don Cohen]

Hi Jym,

Your voltages seem to be right where they are suppose to be with the
processor that you have ,and well within the working variance of the
specifications. Researching your stop error seem to point toward system
ram , taking into account the problems that you have posted . Although this
common stop error could be masked by underlying causes. Try one stick at a
time , then replacing more with successful operation. I too believe the best
way to troubleshoot is to pull out all the nonessential hardware and
software and replace one at a time. Good Luck . Jym

Thanks. I guess this is what I'll be doing for the next few days.....

Best regards,

Don

 

[Ben Myers]

More or less complicated than diagnosis of a human patient? I'll vote for more
complicated because of the infinite complexity of Windows... Ben Myers

 

[Don Cohen]

Hi Doug (and all),

I did the swap of the memory modules, ran memtest86, still no errors.

But then, I rebooted normally, and turned on "Motherboard Monitor". I set
it to always 'stay on top' so I could see the CPU Core Voltage at all times.
When I first turned it on, it registered 1.49V (according to Intel, the spec
for my P4 chip is 1.50V).

I then opened up a number of different applications (Photoshop, ThubsPlus,
Internet Explorer windows, etc.) and watched the Voltage drop down to 1.46V,
1.48V, back up to 1.49V, etc. Back and forth - not quickly, just up and
down. Never went higher than 1.49V.

I opened up a program that shows digital camera pictures, and cycled through
these very quickly. I watched the CPU voltage again drop to 1.46V, up a
little, down again to 1.46V, and then bam!! - the crash occurred, same as
all the other times.

My thinking is that the CPU Core Voltage dropped lower than 1.46V and
triggered the crash. As per Doug's comments, the Voltage Regulator for this
is on the motherboard - hence the conclusion is that the motherboard is the
source of my crashes.

Please critique my thinking here, and confirm or refute what I'm inclined to
conclude.

Thanks.

Don


Its possible, Don. Its also possible that the monitoring circuit is what's
flaky, but that shouldn't cause a reboot, unless your BIOS would support
turning off the computer to prevent damage, but then it should shut down.
When you're dealing with voltages that low, there's not a lot of room for
error.

--
Doug Knox, MS-MVP Windows XP/ Windows Smart Display
Win 95/98/Me/XP Tweaks and Fixes
http://www.dougknox.com

 

[Don Cohen]

And finally:

I removed all but the VideoCard, disconnected my secondary hard drive, and
started it up again.

Motherboard Monitor showed CPU Core Voltage at 1.49V, unchanged from before.
I opened up the same program to quickly cycle through a series of photo
images, and again watched the VCore fluctuate from 1.48V to 1.46V to 1.49V.
When at 1.46V for a short while, the crash again occurred.

All in all, looks like it's time for a new motherboard.

But again, any critical thinking that suggests otherwise is still welcome.

Don

 

[Bob Dietz]

Hi Doug (and all),

I did the swap of the memory modules, ran memtest86, still no errors.

But then, I rebooted normally, and turned on "Motherboard Monitor". I set
it to always 'stay on top' so I could see the CPU Core Voltage at all times.
When I first turned it on, it registered 1.49V (according to Intel, the spec
for my P4 chip is 1.50V).

I then opened up a number of different applications (Photoshop, ThubsPlus,
Internet Explorer windows, etc.) and watched the Voltage drop down to 1.46V,
1.48V, back up to 1.49V, etc. Back and forth - not quickly, just up and
down. Never went higher than 1.49V.

I opened up a program that shows digital camera pictures, and cycled through
these very quickly. I watched the CPU voltage again drop to 1.46V, up a
little, down again to 1.46V, and then bam!! - the crash occurred, same as
all the other times.

My thinking is that the CPU Core Voltage dropped lower than 1.46V and
triggered the crash. As per Doug's comments, the Voltage Regulator for this
is on the motherboard - hence the conclusion is that the motherboard is the
source of my crashes.

Please critique my thinking here, and confirm or refute what I'm inclined to
conclude.

Thanks.

Don


Its possible, Don. Its also possible that the monitoring circuit is what's
flaky, but that shouldn't cause a reboot, unless your BIOS would support
turning off the computer to prevent damage, but then it should shut down.
When you're dealing with voltages that low, there's not a lot of room for
error.

Don, if a patient walked into your office and said, "My blood pressure is
130/85." Would you take him at his word or would you take your own reading?
If the patient brought in his own "blood pressure test kit", would you use
that to take the measurement or would you use your own equipment.

