CPU Voltage sometimes dips to zero for a very short period

  • Thread starter Thread starter JD
  • Start date Start date
J

JD

Hi Experts,

I have a box where the CPU voltage sometimes dips to zero for a few secs
and then returns to the normal of 1.5v.

Has anyone seen this situation before? I am thinking of removing the CPU
and the local electrical connections around it and re-installing them.
The board is Intel and I have the Intel monitor for temperature and
voltage running.

Suggestions appreciated.

JD
 
JD said:
Hi Experts,

I have a box where the CPU voltage sometimes dips to zero for a few secs
and then returns to the normal of 1.5v.

Has anyone seen this situation before? I am thinking of removing the CPU
and the local electrical connections around it and re-installing them.
The board is Intel and I have the Intel monitor for temperature and
voltage running.

Suggestions appreciated.

JD
Click to expand...

If for whatever reason, your power supply was faulty and the voltage
dipped to zero...reinstalling the CPU would hardly solve the problem...

however I suspect there is a problem with your monitoring...
for if the CPU voltage actually dropped to zero your machine would of
course cease to operate!
 
JD said:
Hi Experts,

I have a box where the CPU voltage sometimes dips to zero for a few secs
and then returns to the normal of 1.5v.

Has anyone seen this situation before? I am thinking of removing the CPU
and the local electrical connections around it and re-installing them.
The board is Intel and I have the Intel monitor for temperature and
voltage running.

Suggestions appreciated.

JD
Click to expand...

You don't actually say if the computer continues to work during these "zero
voltage" periods. There is no guarantee that the monitoring software is
telling you the truth, especially if something else on the computer is
interfering with it or preventing it from updating for a short period. The
fact that you "see" zero volts displayed from software running on the
computer certainly suggests that the computer does indeed continue
operating, something it could never do with no power going into it. At the
very least there would be an immediate crash (and probably a reboot) if the
power was dropped for even an instant.
 
John said:
You don't actually say if the computer continues to work during these
"zero voltage" periods. There is no guarantee that the monitoring
software is telling you the truth, especially if something else on the
computer is interfering with it or preventing it from updating for a
short period. The fact that you "see" zero volts displayed from software
running on the computer certainly suggests that the computer does indeed
continue operating, something it could never do with no power going into
it. At the very least there would be an immediate crash (and probably a
reboot) if the power was dropped for even an instant.
Click to expand...

Thanks Philo and John.

I was too concerned to check if the computer was working. All other
voltage meters in the Intel monitor did not flicker to any extent but
the CPU meter did hit zero a few times from 1.5v for a fraction of a
second. On at least one occasion the CPU voltage flickered to zero while
the other meters stayed steady. I didn't hang around long enough to
decide if any damage was done. I just shut down. I have never seen this
happen in a computer up to now.
 
JD said:
Thanks Philo and John.

I was too concerned to check if the computer was working. All other
voltage meters in the Intel monitor did not flicker to any extent but
the CPU meter did hit zero a few times from 1.5v for a fraction of a
second. On at least one occasion the CPU voltage flickered to zero while
the other meters stayed steady. I didn't hang around long enough to
decide if any damage was done. I just shut down. I have never seen this
happen in a computer up to now.
Click to expand...


I suspect the problem is with the voltage monitor. as if the voltage on
your CPU dropped to zero...your machine would cease to function.

Now, if you are having power problems, the I'd suspect the power supply
itself or perhaps it's connection to the mobo
 
kony said:
Disable the CPU voltage detection part of your software as
it is not working properly. As others mentioned it is
impossible that it actually dipped down to zero, more likely
the speed is simply throttling back to save power and
voltage too, that voltage drops some but only enough to dip
below the detection threshold.
Click to expand...

Many thanks Kony. Do you think that Intel would develop a program
that cannot measure the voltage of the CPU, considering that it made the
motherboard, the CPU and the program? This is my first experience of
this problem.
 
Are you sure ? :-)
I suspect the problem is with the voltage monitor. as if the voltage on
your CPU dropped to zero...your machine would cease to function.
Click to expand...

Well, not so sure about that. Drops in the CPU powersupply don't
trigger the reset logic. Removing the power for a few seconds is
likely to toggle some bits inside, but I wouldn't at all be
surprised if the CPU just keeps on running (i.e. more or less
continue when power is restored.)
Now, if you are having power problems, the I'd suspect the power supply
itself or perhaps it's connection to the mobo
Click to expand...

1.5 volt supply for the CPU does not come from the PSU, at least
not directly. It runs from a local regulator, quite close to the
CPU socket on the motherboard and is fed by + 12 Volt.

First, I would look at the CPU datasheet to locate the power
pins, attach a DVM and see if the voltage realy drops.

Alternatively, you could take a look at the pictures on
http://badcaps.net and see if they resemble the caps near the CPU
;-)
 
Gerard said:
Are you sure ? :-)



Well, not so sure about that. Drops in the CPU powersupply don't
trigger the reset logic. Removing the power for a few seconds is
likely to toggle some bits inside, but I wouldn't at all be
surprised if the CPU just keeps on running (i.e. more or less
continue when power is restored.)