At this very moment, software readings on the following system:
Motherboard: Asus A7V Rev3
CPU: 1.3 Ghz (Morgan core)
RAM: 512MB ( 2 x 256Mb SD-133)
Harddisk: Western Digital 80Gb @ 7200RPM with 8Mb cache.
Power supply: Antec Smart Power SL300S
are:
+12.0V Rail: 12.46V
+ 5.0V Rail: 4.96V
+ 3.3V Rail: 3.57V
CPU Core: 1.84V

At the same time, actual readings taken with a $60 Radio Shack 22-163 Digital
Multi-meter:
+12.0V Rail: 12.22V
+ 5.0V Rail: 5.03V
+ 3.3V Rail: 3.37V
CPU Core: (no way to measure)

This particular motherboard ALWAYS gives a low reading for the +12V and +5V
rails. It ALWAYS gives a high reading for the +3.3V rail. The core voltage
ALWAYS reads higher than what it has been set to in the BIOS. That has held
true in the case of three different power supplies.

The software readings off your motherboard may possibly be extreamly accurate.
But I wouldn't bet the team on it.

Personally, my next step would include all of the following.
Unplug power and ribbon cables from every CD/DVD readers and/or writer.
Unplug the printer.
Remove any sound cards.
If there's a modem and you're not using it (you've got broadband), pull it.
Pull any other hardware you don't absolutely need.
Remove one pair of RD-RAM sticks.
Open Device Manager and disable any hardware you've temporarily removed.
While you're in Device Manager disable any non-essential hardware on the
motherboard. Eg. Built-in sound. Network adapter if you're using dialup.

If the system runs stable with minimal hardware, start adding things back one
item per day. If you get a random reboot, pull the last item and run the
system unchanged for a few days. If random reboots seem to disappear, assume
the item you removed is bad and continue adding items one per day. If the
random reboots crop-up again with these other items, start thinking
motherboard or power supply.

In that case, try to get the store you bought the power supply from to
exchange it for another. (Antec TruPower has a great reputation, but maybe
this one was produced on Monday morning after a wild weekend.)

Good luck,
Bob

 

[Jym]

It looks like you have done a good job at troubleshooting the problem . I
would make a call to Gateway and see if they will send you a new motherboard
,( being realistic I know it is a long shot , but it can't hurt).
Jym

 

[Jym]

Bob ,
This might be unrelated to this thread . You have noted that your voltage
software has your core voltage at 1.84V . I too have a 1.3 Ghz (Thunderbird)
and my MSI software has the core running at 1.76V . Is your CPU overclocked
? Is the Morgan core a Thunderbird ? Jym

 

[Don Cohen]

Hi Bob,

Don, if a patient walked into your office and said, "My blood pressure is
130/85." Would you take him at his word or would you take your own reading?
If the patient brought in his own "blood pressure test kit", would you use
that to take the measurement or would you use your own equipment.

Point well taken.

Personally, my next step would include all of the following.
Unplug power and ribbon cables from every CD/DVD readers and/or writer.
Unplug the printer.
Remove any sound cards.
If there's a modem and you're not using it (you've got broadband), pull it.
Pull any other hardware you don't absolutely need.
Remove one pair of RD-RAM sticks.
Open Device Manager and disable any hardware you've temporarily removed.
While you're in Device Manager disable any non-essential hardware on the
motherboard. Eg. Built-in sound. Network adapter if you're using dialup.

If the system runs stable with minimal hardware, start adding things back one
item per day. If you get a random reboot, pull the last item and run the
system unchanged for a few days. If random reboots seem to disappear, assume
the item you removed is bad and continue adding items one per day. If the
random reboots crop-up again with these other items, start thinking
motherboard or power supply.

Well, I've done part one of this: removed modem, sound and firewire PCI
cards, leaving the Videocard as the only card left. I disconnected the
secondary internal hard drive, and rebooted. VCore readings were the same:
fluctuating 1.49V, 1.48V, 1.46V, and while quickly looking through a series
of images, the same crash occurred. That eliminates these hardware devices.

I guess I can do the same thing, disconnecting both CDRW and DVD-RW drives
(would unplugging be sufficient, or does one need to disconnect the data
cables as well)?

In that case, try to get the store you bought the power supply from to
exchange it for another. (Antec TruPower has a great reputation, but maybe
this one was produced on Monday morning after a wild weekend.)

Always a possibility, but sometimes if it looks like a skunk, smells like a
skunk and acts like a skunk, chances are it must be a skunk. I don't know
for sure, but that's the way I'm starting to feel with regard to the
motherboard.

I'll be in touch with Gateway tomorrow to see about replacing the
motherboard, with the option to return if I turn out to be wrong.

Thanks for all your help and advice.

Don

 

[Don Cohen]

It looks like you have done a good job at troubleshooting the problem . I
would make a call to Gateway and see if they will send you a new motherboard
,( being realistic I know it is a long shot , but it can't hurt).

That's on my to do list for tomorrow. I've been in touch with the Service
Manager for the local Gateway Country Stores, and she has been sympathetic
and good to work with. I'll see what Gateway will be able to do here.