1.5 volt supply for the CPU does not come from the PSU, at least
not directly. It runs from a local regulator, quite close to the
CPU socket on the motherboard and is fed by + 12 Volt.

First, I would look at the CPU datasheet to locate the power
pins, attach a DVM and see if the voltage realy drops.

Alternatively, you could take a look at the pictures on
http://badcaps.net and see if they resemble the caps near the CPU
;-)
Click to expand...


No argument from me...
but the OP never mentioned that his machine hesitated.

My feeling is that if the voltage had really dropped ...then the
monitoring software would display the lowest voltage at the cpu when it
lost functionality.

Since the OP shut his machine down (safely?)
the cpu must have been working
 
JD said:
Many thanks Kony. Do you think that Intel would develop a program
that cannot measure the voltage of the CPU, considering that it made the
motherboard, the CPU and the program? This is my first experience of
this problem.
Click to expand...

Have you tried any other programs that monitor Vcore ? For example,
CPUZ might be able to do it. Perhaps Speedfan would be another example.

http://www.cpuid.com/cpuz.php

http://www.almico.com/speedfan439.exe

If you try enough test cases, maybe you'll find only the Intel
software is reading zeros. It's an illogical result - the CPU
can't function at zero volts. If it could, we wouldn't need
power supplies any more.

Paul
 
JD said:
I have a box where the CPU voltage sometimes dips to zero for a few secs
and then returns to the normal of 1.5v.

Has anyone seen this situation before? I am thinking of removing the CPU
and the local electrical connections around it and re-installing them.
The board is Intel and I have the Intel monitor for temperature and
voltage running.
Click to expand...

The voltage monitoring program has to be wrong because otherwise your
computer would have crashed each time the voltage read zero. I've
had problems with such programs. With my old ECS K7VTA3 mobo,
SpeedFan said the +12V rail alternated between +6V and +8V, about 2-3
times a second, while Hmonitor said the voltage was +10.6V. Neither
could have been right because the hard drive had been running fine all
along (digital multimeter read +12.11V or +12.04V), and HDs use +12V
for their motors. Also I tried the HD another system where the +12V
rail really did put out only +10.6V (verified with digital
multimeter), and it wouldn't even spin (old power supply design,
needed a heftier load than a 466 MHz Celeron system would provide).
 
larry said:
The voltage monitoring program has to be wrong because otherwise your
computer would have crashed each time the voltage read zero. I've
had problems with such programs. With my old ECS K7VTA3 mobo,
SpeedFan said the +12V rail alternated between +6V and +8V, about 2-3
times a second, while Hmonitor said the voltage was +10.6V. Neither
could have been right because the hard drive had been running fine all
along (digital multimeter read +12.11V or +12.04V), and HDs use +12V
for their motors. Also I tried the HD another system where the +12V
rail really did put out only +10.6V (verified with digital
multimeter), and it wouldn't even spin (old power supply design,
needed a heftier load than a 466 MHz Celeron system would provide).
Click to expand...

Just before I shut down, the 1.5v CPU Core voltage dipped to zero and
then came
back within a second or two. Meantime the 1.5v. CPU I/O stayed solid as
a rock.
The other 3 voltages 3.3, 5 and 12 also stayed solid.

This afternoon I disconnected the main connections around the motherboard,
then restored them, and started up. I'm happy to say that the box has
been as
solid as before. No dipping now.

My thanks to you all for your attention.
 
kony said:
No way for it to do that, because it's the CPU... bit
different than if the voltage to a sound chip dropped.

The moment the CPU voltage dropped below a stable threshold
the crash would stop everything, and if dipping down close
to zero I don't mean a bluescreen, it wouldn't even manage
that.
Click to expand...



Good point!
 
Good point!
Click to expand...

Beg to differ :-)
Video comes from a totally different system. GPU and often even
it's own RAM. No reason whatsoever to 'drop the picture'.
(It's the reset that does the harm. Remove +5 volt and the reset
circuitry kicks in. Remove the CPU core voltage and .....)
 
JD said:
Just before I shut down, the 1.5v CPU Core voltage dipped to zero and
then came
back within a second or two. Meantime the 1.5v. CPU I/O stayed solid as
a rock.
The other 3 voltages 3.3, 5 and 12 also stayed solid.

This afternoon I disconnected the main connections around the motherboard,
then restored them, and started up. I'm happy to say that the box has
been as
solid as before. No dipping now.

My thanks to you all for your attention.
Click to expand...



Thanks for posting back...and glad you got the problem resolved...
looks like you did have a bad connection
 
The picture could freeze, but it would freeze before the low
voltage event was displayed because the sensing and
reporting logic running on the CPU cannot retrive and
process that data.
Click to expand...
:-)
 
Back
Top