Don

 

[Bob Dietz]

Bob ,
This might be unrelated to this thread . You have noted that your voltage
software has your core voltage at 1.84V . I too have a 1.3 Ghz (Thunderbird)
and my MSI software has the core running at 1.76V . Is your CPU overclocked
? Is the Morgan core a Thunderbird ? Jym

Opps. I meant to write Thunderbird. Morgan core was the Duron that the current
CPU replaced. (Fried when the CPU fan died without any warning.)

Those readings cited below are from AIDA32. I get similar numbers from
PC-Probe (supplied with MB.) The setting in the bios is 1.78V (auto).
I think the offical AMD specification for the CPU is 1.75V.

Some while after I purchase that motherboard, Tom's Hardware did a comparison
of 10 or so different Athlon Motherboards. They said ASUS fudged some of the
default settings to get slightly better performance numbers. In that sense it
may be overclocked by some miniscule amount, but I haven't done anything to
make it so.

On the whole I trust the bios setting more than I trusted the software readings.

Bob

 

[Bob Dietz]

Hi Bob,

Well, I've done part one of this: removed modem, sound and firewire PCI
cards, leaving the Videocard as the only card left. I disconnected the
secondary internal hard drive, and rebooted. VCore readings were the same:
fluctuating 1.49V, 1.48V, 1.46V, and while quickly looking through a series
of images, the same crash occurred. That eliminates these hardware devices.

I guess I can do the same thing, disconnecting both CDRW and DVD-RW drives
(would unplugging be sufficient, or does one need to disconnect the data
cables as well)?

Just the power cable ought to be enough, so far as I know.
But I always unplug both.

Always a possibility, but sometimes if it looks like a skunk, smells like a
skunk and acts like a skunk, chances are it must be a skunk. I don't know
for sure, but that's the way I'm starting to feel with regard to the
motherboard.

I agree the motherboard/cpu/memory smells like a skunk. The power supply is
pretty unlikely, but it's just four screws and unplug a few connectors to try
a different one. (That assumes you've bought it at a local store that will
give you an exchange for the current one.)

I'll be in touch with Gateway tomorrow to see about replacing the
motherboard, with the option to return if I turn out to be wrong.

You've swapped the memory around, but have you - removed half - tested -
swapped the other half in and tested?

When the core voltage dips, what is happening to the voltage on
the +5V and +3.3 volt rails? If those are holding steady, it would pretty well
convince me the motherboard was the only reasonable possibility.

Thanks for all your help and advice.

Don

You're welcome.

Bob

 

[Jym]

Your problem is interesting , ( more than just interesting for you of
course) .
While I monitored my voltages I opened up every program that I could think
of , and no voltage drop in the core or any other voltages. Which leave me
to believe that MB voltages should not fluctuate with the opening of
programs. Jym

 

[Don Cohen]

Bob,

I agree the motherboard/cpu/memory smells like a skunk. The power supply is
pretty unlikely, but it's just four screws and unplug a few connectors to try
a different one. (That assumes you've bought it at a local store that will
give you an exchange for the current one.)

I got it from Best Buy; their return policy is 15 days, but as you suggest
they might be willing to swap one. I'll keep this in mind as a possibility.

When the core voltage dips, what is happening to the voltage on
the +5V and +3.3 volt rails? If those are holding steady, it would pretty well
convince me the motherboard was the only reasonable possibility.

Actually I was watching both the +2.5V and the +3.3V readings. The +2.5V
read as 2.46V and didn't budge at all. The +3.3V read at 3.29V virtually
all the time, with a very rare and brief drop to 3.27V. I just checked
again and the +5V stayed at 5.02V and didn't budge.

I guess your thinking is that if it were some other component it would drop
voltage across the board, whereas if it's only one voltage reading that's
varying, that has to be originating in the motherboard itself.

I hadn't considered that, but it does make sense.

Let me ask one follow-up question:

Assume for the moment that this is in fact a motherboard defect (presumably
in voltage regulation) that is causing these crashes. Consider this:

Starting from when I first bought this system 2 years ago, I (and a number
of others) have experienced random and episodic 'blackout' episodes. Here,
the LCD (DVI connection) would go completely black, and within 2 or 3
seconds at the most, would come back on. Nothing else occurred. No error
messages, no applications were effected, nothing. These have recurred
sporadically - once a week, once a month, once every few months. Always the
same. Trouble-shooting by me and others on the web never identified a
clear-cut cause. I just lived with it, since it was a minor annoyance and
seemingly harmless.

When my first reboot/crash occurred last month, it started the same way -
the LCD went completely black, but instead of just coming back on as it
usually did, the system just rebooted. It's as if whatever was causing
those 'blackouts' now worsened, and caused a reboot.

What do you (or others) think of this theory? I know it's essentially
impossible to prove, but it seems might suspicious to me.

Thanks again. I'll be in touch with the local Gateway Service Manager
tomorrow.

Don